Re: [time-nuts] Allan variance by sine-wave fitting

[Jim Lux]

-Original Message-
>From: Magnus Danielson 
>Sent: Nov 27, 2017 2:45 PM
>To: time-nuts@febo.com
>Cc: mag...@rubidium.se
>Subject: Re: [time-nuts] Allan variance by sine-wave fitting
>
>Hi,



>
>There is nothing wrong about attempting new approaches, or even just 
>test and idea and see how it pans out. You should then compare it to a 
>number of other approaches, and as you test things, you should analyze 
>the same data with different methods. Prototyping that in Python is 
>fine, but in order to analyze it, you need to be careful about the details.
>
>I would consider one just doing the measurements and then try different 
>post-processings and see how those vary.
>Another paper then takes up on that and attempts analysis that matches 
>the numbers from actual measurements.
>
>So, we might provide tough love, but there is a bit of experience behind 
>it, so it should be listened to carefully.
>



It is tough to come up with good artificial test data - the literature on 
generating "noise samples" is significantly thinner than the literature on 
measuring the noise.

When it comes to measuring actual signals with actual ADCs, there's also a 
number of traps - you can design a nice approach, using the SNR/ENOB data from 
the data sheet, and get seemingly good data. 

The challenge is really in coming up with good *tests* of your measurement 
technique that show that it really is giving you what you think it is.

A trivial example is this (not a noise measuring problem, per se) - 

You need to measure the power of a received signal - if the signal is narrow 
band, and high SNR, then the bandwidth of the measuring system (be it a FFT or 
conventional spectrum analyzer) doesn't make a lot of difference - the precise 
filter shape is non-critical.  The noise power that winds up in the measurement 
bandwidth is small, for instance.

But now, let's say that the signal is a bit wider band or lower SNR or you're 
uncertain of its exact frequency, then the shape of the filter starts to make a 
big difference.  

Now, let’s look at a system where there’s some decimation involved - any 
decimation raises the prospect of “out of band signals” (such as the noise) 
aliasing into the post decimation passband.  Now, all of a sudden, the 
filtering before the decimator starts to become more important. And the number 
of bits you have to carry starts being more important.  And some assumptions 
about noise being random and uncorrelated start to fall apart.


It actually took a fair amount of work to *prove* that a recent system I was 
working on
a) accurately measured the signal (in the presence of other large signals)
b) that there weren’t numerical issues causing the strong signal to show up in 
the low level signal filter bins
c) that the measured noise floor matched the expectation
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Timestamps in audio files?

[Jim Lux]

On 12/21/15 3:19 AM, Tim Shoppa wrote:

As an adjunct to the thread about timestamped samples of LORAN
transmissions...

Are there any standard consumer-type audio file formats, that support
absolute time time/datestamps? Would not have to be done continuously, but
something like a time and date stamp inserted nearest each sample on a
second boundary.

I have worked with some analog tape audio formats in the past where
IRIG-type timestamps were written on a separate channel or on a subcarrier.

I know of many proprietary digital recording applications that make WAV's
or MP3's or proprietary codec formats, where the filename includes a
timestamp. Much more interested in standard formats where the timestamp is
embedded in the file itself.



For RF recordings, VITA49 has a standard for timestamps in the packet 
headers (4 flavors of epoch, multiple flavors of time format and precision)


Video file formats seem to draw from older time code things like SMPTE 
and are "relative" (so you're always fooling around trying to figure out 
the offsets).  I spent a few days earlier this year trying to put 
absolute time subtitles on video files using all manner of tools, and it 
was frustrating (ffmpeg, vlc, etc.. all were to no avail).  Trying to 
put UTC time into embedded timecode was also pretty unproductive (most 
tools don't like to see the first frame occurring at a time very 
different from 00:00:00:00)



In fact, in the music file world (e.g. MIDI) you see references to 
absolute and relative time, and there, they are really talking about 
time measured in seconds vs time measured in beats; e.g. whether the 
duration of something  is 1 second, or 2 quarter notes, which might be 
the same if the tempo is 120bpm.



You might look for solutions for people trying to synchronize multiple 
multimedia streams delivered over the internet (e.g. slides and 
accompanying narration or music) because they actually have a need for 
"show this slide at time HH:MM:SS and play this sound at HH:MM:SS" kind 
of synchronization.


I suspect, though, that this kind of info gets encapsulated in the 
transport layer, rather than the underlying files holding the info.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] timestamps on downconverted data streams

[Jim Lux]

On 12/20/15 12:17 PM, Bill Byrom wrote:

Yes, you can do that. My employer (Tektronix) makes RF Signal Analyzers
which sample at a high rate then use a DDC (Digital DownConverter based
on decimation and digital filtering) to produce a much smaller output
I/Q file at a smaller bandwidth and lower time resolution. The
decimation and digital filtering delay acquired signals, so you have to
correct for the group delay in these operations.
-


and, how do they do that correction? Or, do they leave up to you to do it?


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] timestamps on downconverted data streams

[Jim Lux]

Here's an interesting problem.

You have a fast sampler that is collecting samples off-the-air (e.g. the 
end of LORAN) with a fairly wide bandwidth: say 10 Megasamples per second.


Those samples get post processed in a digital downconverter (not 
necessarily in real time) to a narrower band representation at a lower 
sample rate.


You know when the input samples were acquired: e.g. you've got a good 
oscillator, and a reliable sync pulse.  For instance, your handy GPSDO 
(or the ensemble of H2 masers in your garage) might give you a 1pps tick 
good to, say, 20 ns, so you know when your 10 MHz samples were taken (to 
20ns)


Is there a consistent (and standardized) way to calculate and report the 
time of the output samples.


Each output sample is composed of information from multiple input samples.

One could test the system by digitizing a signal with known timing (e.g. 
a 1 MHz sine wave, where the zero crossing is "on the second") and then 
look for the zero crossing in the downconverted output.  Depending on 
the filtering in the downconverter, there's some time vs frequency 
characteristic that could be used to back out any deltas for other 
frequencies.


So you could report the time of the low rate output samples in terms of 
the time of the input sample, at least for the 'center frequency' of the 
downconverter.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Mechanical clock sound pickup circuit

[Jim Lux]

On 12/14/15 9:12 PM, ed breya wrote:

This may be totally ridiculous, but maybe there's another way to get a
balance wheel signal. The X-band Doppler type microwave motion detectors
can pick up various object signals in free air from quite a distance, so
maybe up close there would be enough resolution and penetration of the
metal parts of a timepiece to get a usable signal in and out. It would
tend to accentuate the fastest part of any motion - the balance wheel in
this case. I can picture setting one up with the horn pointed at the
thinnest part, likely the watch face, from maybe a few inches away - or
whatever it takes to not overload the detector. The audio detector
signal (if sufficient) could then be processed in the same way as with a
microphone sound signal.


As it happens, I have a fair amount of recent experience detecting small 
(<1mm) motions using radar.


Yes, remarkably tiny holes will let enough signal in and out, but, it's 
going to be very, very position dependent. You have a lot of multipath 
in this kind of testing, and it's easy to wind up in a null zone.


You might want to look for K-band (24GHz) units: the shorter the 
wavelength, the more phase shift you get from the tiny motion.  To put 
some numbers on it: at 3 GHz, a 1mm displacement gives you about 6-7 
degrees; at 24 GHz, you're going to be getting 50-60 degrees.


You'll be wanting some form of homodyne detector (which has the nice 
property that the phase noise of the source cancels out, so you can have 
a pretty grungy quality oscillator).  The signal you're looking for, 
though, is phase shifts occurring at a 1Hz kind of rate.  Most of the 
cheap "motion detectors" have a high pass filter  (1 m/sec at 3 cm 
wavelength is 66 Hz) and the amplifier chain is AC coupled.


You'll need a good low noise amplifier with a low 1/f knee.

For reference, a receiver gain of about 60 dB gives you a millivolt kind 
of signal from a 1mm displacement with 1mW at 3GHz from a 0.1 square 
meter target at 10 meter distance. You can scale to your situation.



You'll probably want some way to subtract out the static baseline, so 
your high gain amplifier stages don't need enormous dynamic range. In my 
radars, I do this with an adjustable "leakage" path from Tx to Rx.  You 
could probably do it with a movable metal target next to your 
clock/watch and you adjust it for a null.



You probably also want a I/Q output: if you think about the signal 
you're receiving, it's a slowly moving vector that spans a fairly small 
phase angle (because it combines a very large static response from stuff 
that's moving plus a little tiny moving component). If that vector 
happens to point at 90 degrees to your I axis only, then you're great: 
the variation shows up in the I axis. But if the vector happens to point 
parallel to the I axis, the motion is very small.


With I/Q, you can  either do a arctan demodulation, or you can rotate 
the signal to make the variation largest (basically using the sin x=x 
approximation for small x)









Ed
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPSDO and oscillator steering - EFC vs DDS schemes?

[Jim Lux]

(ADEV of 4E-16 at tau of 1000 seconds is a typical state of the art requirement)


…. and has been since the 1970’s when I first started talking with JPL people 
about this :)….


They've gotten a lot smaller and probably draw less power since then.

There's also the testing problem: proving that it has that performance 
is no easy matter.


It might well be that you could build a better one, but that other 
errors in the whole ranging measurement might dominate (movement of the 
antenna on Earth, etc.)


There's a lot of activity centered around making good flight atomic 
clocks with trapped Hg ions.  Those would essentially instantly improve 
over the USO.






All that said, the real question is — can you change the fabrication of the 
crystal in ways that improve it’s stability by
tuning a long way with DDS rather than a short way with reactance (select parts 
plus varicap(s)).



The DDS synthesis thing is nice because you can, to a certain extent, 
control where the spurs and noise are (at the cost of increased logic 
complexity, but that's essentially free, these days), and you can open 
up a trade space on crystals.  You don't have to have certain "special" 
frequencies.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPSDO and oscillator steering - EFC vs DDS schemes?

[Jim Lux]

On 12/9/15 4:37 AM, Bob Camp wrote:

Hi


On Dec 8, 2015, at 11:20 PM, Jim Lux  wrote:

On 12/8/15 3:31 PM, Bob Camp wrote:

Hi

Let’s see:

EFC uses reference out of the OCXO.
EFC comes on the OCXO at no added cost.
16 bit DAC costs ~$2 to $5

Total cost for EFC setup $2 to $5. Net result is a system with
spurs that are how ever far down you wish them to be. (It’s all
about grounding in this case).

Bob



If the OCXO has steering, the Q of the resonator has to be lower than if the 
OCXO wasn't steerable.


If the OCXO has an oscillator attached to the crystal, it has a lower Q than 
the crystal it’s self…..

The contribution of a “normal” (relatively narrow band) tuning circuit is 
actually quite small.


for very high performance oscillators, I'm not sure about that.  Here, 
I'm thinking about things like USOs where the crystal is in a double 
vacuum bottle with multiple heat shields, etc.


There's been several proposals from JHU/APL where a good oscillator is 
teamed with a high performance DDS so you don't have to get a crystal at 
the *exact* frequency you need. A great idea in my opinion 
(historically, the crystal frequency is tied to the channel allocation 
for your spacecraft, and non-adjustable frequency makes using spare 
oscillators from one mission for another one hard)


As good an idea as this is, it seems that (very risk averse) folks seem 
to stick with the "make lots of oscillators and pick the closest one to 
the desired frequency after initial aging".


The recent GRAIL mission that measured the moon's gravity used two USOs, 
one on each spacecraft, with the frequencies slightly different (so 
they're used as both Tx source, and LO for Rx for the signal from the 
other spacecraft).


A high quality DDS USO would have made this easier in many ways (you 
could cherry pick from the dozen or so oscillators for aging and phase 
noise properties, rather than also frequency)







So conceivably (if such things were available) you could get a non-steerable 
OCXO with better (very) close in noise.


Except when you actually wire up that circuit that’s not the outcome.



And then move the frequency with the DDS.  It's fairly straightforward to make 
a DDS circuit that pushes the spurs and such away from the carrier (at the 
expense of higher noise farther out).


Which gets you into a variety of spur and noise issues if you want those spurs 
to be below the noise floor of a good OCXO. Getting them into the -130 to -150 
db down range is far from trivial even
with the spreading stuff.



But hey, that's brand new and exotic.


And it pushes the spurs out to where the noise floor should be -170 or -180 … 
hmmm ….


But there are applications where far out noise isn't as important, for 
instance, in a deep space transponder used for ranging. The transponder 
is basically a phase locked loop with a very narrow loop bandwidth (a 
few Hz).  And the receiver on the ground is also very narrow band, so 
noise that's say, 10 kHz away, isn't a big deal, compared to noise 
within a few Hz, which is.


(ADEV of 4E-16 at tau of 1000 seconds is a typical state of the art 
requirement)







Bob



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPSDO and oscillator steering - EFC vs DDS schemes?

[Jim Lux]

On 12/8/15 3:31 PM, Bob Camp wrote:

Hi

Let’s see:

EFC uses reference out of the OCXO.
EFC comes on the OCXO at no added cost.
16 bit DAC costs ~$2 to $5

Total cost for EFC setup $2 to $5. Net result is a system with
spurs that are how ever far down you wish them to be. (It’s all
about grounding in this case).

Bob



If the OCXO has steering, the Q of the resonator has to be lower than if 
the OCXO wasn't steerable.


So conceivably (if such things were available) you could get a 
non-steerable OCXO with better (very) close in noise.


And then move the frequency with the DDS.  It's fairly straightforward 
to make a DDS circuit that pushes the spurs and such away from the 
carrier (at the expense of higher noise farther out).


But hey, that's brand new and exotic.

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPSDO and oscillator steering - EFC vs DDS schemes?

[Jim Lux]

On 12/8/15 8:32 AM, Attila Kinali wrote:

Moin,

I've been digging through some stuff and stumbled (again) over Rick's
paper on high resolution, low noise DDS generation[1] and got confused.
The scheme is very simple and looks like to be quite easy and reliably
to implement. If I understood it correctly, the critical points are the
DDS, its sideband generation and the LO/RF feedthrough in the mixers.
Nothing that is not known and nothing that is too difficult to handle
(the 10.7MHz filter get rid of most of the feedthrough already and
there has been a lot written on how to design DDS for specific applications).

What puzzled me is, why this has not been used more often to correct
the frequency of OCXOs instead of using some DAC-to-EFC scheme?



Heritage... if you have a design that works, and there's a lot of them 
in the field, and the idiosyncracies are well known and understood, then 
one tends to stay with the old design.


DDS are "brand new", at least in terms of generating low spurs, etc. 
The idiosyncracies are not as well understood.


I think also the power consumption might be an issue.  Most good DDS 
burn a lot of power, compared to a DAC.


There's also systems that depend on smooth sweeps without steps (yes, 
one can design a DDS with a digital ramp generator driving the increment 
in a phase accumulator to get arbitrarily smooth sweeps, but the "off 
the shelf" parts don't do this)


I don't think parts cost is a big driver.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] RG 6 U couplings

[Jim Lux]

On 12/5/15 12:28 PM, Bert Kehren via time-nuts wrote:

At my new home the GPS antenna location has turned in to a  challenge.  May
have to splice RG 6U. Has any one done measurements on  couplings and the
loss associated with them. Right now I am considering a  female, female
coupling. Is there a better alternative?


The couplings have almost no loss.  The data sheet will say something 
like <0.05 dB, but that's because that's what the measurement 
uncertainty is.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPS down converter question

[Jim Lux]

On 12/1/15 6:41 PM, Bob Camp wrote:

Hi

So back to the question …. does a 70 (ish) MHz fancy filter really buy you
anything ahead of the main box? If you will be multiply band limited ahead
of the mixer (antenna and saw), the contribution of the 70 MHz filter will
likely be minimal. Note that I’m comparing a filter that costs money to a
L-C tank that is essentially free rather than truly no filter at all.



One other thing to consider - Most antennas have a *lot* of gain in them.
Back when I took the course, the standard advice was to put as little gain
ahead of your mixer as possible. It’s all tied up in the impact gain has on
the distortion products. That also would drive you towards a filter ahead
of the mixer approach.

Bob






Do you have strong interfering signals close by that might be an issue.. 
70 MHz is in the low VHF TV band.  These days, in the US, TV is all 
digital, so you don't have that big carrier to worry about, but a lot of 
GPS receivers that use 1 bit ADC are really hit hard by a narrow band 
signal that is big enough to "capture" the sampler.


JPL's receivers use a fairly broad L-band filter (covers L1, L2, and L5) 
before the LNA, then narrower filters for each subband before the 
sampler. Strong L-band signals are a real problem if they cause 
intermods in the LNA.  In space, that's not usually an issue, but in 
terrestrial testing of these receivers it is.  And in receivers designed 
for precision geodesy or timing, they typically hate to have narrow band 
filters, because of the delay through the filter.



They're direct sampling, so conversion to IF isn't an issue for JPL, but 
if you are doing a down conversion, if you have a strong L-band 
interferer that is either generating intermods, or is at the image 
frequency, you might run into an problem.


A


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Einstein Special on PBS

[Jim Lux]

On 11/29/15 12:13 PM, Mark Sims wrote:

Another thing to consider is the gravity anomaly caused by that hunk of granite 
beneath your clock (or above it in a mine).   Hmmm, what is the clock shift at 
the top of Mt Everest that is due to the mountain and not the altitude? 
 
___
t



when going deep..
the Homestake mine in South Dakota is something like 2500m (8000 ft) 
deep.  There's a DoE funded research facility somewhere deep for nuclear 
experiments.  There could well be atomic clocks already down there.



https://en.wikipedia.org/wiki/Sanford_Underground_Research_Facility

https://en.wikipedia.org/wiki/Large_Underground_Xenon_experiment is 
about halfway down (I think the lower part of the mine is still flooded)


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] time sync from cellphone TO microcomputer

[Jim Lux]

On 11/21/15 2:43 AM, Attila Kinali wrote:

On Fri, 20 Nov 2015 15:47:27 -0800 Jim Lux 
wrote:


While this is trivial with a GPS receiver, we were thinking about a
very minimalist implementation, with a smart phone as the control
interface.


So, some electronics with a micro controller that do the actual
modulation/demodulation and a smart phone to control that?


something along those lines.. maybe the phone does the mod/demod (there
are PSK31 apps, for instance)..






That brought up another discussion about whether it is possible to
get time sync (at a better than 1 second, but potentially at the
milliseconds level) across bluetooth or WiFi from a cellphone
(Android or iDevice).


All these wireless protocols are relatively low delay. Most of the
handshakes (which cause delay) is done during "connect" (peering or
whatever they call it) and the "master" forces the "slaves" to
follow its clock. Unfortunately, this synchronization is not
accessible from the higher levels. But as I said, the delay is rather
low.


And one could probably do a round trip measurement once you're connected.




If you want to use a uC as receiving partner, i would go for BT/BTLE
rather then wifi as it's easier to setup. There are wifi modules
that behave kind of the computer modems of old (complete with AT
command set) to connect to an AP and intiate a TCP connection. But if
you want to go for a smartphone, it makes more sense to use BT/BTLE
as you can directly connect the smartphone to the device, without
first needing to find it in the network. BT was originally started as
a wirless 9600 serial connection for cell phones and has retained
much of that property in the protocol. You can still use it as such
and that works pretty much with every cell phone today (if it
supports BT). BTLE (or BT smart as it is officially called) is it's
modern cousin of sorts. The on-air protocol has been completely
rewriten and as such should count as a different system, but it
retains an API that is very similar to that of BT.

Although the standards are freely available, they are not an easy
read. I used the book "Bluetooth Low Energy: The Developer's
Handbook" by Heydon 2-3 years ago to get the basic understanding.


yes, BT was what we were thinking..






In terms of modules you can get to day, the cheapest are the BTLE 4.x
modules. Some of them support "regular" BT, but not all. The usual
interface to those is an UART, sometimes SPI. Get at least a 4.1
module, as they fixed a couple of bugs in the protocol of version
4.0.


of course, as I've found, not all phones and iDevices (particularly
older ones, like my old iPad 1) support BTLE




As for delay, that's a bit a more difficult question. BTLE is
optimized for low power use, as such has higher delay than BT. The
poll interval of the slave can be set to anything between 7ms to 4s.
There are a few other parameters that control polling interval, but
my memory is hazzy there.

I've never actually measured what delays one can achieve, but given
that we were able to transfere data in small chunks pretty quicky, i
guess it could not be much over 10ms.

A word about compatibility: All Android Phones support BT, but BTLE
support came only around 2012/13. I think it's safe to assume that
any Android phone after 2013 supports BTLE. With Apples iPhones it
was the other way round. They supported BTLE early on (4s in 2011)
but didn't support BT until much later (don't remember exactly, i
have something like 2013 in my head).







As a stack on the uC we used bluektichen [1,2] which did a pretty
good job. The author is also very helpfull and easy to talk with. I'm
pretty sure he'd help on non-commercial projects as well. (for
commercial projects, his prices are pretty cheap)

BTW: this information might be important to select the right
sub-protocol[3]: --- So far, the most popular use of BTstack is in
peripheral devices that can be connected via SPP (Android 2.0 or
higher) and GATT (Android 4.3 or higher, and iOS 5 or higher). If
higher data rates are required between a peripheral and iOS device,
the iAP1 and iAP2 protocols of the Made for iPhone program can be
used instead of GATT. Please contact us directly for information on
BTstack and MFi. ---

MFi means here "Made for iPhone" which is Apples way of extorting
money from device manufacturers. You have to buy an Apple crypto chip
for that and put it in each device. And they wont talk to you unless
you buy a few millions a year. Stay clear of that.



Useful to know..


So, probably reasonable to get 1 second level sync, sub 1 millisecond is
probably not possible, somewhere in between is plausible.

Well, for the things we were talking about that's probably plenty good.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] time sync from cellphone TO microcomputer

[Jim Lux]

A couple of us were kicking around the idea of a low cost JT65-like HF 
modem implementation, which requires that the station be synchronized to 
1 second.


While this is trivial with a GPS receiver, we were thinking about a very 
minimalist implementation, with a smart phone as the control interface.


That brought up another discussion about whether it is possible to get 
time sync (at a better than 1 second, but potentially at the 
milliseconds level) across bluetooth or WiFi from a cellphone (Android 
or iDevice).


Say you wanted to "set the clock" on an Arduino or embedded processor. 
Could you do it via a wireless (or, I guess, a wired connection).


Has anyone done it?  the iPhone runtime environment is pretty sandboxed, 
and you might not be able to do that.






___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Downsizing dilemma, HP 3335A

[Jim Lux]

On 11/11/15 3:26 PM, Rob Sherwood. wrote:

The EE department at the University of Colorado has an enlightened professor.

http://ecee.colorado.edu/faculty/popovic.html

Zoya required her students to not only get a ham license, but to build a Norcal 
40A.

http://ecee.colorado.edu/~ecen2420/Files/NorCal40A_Manual.pdf



There's actually quite a few EE profs that do this.  Dave Rutledge at 
Caltech also does NorCal radios.


Shar Katz and James Flynn at Cal State Northridge also encourage ham 
licenses, they're more doing wireless comms on VHF/UHF.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Beginners GPS locked frequency counter question

[Jim Lux]

On 10/31/15 7:32 PM, Bob Camp wrote:

Hi

So … how good is the “calibrate and go” (not the tone on second channel) 
approach likely to be?

If it’s a bare crystal or normal XO (not a TCXO) that is supplying the clock, 
the crystal will follow some
fairly well known curves. Which one of the curves it follows will depend 
entirely on the individual sample
you have. What I see on my sound card will not be what you see on a different 
sample of exactly the same
sound card installed in a similar machine.



having just done a bunch of measurements over 4 days on some crystals of 
this type (16 MHz, nothing special), I can shed some light on it.


over half a day sitting on my desk, they fluctuated about 0.1 ppm  (peak 
excursion)


AVAR is about 1E-9 at tau=10sec, rising to 2e-7 at 10,000 seconds (after 
taking out the temperature dependence)


There is a strong temperature dependence.. with temperature changes of 6 
degrees (C) the frequency change is on the order of 6ppm (90 Hz out of 
16 MHz)






It’s a pretty good bet that your device will change by 0.1 to 1 ppm per degree 
C. Yes it can be a bit better. It
can also be a whole lot worse. Your room environment is likely to move 1 to 3 C 
per hour if you have the heat
or air-conditioning turned on. If you are in an unheated garage … who knows. If 
you take the most likely 0.5 ppm/C tempco
and the most likely 2C change you get a “wobble” of about 1 ppm. The period of 
this wobble probably will be
in the 30 to 90 minute range.



yes, that's about right..
Although my office doesn't have that kind of fluctuation..
And I ran the 4 day test with the oscillators in a cardboard box over 
the weekend, so that damps short term variations.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Beginners GPS locked frequency counter question

[Jim Lux]

On 10/31/15 3:50 AM, Chris Wilson wrote:



   31/10/2015 10:46

I have a Racal counter locked to 1 MHz on its rear panel external
input socket from my Trimble Thunderbolt GPS. I derive the 1 Mhz
from a David Partridge divider board. If I also feed the counter
with the 10 Mhz direct output from the same GPS it reads one or two
Hz out. As I assume the counter is working purely mathematically
why would that be please?


As an aside, I work low frequency RF transmissions on 136 Mhz, and
very narrow bandwidth. Can a soundcard be locked to GPS instead of
its own internal crystal for precise frequency output?




you could probably get a DDS of some sort to take the high quality 10MHz 
and turn it into the clock frequency of your sound card. You'd need to 
figure out how to pull the XO off the board and feed an appropriate 
signal in.


There's a fair number of people interested in this, so I suspect someone 
has already done it.


The other approach is to replace the XO on sound card with a VCXO, and 
"discipline" it against the 1pps from the GPS.  It's a matter of using a 
different divisor than 10,000,000.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low noise quartz crystal oscillator by Bruce Griffiths

[Jim Lux]

On 10/28/15 4:29 PM, Adrian wrote:

That's chapter 6 of his book.
http://rubiola.org/indexx-oscillator-noise.html
Just scroll down for the phase noise plots.
The left hand column of plots contains the essentials.

Adrian





what would be nice is some similar simple analysis for lower performing 
oscillators.. (like the oscillator in a PC, or run of the mill 1-10 ppm 
TCXOs)



tvb has some on his website.

I guess there's stuff around, but it's not all gathered together.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] The Pendulum Paradigm by Martin Beech, 2014

[Jim Lux]

On 10/28/15 7:48 PM, Jim Lux wrote:

On 10/28/15 7:23 AM, Peter Reilley wrote:

I have been pondering pendulum clocks.   I was wondering what the ADEV
of a
pendulum would show.   I assume that you could see the errors in the
gear train.
You should see the period of each gear.   You should see the spring wind
down
and being rewound.

Further, would you be able to see the phase of the moon and the tides?
This
is using the pendulum as a gravimeter.   Would it be sensitive enough
for that?



yes.. it's in the sub-ppm range, as I recall.

Period goes as sqrt(L/g)


from wikipedia
lunar tidal acceleration at the Earth's surface along the Moon-Earth
axis is about 1.1 × 10−7 g, while the solar tidal acceleration at the
Earth's surface along the Sun-Earth axis is about 0.52 × 10−7 g


So sqrt(1/(1+1E-7))... about 0.05 ppm



so you'd need an ADEV <1E-9 at a tau of 12 hours



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] The Pendulum Paradigm by Martin Beech, 2014

[Jim Lux]

On 10/28/15 7:23 AM, Peter Reilley wrote:

I have been pondering pendulum clocks.   I was wondering what the ADEV of a
pendulum would show.   I assume that you could see the errors in the
gear train.
You should see the period of each gear.   You should see the spring wind
down
and being rewound.

Further, would you be able to see the phase of the moon and the tides? This
is using the pendulum as a gravimeter.   Would it be sensitive enough
for that?



yes.. it's in the sub-ppm range, as I recall.

Period goes as sqrt(L/g)


from wikipedia
lunar tidal acceleration at the Earth's surface along the Moon-Earth 
axis is about 1.1 × 10−7 g, while the solar tidal acceleration at the 
Earth's surface along the Sun-Earth axis is about 0.52 × 10−7 g



So sqrt(1/(1+1E-7))... about 0.05 ppm

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPS Crystal Frequencies

[Jim Lux]

On 10/26/15 7:06 PM, Brooke Clarke wrote:

Hi:

I understand 10.230 MHz since when multiplied it gives 1176.45,1227.60,
1381.05 & 1575.42 MHz, all GPS carrier frequencies.
http://www.prc68.com/I/DAGR.shtml#GPSs
But I've got a number of GPS receivers that have Rakon unit oscillators
with a frequency of 10.949297.
http://www.prc68.com/I/DAGR.shtml#Polaris_Link
What's the story?



you might choose a frequency that when used as a direct sampler, causes 
the GPS signal to alias to a convenient offset frequency (then all 
Doppler shifts wind up being positive, for instance).



try 38.65 MHz for instance, that makes both L1 and L2 alias down to 
around 9.2 MHz, which is slightly less than 1/4 of the sample rate, and 
nicely accommodates the bandwidth of the signal


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Celestial Navigation instruction being reinstated in the US Navy

[Jim Lux]

On 10/25/15 9:37 AM, jim s wrote:


Somewhat time related.  The Navy realizes that GPS might not always
work.  I don't imagine that aircraft in the US Air Force will be able to
do this very reliably, and the article doesn't mention that service. I'm
guessing that a lot of strategic Air Force aircraft have star trackers
that will work some of the time w/o GPS (at night).



There's an excellent set of CD-ROMs with about 50 papers on celestial 
nav and time keeping from the Institute of Navigation.


https://www.ion.org/publications/upload/CelestialNavTOC.pdf

Papers in there about all manner of star trackers and celestial nav, 
from prehistory through the Renaissance era, to modern computerized 
celestial nav boxes, etc.


$50, as I recall.

Celestial nav during the daytime isn't all that hard, if you have a 
suitable telescope.  With a 28x telescope on a theodolite, you can see 
Polaris, for instance.  The trick is in finding it first.





http://www.latimes.com/nation/la-na-celestial-navigation-20151025-story.html


Thanks
Jim
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] looking for rakable GPSDO

[Jim Lux]

On 10/19/15 7:10 AM, Stéphane Rey wrote:

Hello,

I'm looking for a 10 MHz output GPSDO with external antenna which would
be rackable. Symmetricon doesn't seem to propose some neither Keysight.


MicroSemi now has the product line.. They've got tons of rack mounted 
GPS disciplined stuff.
We have a lot of XL-GPS at JPL (and their predecessor units like the 
XL-DC, etc.)


You can get them with a variety of internal oscillators in terms of 
phase noise, multiple outputs, different frequencies, etc.


you can get better ovenized oscillators as an option (1E-11 at 100 
seconds) vs the regular OCXO at around 2E-10 at 100 seconds.




Found some stuff  in Oscilloquartz. Any other brand to suggest ?

Thanks & cheers
Stephane

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Digital Phasae Lock Loops

[Jim Lux]

On 10/17/15 6:17 AM, Alex Pummer wrote:


actually, that is a ketch 22, if the loop bandwidth is to low, you will
have low noise , but it may will not lock at all, an other way to try to
filter out the noise, also you may make the loop filter digital, but
leave the the PLL analog, that could have  the possibility to have the
advantage to be able to change the loop bandwidth  increase for locking,
and reduce after the detected locking


changing  loop bandwidth between acquisition and tracking, or, 
similarly, (effective) loop bandwidth that changes with SNR are pretty 
common strategies.


In the deep space transponder world (where you are acquiring and 
tracking a very narrow carrier at -155 or -160 dBm against noise of 
-170dBm/Hz) you also want to know what order filter you should be using 
in the tracking loop.  If there's an expectation that the frequency 
being tracked is changing (e.g. Doppler), then a low order loop may not 
be the best choice.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Digital Phasae Lock Loops

[Jim Lux]

On 10/16/15 7:45 AM, Martyn Smith wrote:

Hello,

I want to design a digital phase lock loop.

I intend to lock a 10 MHz ultra low noise oscillator that we make to an 
external frequency standard.

I need a digital PLL as I’m trying to get a loop bandwidth < 0.1 Hz.



are you locking the oscillator via a voltage to the oscillator?
How would you derive the error signal between your 10 MHz and the 
external standard?   Some sort of phase/frequency detector?
I would think that any of the GPS disciplined oscillator schemes that 
have been discussed on the list would be a good start.


You could also use your oscillator as the clock for an ADC that 
digitizes the external standard (or vice versa), which basically makes 
the ADC the "mixer".


Once you've got an error signal, it's a matter of an appropriate loop 
filter driving a DAC.






Has anyone had any experience of Digital PLL’s or can point me to any documents 
published?

Regards

Steve
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Iridium source?

[Jim Lux]

On 10/7/15 1:16 AM, Tom Van Baak wrote:

Stu,

Thanks for the heads up. If you can leak anything from the upcoming paper 
please let us know.

Since you're an Iridium expert, would you be able to answer the question? The 
OP (John Todd) was asking about reception deep indoors, where GPS signals fail. 
In case you deleted it, the posting is here:
https://www.febo.com/pipermail/time-nuts/2015-October/094146.html




I would worry about the varying propagation time through the building. 
Over the past couple years, I've been propagating 3 GHz signals through 
a variety of buildings, piles of rubble, and so forth.


There's two effects I can think of with respect to timing:
a) the propagation path is highly scattered, and there is a lot of 
fading (on a scale comparable to wavelengths.. so if the satellite is 
moving at 7 km/s, and the wavelength is 20 cm, you're going to get fades 
at 35 kHz).. More to the point, the phase shift is wildly fluctuating in 
the short term, but pretty stable in the long term average sense.


b) The "average" path length is going to change as the satellite to 
ground terminal angle changes, because you're going through more or less 
of the building (unless the building happens to be a big symmetrical 
radome that you are in the center of)


Dense rubble has a propagation velocity of about 50-70% of free space. 
Buildings are pretty close to free space: the effect is one of bouncing 
around in an indirect path, rather than going through walls, although 
there's some of both effects.


In any case, at 1 foot/nanosecond in free space, going through, say, 3-4 
feet of construction material at 0.70 Velocity is going to add a 
nanosecond to your time delay.


The attached is a modeled gaussian pulse propagating through 5-7 meters 
of mixed rubble (I don't recall the exact geometry.. the rubble pile is 
5x5x3 meters in this model, and I think the transmitter was at one 
corner and the receiver is buried near the opposite corner)



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] advice for buying VCXO

[Jim Lux]

On 10/6/15 12:46 PM, Can Altineller wrote:

Hello,

I was hoping for a cheap VCXO. I think those ovenized ones are more
expensive.

Where would I get a brand new VCXO, that has < 1ppm, and specs that are
true. (as in the declared spec would match the actual performance, these
days when we buy from ebay, or aliexpress this is hardly the case)



There's tons of vcxos from Digikey.. just searching for VCXO and 
restricting it to 10 MHz turns up about 17 parts.


choosing the first one in the list
http://www.foxonline.com/pdfs/FVXO_HC53.pdf

50ppm pulling range

This happens to be a ceramic resonator, so the close in noise is 
probably not so hot.


But some checking at Digikey and Mouser and the like should turn up 
soemething.



How about an Abracon ASVV series SMT at 20 MHz for $2.98 Qty 1?

or a VCTCXO AST3TQ-V-10.000MHz-28 for $19.25 Qty 1 from Digikey.
280 ppb from -40 to +85
10 ppm pulling range..


Abracon is in southern california, for what it's worth.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] advice for buying VCXO

[Jim Lux]

On 10/6/15 5:46 AM, Can Altineller wrote:

Hello Jim,

Yes i need a VCXO or even a VCTCXO. However when I search them on ebay,
unfortunately, since they included every term for XO, I am lost in the
noise.

Any recomendations for buying VCXO's that are surplus, and not ebay?





do you want a Ovenized VCXO? or any VCXO? Is there a particular 
performance you need?  There's lots of inexpensive brand new VCXOs 
available.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] more teensies

[Jim Lux]

On 9/25/15 11:40 AM, Jim Lux wrote:

I ran 5 teensys in parallel, driven from the same Rb source for an hour..
They track reasonably well.

and AVAR from the same run

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] teensy as time capture device

[Jim Lux]

On 9/24/15 11:02 AM, Hal Murray wrote:


xne...@luna.dyndns.dk said:

External Oscillator (the system clock clock) ,  or External Timer clock
(limited to system clock/4)


That sounds like they are running the external signal through a synchronizer
and then doing all the logic on the system clock.  That would add a lot of
high frequency jitter but work as expected at longer time scales.

It would be fun to see if you could pick up hanging-bridge type artifacts.




Almost certainly.

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] teensy as time capture device

[Jim Lux]

On 9/24/15 7:13 AM, cfo wrote:

On Thu, 24 Sep 2015 05:30:35 -0700, Jim Lux wrote:




What would be interesting is if there's a pin on the Arduino/Teensy that
you could feed a high quality oscillator to, and then do counting with
that.  The K20 microcontroller has a mindbendingly large number of
features and alternate pin functions.


As i see it, from the DS. , it seems like there are 2 options.

External Oscillator (the system clock clock) ,
or External Timer clock (limited to system clock/4)

Electr. specs
https://www.pjrc.com/teensy/K20P64M72SF1.pdf

Family Ref
https://www.pjrc.com/teensy/K20P64M72SF1RM.pdf


Seems like the EXTAL pin (main oscillator) accepts an Oscillator signal
(max VDD + 0.3v) (Electr specs pg.10)

Ext Osc Connection (Family ref. pg. 526)

Also see PLL jitter (Electr specs pg.27 table 14)


There is a possibility to clock the FlexTimerModule by an external clock
(FTM_CLKIN/EXCLK - Max freq is Mainclock/4) ,
see (Family Ref. pg. 113, 214 & 774)


If using EXTAL and USB , you want to make sure that the USB part still
gets it's 48Mhz (via the PLL mul/div).
Maybe divide 10Mhz (XTAL) by 2 , and mul by 48, in the PLL.


I would think that the FTM_CLKIN would be the way to go.. keep the core 
running at 48 and run 10 MHz in as EXCLK (since it's less than Mainclock/4)


Not sure I'd want to fool with the PLL programming: that's something 
where there might be other stuff that makes assumptions about how it's 
configured.



Now I have to look if the right pins come out to something I can see.
This is the MK20DX256VLH7 configuration.
Only FTM1 (0,1) and FTM2 (0,1) are configured I think.
"There are two external FTM_CLKINx pins that can be selected by any FTM 
module via the SOPT4 register in the SIM module."


Looks like pin 32 and 33 on the chip




Tnx
Jim
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] first teensy3.1 data

[Jim Lux]

On 9/23/15 2:50 PM, Magnus Danielson wrote:

Jim,

There is some systematic bumps in there which make me wonder what
happens here. Care to share data/plots for phase?


Sure..
This is just the part sitting on my desk with the source hooked up with 
clip leads. The run is very short (5-10 minutes) so there's, at the 
least, some thermal effects.



Over that 10 minutes, with the Rb, the frequency slowly decreases (about 
0.1 ppm over that time span).


This thing is also powered off the USB from the computer it's hooked up 
to, and the environment is hardly quiet.


I'll get some data files and a better test setup.



You want to understand the systematics when it looks like that.

Cheers,
Magnus

On 09/23/2015 09:09 PM, Jim Lux wrote:

a bit more than 5 minutes of data.

Now to go get a real 1pps source that's decent.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] teensy as time capture device

[Jim Lux]

On 9/23/15 10:08 PM, Poul-Henning Kamp wrote:


In message <5602f65b.6050...@earthlink.net>, Jim Lux writes:


I've got a teensy3.1 hooked up to a 33120 function generator (not
exactly a super stable device) and generating period data for a 1 Hz
square wave.


That is actually a particularly bad way of doing it, because the
33120 will generate a sine and run it through a zero-detector.

You get much better performance by defining a ARB function which is
a square.




Interesting.. in any case, it was more about seeing if it worked, and 
the 33120 was handy.


Now I've got the Rb hooked up, and it's a 10 microsecond long pulse.


I'm not sure how useful, overall, this is.   The crystal on the Teensy 
is probably nothing special.


I did do a quick run with the Teensy in 2x overclock mode (96 MHz CPU 
instead of 48MHz): the code returns counts that are close to 48,000,000, 
so either the counter runs off the crystal, or the code is scaling it 
internally.


What would be interesting is if there's a pin on the Arduino/Teensy that 
you could feed a high quality oscillator to, and then do counting with 
that.  The K20 microcontroller has a mindbendingly large number of 
features and alternate pin functions.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] more teensy timer stuff

[Jim Lux]

Got the Rb hooked up.
1E-8, 1E-9 kind of AVAR (after linear trend removal) for 10 minute run.


next, we'll try running it clocked at 96 MHz
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] first teensy3.1 data

[Jim Lux]

a bit more than 5 minutes of data.

Now to go get a real 1pps source that's decent.
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] teensy as time capture device

[Jim Lux]

I've got a teensy3.1 hooked up to a 33120 function generator (not 
exactly a super stable device) and generating period data for a 1 Hz 
square wave.


The period is in "ticks" of the 48 MHz clock, so my thinking is that if 
I hook up a good 1pps, what I'm really measuring is the frequency of the 
CPU clock, so I can just accumulate some data, and then load it into 
timelab.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] algorithms and hardware for comparing clock pulses

[Jim Lux]

On 9/21/15 12:13 PM, Robert LaJeunesse wrote:

The Teensy 3.1 (http://www.pjrc.com/store/teensy31.html ~$20) has a Flex Timer 
Module that appears to allow a single counter to be captured into independent 
registers from independent inputs. Not sure, but PJRC tends to run the clock 
fast (96MHz) so relative timing resolution should be much better than 0.1 
microseconds.



A bit of googling:  FreqMeasure library for the Teensy can time stamp 
zero crossings on pin 3
Apparently, the code is there to use up to 4 pins, but it's not really 
set up for multiple pins.
It uses the hardware to capture, and then an ISR to unload the register 
and buffer them up.


It also works on Arduinos.


https://www.pjrc.com/teensy/td_libs_FreqMeasure.html

the diagram in the K20 manual shows a two stage D flip-flop synchronizer 
driven off the system clock, and then a simple rising/falling edge (or, 
really, a 0->1 or 1->0 transition detector, which then latches the counter.

(Figure 37-175 in the manual)

Is this "time-nuts" precision capable. I've not tried to drive an 
Arduino or teensy with an external clock, which I think might be a 
starting point.




Bob LaJeunesse


Sent: Monday, September 21, 2015 at 1:45 PM
From: "Can Altineller" 
To: "Discussion of precise time and frequency measurement" 
Subject: [time-nuts] algorithms and hardware for comparing clock pulses

Dear Time-nuts,

  ...
I probably need a hardware to measure the time pulses more precisely.
  ...
Are there any solutions to this problem?
  ...
Best Regards,
Can Altineller
___

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] algorithms and hardware for comparing clock pulses

[Jim Lux]

On 9/21/15 12:13 PM, Robert LaJeunesse wrote:

The Teensy 3.1 (http://www.pjrc.com/store/teensy31.html ~$20) has a Flex Timer 
Module that appears to allow a single counter to be captured into independent 
registers from independent inputs. Not sure, but PJRC tends to run the clock 
fast (96MHz) so relative timing resolution should be much better than 0.1 
microseconds.



I was looking at the discussion of the Flex Timer on the teensy forum..
One might want to be careful about the quality of the clock fed to those 
timers.  It's the regular old CPU clock, but it runs through a DPLL.
(e.g. the crystal is a 48 MHz crystal, and converted to either 72 or 96 
MHz as you select..)


I run my teensys at 48MHz, so I could hook up a 1pps to a pin and log 
some data pretty easily.  I suspect that the ADEV will be dominated by 
the CPU crystal, so I can use any convenient 1pps.





Bob LaJeunesse


Sent: Monday, September 21, 2015 at 1:45 PM
From: "Can Altineller" 
To: "Discussion of precise time and frequency measurement" 
Subject: [time-nuts] algorithms and hardware for comparing clock pulses

Dear Time-nuts,

  ...
I probably need a hardware to measure the time pulses more precisely.
  ...
Are there any solutions to this problem?
  ...
Best Regards,
Can Altineller
___

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] algorithms and hardware for comparing clock pulses

[Jim Lux]

On 9/21/15 12:13 PM, Robert LaJeunesse wrote:

The Teensy 3.1 (http://www.pjrc.com/store/teensy31.html ~$20) has a Flex Timer 
Module that appears to allow a single counter to be captured into independent 
registers from independent inputs. Not sure, but PJRC tends to run the clock 
fast (96MHz) so relative timing resolution should be much better than 0.1 
microseconds.


I have, literally, a box full of teensy 3.1s at work.
If there's a quick test that would help answer any questions, I'm 
willing to set them up.
I have, in the lab with the Teensys, a SRS Rb and a Wenzel OCXO (not a 
superduper, just the streamline).


Is there a Flex Timer Module in the teensyduino library?



Bob LaJeunesse


Sent: Monday, September 21, 2015 at 1:45 PM
From: "Can Altineller" 
To: "Discussion of precise time and frequency measurement" 
Subject: [time-nuts] algorithms and hardware for comparing clock pulses

Dear Time-nuts,

  ...
I probably need a hardware to measure the time pulses more precisely.
  ...
Are there any solutions to this problem?
  ...
Best Regards,
Can Altineller
___

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] looking for SMT oscillator SC cut, with no oven

[Jim Lux]

On 8/29/15 7:19 PM, Alex Pummer wrote:

Hi Bob,

go to your local city library get membership[ here in California it is
free] , and ask them to get from the university  library, it will take
some  time than they cal you the your stuff is there, you could have it
for  four weeks if you need you could extend it for an other four weeks,
the engineering library of the university of Berkeley is open to
everybody, you can not take it out without additional formality, but you
could read, copy, scan it there,
I assume that works similarly in your state/ city/ university library,
If you have a specific title, let me know, it will not happen right
away, since I am working on five projects [for clients] also I am [life]
member of the IEEE, where is not everything free any more, but people
are reasonable
73




As a Californian, I thought similarly.. all the UC libraries are open to 
the public and you can get free access to online resources (e.g. IEEE 
Xplore) via free public workstations; although printing stuff costs 
money.  There might be visiting hour restrictions for the general public 
(no showing up at 3 AM), and most of them do require some kind of photo ID.


However, a bit of casual browsing shows that this is decidedly NOT the 
case in other states.  The Ohio State University, as far as I can tell, 
requires you to be a member of "Friends of the Library", which is not 
free.  It was unclear whether free access to Univ of Washington 
libraries includes online access (on-site).


Fascinating.

Local public libraries vary (even in California) on their interlibrary 
loan/ability to request copies of articles. It depends on local budgets 
and politics.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] equations for crystals, frequency vs temp vs angles

[Jim Lux]

Someone was asking about changing cut angles and the effect.  You might 
find some useful stuff in Mark Haney's thesis "Design Technique for 
Analog Temperature Compensation of Crystal Oscillators"


http://scholar.lib.vt.edu/theses/available/etd-11262001-111453/unrestricted/etd.pdf

Here's a matlab snippet:
% Method 1: Use AT coefficients given by Gerber & Ballato [p101]:
% ref: (Gerber & Ballato, "Precision Frequency Control, Vol. 2", 1985)
% Angle determines frequency tolerance.
Theta_0 = 35.25; %ref angle that produces a zero TC slope at
 %the ref temp
del_Theta = 1.7/60;  %# of minutes off ref angle to produce desired
 %curve
A1 = -5.08e-6 * del_Theta;
A2 = -0.45e-9;
A3 = 108.6e-12
curve1 = A1*(T-T0)+A2*(T-T0).^2+A3*(T-T0).^3;
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] looking for SMT oscillator SC cut, with no oven

[Jim Lux]

On 8/27/15 4:46 PM, Bob Camp wrote:

Hi


On Aug 27, 2015, at 3:58 AM, Hal Murray  wrote:


kb...@n1k.org said:

Is there anything fundamental about SC that forces the turn over

temperature

to be high?



Simple answer yes. More complicated answer : that depends.



The crystal curve on an AT or an IT centers roughly at room temperature.
When you fiddle the  angles to get a stress compensated blank, that center
point moves up to the 90 to 100 C range.


Thanks.  I guess I thought there was an extra degree of freedom so you could
pick the turn over temperature.


Life would be so much simpler if that was true ….

There are indeed a range of cuts you could make. Working out the in’s and outs
of any one of them is a megabuck sort of endeavor. You can predict that this or 
that
will happen. That only gets you just so far. There are a lot of fine details 
that
you can only find out by experiment.



The graph at the bottom of this URL
  http://www.4timing.com/techcrystal.htm
shows that there are actually 3 turn over temperatures.


The Beckman graph at the bottom of that page shows a number of curves that
have no turnover (those below 0 angle) . For the ones that do have a turnover, 
each
one has an upper turn and a lower turn. The magic point in the middle that they 
all
go through is generally called the inflection temperature.

Lots of make your head hurt info at:

http://www.ieee-uffc.org/frequency-control/learning/fc_conqtz2.html

I don’t see anything on a quick Google search that actually give
the Beckman constants.




there's some C code out there that models AT (and also other cuts).. 
I'll see if I can find it. I found it in a PhD dissertation on designing 
temperature compensation neworks, as I recall.


It's not necessarily reality, but it's a model that you could probably 
use to predict a range of variations.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] looking for SMT oscillator SC cut, with no oven

[Jim Lux]

On 8/26/15 4:46 PM, Alex Pummer wrote:

But if he needs 100dBc at 10Hz that is Wenzel's stronghold
[https://twitter.com/ultralownoise]
look that:  http://www.wenzel.com/wp-content/parts/501-04517.pdf



Yep.. got one of those sitting on my desk (or one that's very similar).. 
but it's a 2x2" block that draws many watts..


I want something that is 0.5x0.5" and draws 100-200 milliwatts or so.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] looking for SMT oscillator SC cut, with no oven

[Jim Lux]

On 8/26/15 2:38 PM, Gerhard Hoffmann wrote:

Am 26.08.2015 um 22:04 schrieb Javier Herrero:


I suppose that one of the alternatives that you've explored are the
ABLNO from Abracon http://www.abracon.com/Precisiontiming/ABLNO.pdf


looks just like this one from Crystek:

< http://www.digikey.de/product-search/de?keywords=cvhd-950 >


Yep, that's one (and similar ones) that's my current "best I found in an 
hour of googling"




but that fails the specs, also. If you have a good quality 5 or 10 MHz
source in your
system, you can lock the VCXO to it and clean it up close to the carrier.


Nope.. this oscillator is the ADC clock: it's a direct sampling receiver 
to look at narrow band signals in the 3-30 MHz range.

Maybe Axtal has something.

I'll look..





They say that they are 3rd overtone, but it seems more an AT-cut than
a SC, and anyway is around 10dB poorer


SC requires high temperature, that does not go together well with SMD
and low power.


SC only requires high temperature if you want to operate close to the 
turnover to minimize temperature effects.


I've got a GPS 1pps to count my oscillator, so the sampled data can be 
post processed to take out the frequency variations.  You'd get a bunch 
of digital samples and the timestamps when the 1pps occurs.



I'm kind of hoping someone has run across a SMT OCXO where there's a 
separate oscillator and oven power pin.

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] looking for SMT oscillator SC cut, with no oven

[Jim Lux]

On 8/26/15 1:28 PM, steve heidmann via time-nuts wrote:

Rakon has always impressed me .



I'll take a look. The online datasheets don't have phase noise data for 
close in frequencies (at least the 3 I looked at).. some give a 
"integrated jitter" but it's for 12kHz and out, and I've noticed there's 
lots of "ultra low noise, low jitter" oscillators out there that have 
very good noise performance from a few kHz out, but are pretty bad at 10 
and 100 Hz.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] looking for SMT oscillator SC cut, with no oven

[Jim Lux]

On 8/26/15 1:04 PM, Javier Herrero wrote:

Hello, Jim,

I suppose that one of the alternatives that you've explored are the
ABLNO from Abracon http://www.abracon.com/Precisiontiming/ABLNO.pdf

They say that they are 3rd overtone, but it seems more an AT-cut than a
SC, and anyway is around 10dB poorer than your requirements. An I
suppose that to make surgery in an AOCJY (that fully meets your
requirement) to remove the oven will not be adequate :) Also it is a bit
bulky...



exactly.. I've thought about delidding a OCXO and cutting the trace.
That's a fairly expensive operation, maybe? (by the time we find a tech 
to do it, write the procedure, etc.)  It would turn a $50 oscillator 
into a several thousand dollar oscillator.  Still cheaper than designing 
a new oscillator from scratch.


Or if someone knows of an OCXO where the oven power is separate from the 
oscillator power, that would make it easy.


Darn these highly integrated parts..

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] looking for SMT oscillator SC cut, with no oven

[Jim Lux]

For a project at work, I'm looking for a good close in phase noise 
oscillator (better than -100dBc@ 10Hz, -120dBc would be nice) at 100 MHz 
in a SMT form factor.  But it doesn't need good temperature stability. 
There's tons of SMT OCXOs out there with reasonably good performance, 
but they draw "watts".  My application is actually quite temperature 
stable already AND I have an external reference to measure against.


Most of the lower powered oscillator modules are TCXO, and have, maybe, 
-80dBc at 10MHz.


I guess we could go to a discrete design with a crystal and amplifier, 
but a little clock module would be a simpler solution.






___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPS outage? 19:00-2100 UTC Tue

[Jim Lux]

On 8/19/15 10:12 PM, Hal Murray wrote:


t...@radio.sent.com said:

I think that Menlo Park is somewhat under 300 NM (nautical miles) from China
Lake (depending on exactly where the test was located), and the expected
interference range was about 252 NM. So you might have been at the edge of
the affected area. See: http://www.navcen.uscg.gov/pdf/gps/gpsnotices/
GPS_Interference.pdf




On the other hand, there is a 4000 m tall mountain range (Sierra Nevada) 
and a lot of other stuff between the two. I'm pretty sure that the 
radius of the circle is for *airborne* platforms, where the jammers are 
in line of sight.  (radar horizon at 30,000 ft is about 250 miles)


It would be truly amazing to get L-band propagation from China Lake to 
Menlo Park.


I'd go for the "trucker with a jammer" scenario
http://gpsworld.com/personal-privacy-jammers-12837/

or "malfunctioning equipment oscillating in band" scenario (like the 
Monterey Bay event)


http://gpsworld.com/the-hunt-rfi/



Thanks.  My measurements on Google maps are are close to 250 miles.  That's
to the town rather than the boonies out back.

My antennas are all in poor locations (indoors) rather than good locations so
I'm probably more sensitive than some of their target users.

What's the BNM column on that chart?

What sort of gear would I need to detect a local jammer?




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Cavity frequency air filled vs vacuum?

[Jim Lux]

On 8/18/15 10:30 PM, Bruce Griffiths wrote:

The ratio of the resonant frequency of the evacuated cavity to that
of the air filled cavity increases by the square root of the relative
permittivity of the ambient air or around 300ppm or so. Bruce






I believe that there are systems that measure humidity by this means. 
Water content has a strong effect on permittivity.  Although, since 
there's just not that much water in humid air, it's probably in the few 
ppm range, but easily measureable

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Serial Ballpoint issue again

[Jim Lux]

On 8/11/15 5:20 AM, David C. Partridge wrote:

I just installed Windows 10 (yes I know how rash), and now my Thunderbolt is
being detected as a Microsoft Serial Ballpoint Mouse (yes, just like before
inder Windows 7).  I had set something up on Windows 7 in the boot.ini to
stop this, but for the life on me I can't remember what it was.

Please could someone who's got this setup on Windows 7 remind me what the
magic incantionation is.



On win 7, I go into device manager (rt click on computer, properties, 
etc.) and disable "Microsoft Serial Mouse".


There's probably a command line way to do it it too, using the command 
line device manager tool.  You'd need to know the name of the device, 
but once you know it, you're set for ever.


Does Win10 use a similar scheme?  It's very likely: it's not like MS 
rewrites the OS kernel.. the version changes are more UI focused. Device 
manager in Win7 isn't a whole lot different than it was in NT4.0




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] wtd: WWVB info

[Jim Lux]

On 8/9/15 7:57 PM, John Allen wrote:

Hi Jim -

You wrote:
At some point, multiproject wafers (like MOSIS) might become a hobby
product.  So far, it's in the "several kilobuck" minimum purchase, and,
as well, the tools aren't easy to come by. Or, more properly, good
design tools are expensive, tedious design tools are free.. you CAN
layout an IC for MOSIS with a ruler and and a quadrille pad.. get your
copy of Carver and Mead and have at it)

I assume that you mean Carver Mead and Lynn Conway, Introduction to VLSI System 
Design 1978.




Yup, Mead and Conway.. I misspoke/mistyped..
that's the one.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] wtd: WWVB info

[Jim Lux]

On 8/9/15 4:33 PM, Bob Camp wrote:

Hi

If you never have tried to keep an IC in production, there are some basic things
that may not be very obvious:




There's always Rochester Electronics.. "leaders in the trailing edge" 
(no kidding, that's their slogan)..


They buy old fabs, masks, etc, and keep producing small runs of older 
parts.  For a price.


At some point, multiproject wafers (like MOSIS) might become a hobby 
product.  So far, it's in the "several kilobuck" minimum purchase, and, 
as well, the tools aren't easy to come by. Or, more properly, good 
design tools are expensive, tedious design tools are free.. you CAN 
layout an IC for MOSIS with a ruler and and a quadrille pad.. get your 
copy of Carver and Mead and have at it)


For digital stuff, small boards with FPGAs or microcontrollers on them 
are probably the sweet spot for small runs.


The same is not true for analog.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] wtd: WWVB info

[Jim Lux]

On 8/7/15 1:40 AM, Hal Murray wrote:


kb...@n1k.org said:

Well, at least *some* of the chips out there do not make it to 96 KHz when
sampling at 192 KHz. It’s  been a few years since I dug into them. Back then
a chip that had an internal filter that went to 96K was very much the
exception rather than the rule. If the only point of 192K is getting to a
96K bandwidth, a lot of the chip guys missed out on it ….


Where did 192 KHz come from?  Why is anybody interested in anything that far
over 2*44 KHz?


There's lots of high resolution parts at that rate..
And if it's that fast, odds are the built in sample/track/hold is good 
enough that you could directly sample the 60kHz without much trouble.


With a slower ADC and a good analog BPF and a good sample/hold, you 
could sample at a few kHz, but that shifts the design burden to the BPF 
and the sample/hold.




It's common to have an audio ADC run much faster than Nyquist, but that's a
hack to make it easier to build the cut off filter.  You build a simple
analog filter and a sharp digital filter and decimator so the output is 2x
the target frequency.  You get what you want without a fancy analog filter.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] wtd: WWVB info

[Jim Lux]

On 8/5/15 8:03 PM, Donald wrote:

On 8/5/2015 6:44 PM, Jim Lux wrote:


I'm not sure it would buy you much.. you'd have something running at
240kHz switching the inputs to the detector?

It's MUCH easier to just digitize the 60kHz with a high resolution
converter.  And have a nice BPF in front of the digitizer.

The tayloe/quadrature sampling detector (a key part of the Flexradio
original design) is more convenient if you're making a direct
conversion receiver that needs to tune up to 10s of MHz, since it
allows you to use a slow ADC with lots of bits.  It's basically a I/Q
mixer and will take a lot more parts in total than just getting a
192ksps 24 bit converter.


Thanks for the numbers.

I was uncertain how to scale the sample rate for the I/Q.



if you want to turn a regular single channel data stream and turn it 
into IQ, an easy way is to do the ++-- approach.


complex sample 1 = real sample 1 + j real sample 2
complex sample 2 = -real sample 3 - j real sample 4
complex sample 3 = real sample 5 + j real sample 6
complex sample 4 = -real sample 7 - j real sample 8
and so forth..

This will give you an I/Q sample stream centered at samplingrate/4.  So 
for 192 ksps (real), fs/4 = 48 kHz, so your 60 kHz signal will be at an 
apparent frequency of +12 kHz (sampled at 96k i/q samples/sec)


As mentioned in a different post, I'd look at a filter/decimate, which 
might be computationally faster.  Say you need 1kHz bandwidth (which I 
think is pretty wide for WWVB)

http://www.nist.gov/pml/div688/grp40/upload/Bin-2719.pdf


so you're looking at 192:1 decimation.  with 192 taps, I'll bet you can 
make a fairly decent BPF (and if you're clever, probably bring 60kHz 
down to close to DC at the same time, by combining a synthetic LO with 
the taps)


So you'd be looking at 192 multiplies and adds per 1kHz output point, 
which is a pretty "loafing along" CPU load (e.g. a whole lot of 
microcontrollers could do it)


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] wtd: WWVB info

[Jim Lux]

On 8/5/15 8:27 PM, Donald wrote:

On 8/5/2015 7:55 PM, Bob Camp wrote:

Hi

Analog Devices has some very nice ADC’s that are directly targeted at
doing this general sort of thing. They do not have any “odd” filtering
approach
that creates issues. Some of the early 192 KHz audio parts did not do
very well
past 1/4 the clock rate.

What type of FFT horse power would be needed at 192 Khz ?



not very much..

It's a fairly narrow band signal, It might be more computationally 
efficient to do a Bandpass FIR and decimate.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] wtd: WWVB info

[Jim Lux]

On 8/5/15 12:41 PM, Donald wrote:

On 8/4/2015 9:36 PM, Hal Murray wrote:

kb...@n1k.org said:

So far there have not been any home brew design radios show up that will
demodulate and lock to the new data format. There is plenty of info
on the
transmit format. The demodulation approach is not crazy hard. That said,
there’s still a lot of work to get a receiver running.

Has anybody looked into a software approach?  What sort of front end
would
you want?


I have been looking at building a "Tayloe Detector".

I have not seen any sites that have shown a 60Khz receiver based using a
Tayloe Detector tho.


I'm not sure it would buy you much.. you'd have something running at 
240kHz switching the inputs to the detector?


It's MUCH easier to just digitize the 60kHz with a high resolution 
converter.  And have a nice BPF in front of the digitizer.


The tayloe/quadrature sampling detector (a key part of the Flexradio 
original design) is more convenient if you're making a direct conversion 
receiver that needs to tune up to 10s of MHz, since it allows you to use 
a slow ADC with lots of bits.  It's basically a I/Q mixer and will take 
a lot more parts in total than just getting a 192ksps 24 bit converter.




If anyone has experience with this type of direct conversion receivers,
please share any experiences.

don
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 5 x 2 x 3 = 30 MHz

[Jim Lux]

The whole USO is about $1M, the vast majority of which is labor.  

Analysis, testing, paperwork, etc

Jim






 Original message 
From: Bob Camp  
Date: 07/18/2015  02:10  (GMT+00:00) 
To: tim...@timeok.it, Discussion of precise time and frequency measurement 
 
Subject: Re: [time-nuts] 5 x 2 x 3 = 30 MHz 

Hi

The basic space grade crystal is in the $400 to $1000 range (depends on 
quantity). After they
do the 10:1 sort, the resulting ones are (effectively) $4K to $10K each. The 
cost just goes 
up from there.If you count in the labor, it probably adds another 30% to the 
cost. 

If you don’t do the sorts, the basic crystal is not (on average) much better 
than any other 5MHz
unit with the same blank size. Yield through the whole process is dependent on 
the batch they get.
It *may* be << 10%. 

Bob

> On Jul 17, 2015, at 5:50 AM, tim...@timeok.it wrote:
> 
> 
> I an curious to know the price of this crystal.
> 
> Luciano
> 
> 
> 
> On Fri 17/07/15 00:02 , KA2WEU--- via time-nuts  wrote:
> 
>> I am working with this .. amazing device, Ulrich N1UL
>> 
>> ___
>> time-nuts mailing list -- time-nuts@febo.com
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts [1]
>> and follow the instructions there.
>> 
>> 
>> 
>> Links:
>> --
>> [1]
>> http://webmail.timeok.it/parse.php?redirect=https://www.febo.com/cgi-bin/ma
>> ilman/listinfo/time-nuts
>> 
> Message sent via Atmail Open - http://atmail.org/
> ___
> time-nuts mailing list -- time-nuts@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] OCXO with ADEV of 1E-13 on Pluto Mission

[Jim Lux]

On 7/16/15 8:17 AM, John Stuart wrote:

Here is an interesting link to the New Horizons Mission to Pluto radio
system design.
Note last section describes an OCXO with ADEV = 1E-13 at 1s, and aging rate
of <1E-11 per day.



That's no ordinary OCXO. That's a USO made at APL.  The crystal is in a 
special low stress holder, in a vacuum bottle with a very good 
temperature controller, etc.



http://www.uhf-satcom.com/amateurdsn/Paper-969.pdf



I wonder if their spares will show up on eBay?

Nope, they get repurposed onto subsequent spacecraft.

GRAIL spares are being used in GRACE follow on, etc.

They are sort of the ultimate in crystal oscillators (at least 
domestically produced in the US).



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Number of GPS sats in the sky?

[Jim Lux]

On 7/10/15 6:23 AM, Azelio Boriani wrote:

From 

Block IIA: 3 operational
Block IIR and IIR(M): 19 operational
Block IIF: 9 operational
so they should be 31 satellites working.



Plus various and sundry WAAS and similar signals?


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] beaglebones, time, web services

[Jim Lux]

On 7/7/15 9:59 PM, Brian Inglis wrote:

On 2015-07-04 07:13, Jim Lux wrote:

I've got a project I'm working on to make a sophisticated sundial with
moving mirrors.  I've got a batch of Arduinos that move the mirrors to
the appropriate places, given the current sun angle, etc.
I've got a beaglebone that runs some python code to calculate sun
angle based on time
The beaglebone will have a GPS feeding it to get time.
BUT now, I'd like to add a web interface, so that it can be
manipulated by a mobile device using a browser.
One way I can think of is to run some sort of limited web server.
there are a couple that come with the beaglebone, including the python
"simplehttpserver".
But I'm sort of stuck on the interface between the webserver and the
other code running.
I've done this kind of thing where the one task goes out and updates
files in the tree that's being served by the web server, and that
works fine for "status display" kinds of things that don't update very
quickly. It's also nicely partitioned.
but I want to be able to change the behavior of the system (e.g. by
having the server respond to a PUT or something)
Is the best scheme to go in and modify the webserver code to look for
specific URLs requested, and then fire off some custom code to do what
I want?


May want to start with a control web page with an HTML FORM element and
embedded input elements - easy even if you have not done much form
design and entry implementation.
Submit target can be any URL designating a Python CGI script, which
generates at least a Content-type header and HTML on stdout returned to
the browser.
HTML output normally includes a copy of the original FORM (with values
passed selected for editing) as well as HTML output and maybe inline or
linked graphics.
You only need a web server that supports the CGI interface, with some
way to configure it and say where the scripts are.
See Python cgi, html, http module docs to DIY.




Yes, that seems to be the way..
The interesting thing is that the cgi needs to return reasonably fast, 
or the user client will timeout, so it's not a good way to do something 
that takes a long time.  Great for "put parameters in a file" or "send 
short command out IO device", not so great for "start long running 
process that needs to continue after user has gone on to do other things.


So it comes down to lashing up some sort of interprocess communication, 
whether it's a named pipe, a file that is shared between two processes, 
IP sockets, etc.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] US export regulations for TICs

[Jim Lux]

On 7/7/15 6:28 AM, Attila Kinali wrote:

Moin,

As we need a need a proper TIC here to do our research, we are
going to buy one from ebay form a seller in the US and let a friend
who is in the US at the approriate time and can pick it up to bring
it back in the plane.

Now the big question is, are there any export regulations regarding
such equipment and if yes, where do I find it? (my search didnt show
up anything approriate). Yes I know it's a boat anchor and that takeing
it in a plane is kind of iffy, but it's better than shipping it.



Ah, the complexities of Export Controls, with which I deal every other day.

I'll summarize a bit here, for the benefit of other list members who 
might contemplate this.  If it's straying too far off the list topics, 
let me know...


Ebay sellers in general don't know anything about export controls, other 
than some just don't sell to foreigners (which they think neatly solves 
that problem it doesn't).


In export control, there are US Persons and non-US persons.  The former 
are US Citizens and legal permanent residents (green card holders), who 
are NOT representatives of a foreign entity.  You can be a US citizen, 
but work for, say, Thales-Alenia Space Italia, and be a non-US person.


If you are transferring export controlled goods (or information!) to a 
non-US person, you *may* need an export license, depending on what the 
goods are.  Where the transfer takes place is immaterial (so if your 
friend going to do the pickup isn't a US person, the export would occur 
when he or she put their hands on the goods).


Now to whether you need a license.  There are two kinds of export 
controls in the US: US Munitions List (ITAR), run by the department of 
state, and Commerce Control Regulations (CCR) run by the Department of 
Commerce.


USML stuff almost always needs a license, a non-trivial process 
requiring, typically, an "end user certificate" describing who the 
ultimate recipient is (e.g. you can't say "I'm buying it for myself", 
when, really, you're transferring that 17-axis milling machine to a 
"designated country")


You can easily google the USML (the version at fas.org is a bit out of 
date): it's fairly straightforward, and MOST of the stuff will say 
"specially designed for military purposes", so a piece of test equipment 
that has multiple uses probably won't be controlled, unless it's 
designed to be bolted into a fighter or ship or carried by an infantry 
soldier.


There are some "no conceiveable dual use" things on the USML: GPS 
receivers that work at more than some high altitude or at Mach 5 are a 
nice example.


For USML/ITAR stuff ALL exports need a license.

Then, there's the EAR/Commerce rules.  For these, which are much broader 
and cover more commodities and things, there's a "which country is it 
going to" distinction.  Shipping fast CPUs to the UK, no problem; 
Shipping to North Korea, can't do it.



USML is here https://www.pmddtc.state.gov/regulations_laws/itar.html

I would suggest looking at Category 11 (Military Electronics)
for which there was a recent amendment:
https://www.pmddtc.state.gov/FR/2014/79FR37536.pdf


the Export Administration Regulations (EAR) are here
https://www.bis.doc.gov/index.php/regulations/export-administration-regulations-ear

Chapter 3 might be of interest. It has rules on RF amplifiers, arbitrary 
waveform generator, signal analyzers, etc.  I've never looked for 
counters or TICs.






Thanks in advance
Attila Kinali



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] beaglebones, time, web services

[Jim Lux]

On 7/6/15 3:19 PM, Tom Harris wrote:

Since you want simple just use a CGI script written in your language of
choice. Very easy technology to learn, Python has support libraries out of
the box if you want. You have a webpge with carious simple controls on it
like buttons etc, you click a special button that posts a request to a URL,
the webserver runs a script that generates the response, the webserver
serves it out, your browser displays it. Why bother with learning a
framework? Messing about with mechanics is far more fun!






The only hiccup with the cgi approach (and with "directly code the 
action in the guts of the server" like with flask) is that the 
subprocess that's spawned has to complete before control returns (e.g. 
to serve stdout to the user). So if you want to fire off a task that 
will run in parallel with the webserver's other stuff, you need to have 
some sort of interprocess communication (e.g. a named pipe, socket, 
file, MPI communicator, etc.).  (or you do something like run "at" or 
"batch", which is basically using a file as a interprocess 
communication, and the at daemon watches the file)




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] beaglebones, time, web services

[Jim Lux]

On 7/5/15 8:43 AM, Bob Camp wrote:

Hi


On Jul 5, 2015, at 8:46 AM, Jim Lux  wrote:

On 7/4/15 7:53 PM, Hal Murray wrote:


jim...@earthlink.net said:

Exactly... I've got an array of mirrors on az/el mounts (two servos
stacked) and the reflection from the mirrors on the wall forms the display.


How many pixels in that display?  Or what is the unit of quality measurement?

What sort of ADEV are you aiming for?  If your goal is solar time rather than
TAI or UTC, you should be able to get pretty good.




Prototype is 6 pixels to demonstrate concept and work out the bugs. Long term, 
probably several dozen.

Time Accuracy? better than a second

Turns out, having done some experimenting, the real issue is angular accuracy. 
RC servos aren't all that great, and have significant jitter (probably not an 
issue in their design application which tends to have good mechanical low pass 
filtering).  They're cheap and easy to use (as in, I had a bunch in the garage 
I could cannibalize out of another project).

But if you have 3x3 inch mirrors (call it 7.5 cm), and want to create a picture 
on the wall that's, say, 10 meters away, you really need angular pointing of 
0.007 radians.. that's about 1/2 degree.  An RC servo has roughly 270 degree 
rotation corresponding to 256 steps of PWM (in the Arduino implementation).


Probably a good place to use the “drive a stepper as a selsyn trick. Steppers 
are dirt cheap these days and you can either program the drive yourself or get 
chips that will do it for you. You have essentially zero load and zero 
acceleration. There is no need for anything big.


Indeed, microstepping might be the way to go in a production system.

But steppers don't have convenient mechanical mounting stuff like RC 
servos do. I could assemble my prototype with zip ties, double sided 
foam tape and a few screws.   For a stepper scheme I'd need to design 
and build (e.g. fabricate) bracketry.  It's also more complex than just 
plugging a servo into a pin on the Arduino; that's pretty easy.


And then you also get into the "do you really want to use an arduino, 
why not program a X microcontroller  on a custom board you've designed 
for the purpose with all the driver components, etc."


If I were building up a full scale system, that's probably what I'd do. 
 BUT, in the mean time, my 6 RC servo az/el thingys are good to fool 
with and get a feel for various configurations and what the design 
issues on a larger system would be.



The virtue of the BBB and Arduino scheme is mostly that it can be 
cobbled together without much work. And you can leverage large consumer 
equipment volumes for the actuators, servos are <$10 each in any sort of 
quantity; it would be hard to find a packaged motor/gear train with a 
feedback pot for that much (leaving aside surplus).


I used to have a box of small 200 step/rev motors (floppy drive 
positioners), but they had a weird sized shaft, so we're back to the 
fabrication of mounts: the servo has a nice splined nylon shaft that 
mates with cheap other injection molded stuff.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] beaglebones, time, web services

[Jim Lux]

On 7/4/15 7:53 PM, Hal Murray wrote:


jim...@earthlink.net said:

Exactly... I've got an array of mirrors on az/el mounts (two servos
stacked) and the reflection from the mirrors on the wall forms the display.


How many pixels in that display?  Or what is the unit of quality measurement?

What sort of ADEV are you aiming for?  If your goal is solar time rather than
TAI or UTC, you should be able to get pretty good.




Prototype is 6 pixels to demonstrate concept and work out the bugs. Long 
term, probably several dozen.


Time Accuracy? better than a second

Turns out, having done some experimenting, the real issue is angular 
accuracy. RC servos aren't all that great, and have significant jitter 
(probably not an issue in their design application which tends to have 
good mechanical low pass filtering).  They're cheap and easy to use (as 
in, I had a bunch in the garage I could cannibalize out of another project).


But if you have 3x3 inch mirrors (call it 7.5 cm), and want to create a 
picture on the wall that's, say, 10 meters away, you really need angular 
pointing of 0.007 radians.. that's about 1/2 degree.  An RC servo has 
roughly 270 degree rotation corresponding to 256 steps of PWM (in the 
Arduino implementation).

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPS/UTC time

[Jim Lux]

On 7/4/15 10:25 PM, Brian Inglis wrote:


Thanks for the good and interesting refs.
One of the interesting points was that normal variations are multiples
of those caused by earthquakes, and annual variations are up to 1ms and 1m.
Another was that the jet streams produce large short term variations
caused by temperature differences.
Will this always turn up as an issue with all oscillators? ;^>

There's a fair amount of noise (some looks periodic) in that plot from 
USNO of day length


So, what *is* the ADEV of the earth's rotation?

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPS/UTC time

[Jim Lux]

On 7/4/15 2:01 PM, Tom Van Baak wrote:

(the Japan earthquake in 2011 sped the earth up by 1.8
microseconds/day. The Sumatra quake on 26 Dec 2004 had a bigger
effect: 6.8 microseconds)


Hi Jim,

Just in case you didn't know -- these are theoretical results only.
There's a guy at JPL (Richard Gross) who does the calculations and
any time there's a big seismic event he runs the simulations and out
comes a number. That's pretty cool but the numbers so far are always
smaller that what VLBI can actually measure. Still, it makes a nice
press release and physics lesson.


Yes, that's true.

I wonder, though, if over time, they can measure it: they're collecting 
a lot of GPS data from around the world, and the GPS constellation is 
reasonably fixed in inertial space.


Although, if the earth rotation changes, and earth's mass distribution 
isn't perfectly symmetric, then that will change the orbits of the 
satellites.


I could envision that they are building a very complex model of GPS 
orbits and earth underneath it, and they could somehow see the transient 
in the model parameters.  (as the papers you cited show)



The measurements are always getting better.  There's also VLBI of 
stellar sources, and that's getting better too, as they model out all 
the perturbations.



One can hope..

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPS/UTC time

[Jim Lux]

On 7/4/15 12:18 PM, Brian Inglis wrote:


(the Japan earthquake in 2011 sped the earth up by 1.8
microseconds/day.  The Sumatra quake on 26 Dec 2004 had a bigger
effect: 6.8 microseconds)


By my calculations, that should mean the earth rotated 3mm farther/day
at the equator after Sumatra.
Anyone know if, or how much, these perturbations affect the orbit of the
earth or other planets, or where to find out?




The folks at JPL who do these calculations could probably tell you..

http://scienceandtechnology.jpl.nasa.gov/people/r_gross/ is probably 
your man.


http://www.iers.org/IERS/EN/Organization/ITRSCombinationCentres/JPL/jpl.html


Off hand, the center of mass of the earth shifted slightly (in addition 
to the rotation rate change, there was a axis of rotation change), so 
there would be a gravitational effect on other heavenly bodies.


Whether it is measurable is another thing.  That's more of a celestial 
mechanics thing.  A different group at JPL who do that, but if you email 
Richard Gross, he might tell you who does that.



Jim
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] beaglebones, time, web services

[Jim Lux]

On 7/4/15 1:42 PM, Simon Marsh wrote:




Pretty much every webserver ever written allows you to run a script in
response to a request. Nowadays there are frameworks that integrate
closely with the language of your choice and do all the heavy lifting
for you.

If fact, the problem is really too much choice, here's a list of
frameworks from the Python wiki:
https://wiki.python.org/moin/WebFrameworks


Yep.. that's exactly the problem... So I was consulting the hivemind 
here... we tend to be building little widgets that are more than a 
blinky light, but also aren't serving airline reservation systems.






If you want lots of functionality then head for the top of the list, but
these are overkill for what you are trying to do.

Scroll down to the 'Non Full Stack Frameworks' and pick one that makes
sense to you. These should all allow you to route a URL to some Python
code, and the process should be simple enough that if you spend more
than 15 mins to get an example up and running then just ditch it and
move on to the next one.


That's where I am...

flask does reasonably good..
I haven't tried firing off a second thread yet..






One caveat, if you are planning to put this on the public internet then
it's a very good idea to proxy the service behind a 'full-fat' webserver
(e.g. apache) that can safely manage access, load, security etc. I
wouldn't expose a BBB directly to the Internet, especially one that is
controlling expensive physical things.



Nope.. just local access from *my* phone on *my* network

And the hardware isn't breakable(!)  at least not by anything that any 
of the processors can do.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] beaglebones, time, web services

[Jim Lux]

On 7/4/15 12:31 PM, Bob Camp wrote:

Hi

As silly as it sounds, having a separate board for the user i/o is probably the 
best way to go.
You already have an empire of devices that (somehow) chat with each other. The 
barrier of
“it’s all on one device” has been broken even before i/o has been added.


Well, sure.. that's what the BBB is for.. mostly the UI.  but at some 
point, the web server has to tell someone else what to do.  You can fire 
off another process, send a message to another processor (or thread), or 
whatever.


And I was looking for a simple(!) webserver that supports this level of 
sophistication.  There's plenty of very lightweight examples out there 
(that run on Arduinos for instance) that are basically single threaded.. 
you intercept the "GET /myfunction" (or whatever) and that turns into a 
"call abc(parameters)"... and while "abc()" is running, the webserver 
isn't.  That's fine for "setting parameters", but not good if the abc 
process is going to take minutes.







Once you get past that part of it, it’s all a bunch of “that depends” and 
personal preference. There is very little
right and wrong. For very little money, you can go from a single core to a quad 
core device on your i/o
processor. The same is true of RAM and flash. If this is a one up (or few 
dozen) sort of thing, optimizing the
board probably makes less sense than attacking the (inevitable) multitude of 
Ardunio gear controlling the
rest of it.

=

Assuming that we’re not already way off track - I’d use a “real” web server to 
feed the user. You get the full
range of modules that way. You can handle anything you decide you need as the 
feature list expands. I’d back
it up with Python, just because it seems to work fine and I already have worked 
my way up the learning curve. Others
would (I’m sure) recommend languages that they are more familiar with. They all 
will get you to the same end
result. If you want to be cool, there’s always Node.js …



Yeah, that's where I'm heading.. but I was looking for something between 
"single threaded webserver with direct calls" and "install apache"



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] beaglebones, time, web services

[Jim Lux]

On 7/4/15 11:55 AM, Mike Magin wrote:

Somewhat new time-nut here (had one of the Samsung-branded Z3805s for a
few months as a house 10mhz ref, but it really got out of control when
I acquired a Wavecrest DTS, multiple frequency counters, an old Astron
1250a, a Lucent RFTG-u pair, etc.), thought I should finally de-lurk
since I can perhaps offer some useful opinion on this.  Comments inline.

On Sat, Jul 04, 2015 at 06:13:06AM -0700, Jim Lux wrote:

I've got a project I'm working on to make a sophisticated sundial
with moving mirrors.  I've got a batch of Arduinos that move the
mirrors to the appropriate places, given the current sun angle, etc.


I guess you are making a human-readable sundial,


Exactly... I've got an array of mirrors on az/el mounts (two servos 
stacked) and the reflection from the mirrors on the wall forms the display.






I was thinking recently

about building a computerized mean-solar-noon tracker, just to see what
sort of accuracy I can get.  Haven't decided between a cheap fisheye
camera behind a dark filter (welding lens?) versus a sort of slot-lens
(like a pinhole, but to tolerate the seasonal change in north-south
elevation) with a wide-range light sensor (CdS or a modern ambient light
sensor IC).


How accurate do you need to be..

two/four solar cells with a hole that projects an image of the sun on 
the cells, and compare the outputs (classic sun sensor for a satellite)


spinning mirror and fixed solar cell, timing of pulse tells you where 
the sun is


camera with a wide angle lens, and then do multi-pixel centroiding (0.1 
pixel is easy)


camera sensor with a plate with tiny holes in it, with the spacing of 
the holes slightly different than the pixels, so the "sun spots" have 
slightly different coverages of each pixel.








I've got a beaglebone that runs some python code to calculate sun
angle based on time

[...]

Or are there libraries that make this more cookbook? (the little
"getting started with beaglebone" book talks about flask)


In a previous contract job, I did some work with Flask, it's pretty
nice, especially for the basic case of "make this subset of the URL
space be handled by this function".

I haven't set it up from scratch, but the Flask documentation seems pretty
good, and if you're already familiar with Python, I'd highly recommend it.


Well, I got it to do the beginner "click here and turn on/off the LED" 
thing..


What I'd like, though, is to separate the "web serving" thread and the 
"doing stuff thread". I suppose I can fire off a thread from the 
webserver python.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] beaglebones, time, web services

[Jim Lux]

On 7/4/15 11:45 AM, Bill Dailey wrote:

Pysolar

Sent from mobile



"Pysolar: staring directly at the sun since 2007"


excellent.. thanks..

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] beaglebones, time, web services

[Jim Lux]

I've got a project I'm working on to make a sophisticated sundial with 
moving mirrors.  I've got a batch of Arduinos that move the mirrors to 
the appropriate places, given the current sun angle, etc.


I've got a beaglebone that runs some python code to calculate sun angle 
based on time


The beaglebone will have a GPS feeding it to get time.

BUT now, I'd like to add a web interface, so that it can be manipulated 
by a mobile device using a browser.


One way I can think of is to run some sort of limited web server. there 
are a couple that come with the beaglebone, including the python 
"simplehttpserver".


But I'm sort of stuck on the interface between the webserver and the 
other code running.


I've done this kind of thing where the one task goes out and updates 
files in the tree that's being served by the web server, and that works 
fine for "status display" kinds of things that don't update very 
quickly. It's also nicely partitioned.


but I want to be able to change the behavior of the system (e.g. by 
having the server respond to a PUT or something)


Is the best scheme to go in and modify the webserver code to look for 
specific URLs requested, and then fire off some custom code to do what I 
want?


I'm not particularly interested in javascript, and would prefer python.


Or are there libraries that make this more cookbook? (the little 
"getting started with beaglebone" book talks about flask)


There's quite a few websites out there where someone has done some sort 
of "home automation", but they tend to be a bit light on the analysis of 
pros and cons of implementation architectures: "I built X using Y and Z 
and it sort of works".



Actually, along a similar line.. my "solar position" code isn't very 
pretty (it's sort of replicating some code I wrote in Basic a long time 
ago, with some changes from stuff I cribbed from ccmatlab).  If someone 
knows of a python package that just "does this", I'd love to hear about 
it.  Either Az El, or X,Y,Z in ECI or ECF would do.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPS/UTC time

[Jim Lux]

On 7/3/15 9:45 PM, Brek Martin wrote:



Hi Guys,

I feel like I missed “The Big Thing” in time keeping land. I should have 
watched my Ublox LEA-5T.
What is the difference if it is in the reporting mode for GPS or UTC time?
If they skip a second UTC, surely the GPS time isn’t run incorrectly forever.




Who's to say which is "correct", UTC or GPS?

GPS time is derived from TAI; both are monotonically increasing, 
continuous, and constant rate.  These are nice attributes for something 
you're going to use for time stamping, or controlling.  No gaps, no 
jumps, etc.


UTC (and local civil time, and GMT, etc.) have leap seconds, to adjust 
the time scale to the motion of the Earth;  so that the sun is highest 
at noon (after accounting for the equation of time).


While that's somewhat convenient, I doubt anyone would really object to 
noon being a few tens of seconds away from the zenith crossing when 
standing on the line. TAI is ahead of UTC by 36 seconds.  They add a 
leap second every year and a half, so I guess in 100 years, we'll have 
drifted some minute or so away.  (the earth has slowed down in the last 
200 years.. a day is now 86400.0015 seconds long, although it's faster 
now than it was in the 70s, when it was 86400.003 seconds)


http://tycho.usno.navy.mil/leapsec.html

(the Japan earthquake in 2011 sped the earth up by 1.8 microseconds/day. 
 The Sumatra quake on 26 Dec 2004 had a bigger effect: 6.8 microseconds)



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] windows and leap seconds

[Jim Lux]

https://support.microsoft.com/en-us/kb/909614

basically: ignore the leap second, so your computer is "fast" relative 
to UTC, and next time you resynchronize time it adjusts


Windows, these days, runs NTP

http://blogs.msdn.com/b/mthree/archive/2015/01/08/leap-seconds-010815.aspx
http://blogs.msdn.com/b/mthree/archive/2015/01/14/leap-seconds-011415.aspx

for more info
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] magnetic electronic components

[Jim Lux]

On 6/23/15 4:02 PM, Tim Shoppa wrote:

Yes, for ferrites, many (all?) of the Amidon FT-xxx parts are perfectly
standard Fair-Rite cores available from full-line distributors like Mouser,
Newark, etc.

Iron powder cores are not stocked by any of the standard distributors that
I know of, but kitsandparts.com has good prices and quick delivery.


Powdered metal.. typically from micrometals.

Lodestone Pacific is one of their distributors.  I've not bought small 
quantities of Micrometals parts from them, though.


(lots of european distributors)




Sometimes Amidon is the best or only place for a particular part.

Good online source on ferrite transformers with measurements and
distributor part numbers: Clifton Labs. e.g.
http://www.cliftonlaboratories.com/ferrites,_inductors_and_transformers.htm

Tim N3QE


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] magnetic electronic components

[Jim Lux]

On 6/23/15 4:25 AM, Tim Shoppa wrote:

"Experimental Methods in RF Design" has a half-dozen pages specifically on
the choices of powdered iron and ferrite materials, and lots of working
circuits and designs with measurements. Aka EMRFD.
http://www.arrl.org/shop/Experimental-Methods-in-RF-Design

Here in the USA, iron powder and ferrite cores of many different materials,
sizes, and a few shapes are available from Amidon and kitsandparts.com.
Many useful ferrite cores for multi-turn transformers and chokes, are sold
as "EMI beads" by Mouser and Newark and other mainline distributors. I
don't know too much about easy availability in EU.



I don't know that I'd recommend Amidon as a source.  Back when 
mail-order was king, Amidon did hams a real service by buying in bulk 
and selling in small quantities.  The price was high, but there was no 
other source.


Amidon has gone through a lot of business changes over the last 20-30 
years (making magnetic tape heads and then not, overseas manufacturing, 
etc.).


They're not the same company as Bill Amidon sitting in his garage in the 
San Fernando Valley putting cores in little paper or plastic envelopes 
with that folded up tissue paper instruction and data sheet with all the 
handy design equations and graphs.


It used to be tough to get databooks from large manufacturers as a 
hobbyist. The sales reps would hand them out after qualifying you as a 
potential lead. Bill did everyone a great service in essentially 
redrawing and republishing all the needed data in a handy form.



These days, most of the parts, (e.g. made by Fair-rite,Ferroxcube, 
Philips), etc are available from Mouser, DigiKey and other distributors 
readily.


Furthermore, the design information is readily available on the web 
(e.g. from Fair-rite) or in various mailing lists.








___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] magnetic electronic components

[Jim Lux]

On 6/22/15 12:02 PM, Attila Kinali wrote:

Hi,

I was looking up some stuff and realized (again) that I don't know
anything about how magnetic electronic components (inductors/solenoids,
transfomers, baluns, ferrite beads...) work. Yes, I can calculate
the inductance, I know how to get from the AL value to number of
windings. But I don't know anything about the practical issues
or where they come from. Unfortunatelly, this knowledge seems to
generally rare among EEs (at least everyone I asked in the last
couple of years) and books about it are either long out of print
(with no pdf available) or more geared towards the physics student.



the best, and probably the only, book is the one by E.C. Snelling.
http://www.amazon.com/Soft-ferrites-properties-applications-Snelling/dp/0592027902

1969 edition is
https://archive.org/details/SNELLING__SOFT-FERRITES__1969

and it's not like the properties of magnetic fields have changed.



So, does anyone have any recomendation where I could read up
on this? Books, pdfs, webpages,... anything.

Also something that covers more the application side, ie how to
use ferrite beads/toroids to build devices, would be appreciated.

Thanks in advance

Attila Kinali



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] potential source for cheap copy of labview

[Jim Lux]

On 6/21/15 11:28 AM, Don Latham wrote:

Just for fun, went to the site.  $149 for basic, but by the time I added all
the toolboxes I thought (!) I needed, I was over $750. sigh.
Don



Hence the popularity of the student license (or Octave)

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] potential source for cheap copy of labview

[Jim Lux]

On 6/20/15 3:03 PM, Dave Daniel wrote:

I wish the MathWorks would resume that practice. Back in the late 90s
they would sell one licenses for MatLab and SimuLink for an affordable
price if one singed an agreement that restricted one to personal
(specifically, non-commercial) use. My copy from back then is so old
that it won't run on Windows 7.



Mathworks still does a variety of low cost licenses, including a $149 
for Matlab "home" license + $45 for add on products. (not for academic, 
commercial, govt, or organizational use)


They also have a $49/$99 student license "in conjunction with coursework 
at a degree granting institution".  I suppose that you could sign up for 
a class at the local community college.(that's gone up a lot with a 
bunch of added fees around here)


The new matlab has drivers/simulink blocks to handle a lot of hobby type 
hardware platforms (RPi, Arduino, LEGO Mindstorms NXT)



One can also use Octave, which is very, very similar to Matlab (I go 
back and forth between the two all the time).  Octave doesn't 
necessarily have all the nice toolboxes that Matlab has. And, the 
plotting is done differently (which is a significant issue, since a lot 
of what I use matlab and octave for is generating nice looking plots).




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Jim Lux]

On 6/19/15 9:30 PM, Bob Camp wrote:

Hi

One of the most interesting things about the last paper mentioned:

On Jun 19, 2015, at 8:57 PM, Charles Steinmetz  wrote:

Rick wrote:


However, a better tutorial would be the one written by
HP's Dieter Scherer which was published in Microwaves
& RF Magazine (or possibly Microwave Journal).  I believe
the same content was available from HP as an Ap Note or
something.


I suspect the paper you are referring to is:

Design Principles and Test Methods for Low Phase Noise RF and Microwave Sources 
(Scherer, 1978)

It is available on Didier's site at:



There are two other Scherer papers there:

Generation of Low Phase Noise Microwave Signals (Scherer, 1981)


The "Art" of Phase Noise Measurement (Scherer, 1985)



^ this one


Best regards,

Charles



is that it has a phase noise plot on an open loop microwave source down to 10 
Hz. Not quite the VCO Jim
was looking for, but close ….




page 25 shows down to 100 Hz, "Typical Free-Running Source at 10 GHz". 
Eyeballing it it looks like 20dB/decade from 10-100k, a bit more from 
1-10k, and almost 30 dB/decade from 100Hz to 1k. I wonder what the 
source was?


The SAW is definitely in the 30dB/decade bucket


This makes a good case for the "30dB/decade very close in"



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Jim Lux]

On 6/18/15 1:46 PM, Brooke Clarke wrote:

Hi Mark:

Is there any documentation on the HB100?

first hit on google for "HB100 microwave sensor" (recognizing that what 
comes up as *my* first hit will probably be different than *your* first 
hit)...

https://www.openimpulse.com/blog/products-page/product-category/hb100-microwave-sensor-module/

links to some data sheets.

There's people building arduino shields for these things.  They're also 
available with more electronics on the back that does the "motion 
detector" kind of function with a single bit output.



I haven't found a $10 unit that has both I/Q yet

The old MA/Com gunnplexers didn't have I/Q either, but they did have a 
varactor for modulating.


Of course, you can also modulate a gunn by just changing the bias 
voltage. Wretchedly non-linear, but that's what a DAC and lookup table 
is for.


Just for fooling around, I'd love something like the HB100, with the 
ability to modulate it over a few hundred MHz (leaving out DROs as an 
option) and with I/Q detection (doesn't have to be perfect quadrature, 
"about 90 +/- 20 degrees" is good enough: the errors are easy to 
calibrate out)


Probably can't do it for $10, though.  More like $100 (in small quantities)




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Jim Lux]

On 6/18/15 10:05 AM, Mark Sims wrote:

If you want to play with a homodyne doppler radar,  search Ebay for "hb100 microwave 
sensor".  It is a cute little 10 GHz doppler module  that costs around $6.   It can 
be operated in continuous or pulsed mode.  The output does require a couple of op-amps to 
get a TTL level output.
___



Those a very cool (and, by the way, easy to blow up with ESD)

The real failing of them is that they don't give you I/Q output so you 
can't distinguish motion towards and away.  If they had added another 
mixer, life would have been really nice.


As a practical matter, if you want to fool with building your own radar 
from scratch, for ranges of up to, say, 10 meters, 1 mW radiated power 
and about 50dB receiver gain works pretty well.  You can get your gain 
either at RF (e.g. with MMIC amps) or post mixer with op amps.  That 
will get you an output of "tenths of a volt"  (the HB100 puts out 
"microvolts")


The HB100 uses a DRO and puts out about 10-15 dBm


For more radar fun, use a VCO so you can step the frequency, and then 
you can do range synthesis (FMCW radar), and if you're really ambitious, 
you can build a SAR.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Jim Lux]

On 6/18/15 3:46 AM, Bob Camp wrote:

Hi

Be careful when you do find the data.

When you go very close in on something like a VCO, you get much higher phase 
noise
than we normally worry about. Some of the “assumptions” that underly the 
measurements
are no longer true. Small angle of modulation is one, but there are a few 
others.


yeah that's a big one.. There's a slide in a recent Keysight/Agilent 
presentation on phase noise measurement that has a line on it at, I 
think, 30dB/decade going through -30dBc@1Hz or something as the 
threshold where small angle approximation is no longer valid.



 A simple

“plug it in and read it” may not give you valid information without some 
corrections.  Coming
up with a supply that’s good enough (at 1 Hz) would be a major challenge ...


Well, it's not like we're looking for 1E-10 ADEV in this sort of 
application.


Mostly, it's a "what should you expect".. and also the whole "we all 
know gunn oscillators are noisy, but how noisy, really"



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Jim Lux]

On 6/17/15 6:15 PM, John Miles wrote:

Also see http://www.ke5fx.com/gunnpll.html , a quick and dirty but
successful attempt at locking a Gunnplexer in a relatively low (1
kHz) loop bandwidth.  The inband noise is likely too high for good
performance in a radar application, but the basic idea is workable
enough.  Unfortunately I tried this experiment before I had any way
to generate log plots. :(

I can't think of many noise-critical applications where a microwave
VCO is used without some form of phase locking.  I'd think that a
homodyne architecture would still need a synthesized source, just
because the waveform being received is delayed relative to the one
that was sent.


Yeah, but at short ranges (<100m) that delay is on the order of <1 
microsecond. So low frequency phase/frequency variations cancel 
themselves out.


Police radar speed detectors are a fine example. Gunn oscillators work 
just fine here when operated open loop.




 Without a clean source, I'd imagine that you'd have

to do autocorrelation between the outgoing and incoming channels
rather than simple/cheap baseband mixing.  An obvious question would
be whether it's cheaper to add another digitizer and correlator to
your pipeline than it would be to clean up your source...


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Jim Lux]

On 6/17/15 1:08 PM, Magnus Danielson wrote:

Jim,

John Miles have been a bit active:
http://www.ke5fx.com/brick/brick.htm

Just to give you a start-sample.



those seemed to be all PLL outputs.. I didn't see the bare VCO data.

And, I'm really interested in the 1 Hz to 100 Hz kind of range.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Jim Lux]

On 6/17/15 1:08 PM, Richard (Rick) Karlquist wrote:



On 6/17/2015 8:22 AM, Jim Lux wrote:

I'm looking for some representative data for inexpensive microwave VCOs
(in the 2.5-6 GHz range, in general).  Not in a locked loop situation,


If the phase noise data you have goes to a low enough frequency to
get below the 1/f corner (which is the case for the example you cited)
then it is a very safe bet that the noise will go up by 30 dB/decade
below that.


That's sort of what I was thinking..  But a couple good measurements is 
worth hours of dialectic.




Having said that, if an ordinary engineer had asked me this question,
I would think that he needed some coaching on how to clean up the
VCO with a synthesizer of sufficiently wide loop bandwidth.


Actually, in a homodyne FMCW radar, there's no loop, so you have just 
the bare noise of the oscillator.  Actually, it looks a lot like a 
"delay line" type phase noise test set, where the length of the delay 
line is some tens of ns (e.g. free space propagation).





 However,

you are very knowledgeable, so I will assume you are going to do
that and just want to predict the phase noise after clean up.  The
trick (as most time nuts know) is to use a small enough capacitor
in the loop filter so that you get clean up at a 40 dB/decade rate
so you can actually make some headway against the 30 dB/decade
1/f slope.

I have been through this exercise innumerable times and also taught
it to many others, and it seems to be very predictable.

In the unlikely event you use the VCO open loop, you'll have lots
of problems with microphonics, power supply noise, and even magnetic
fields from power transformers, as well as load pulling and thermal
drift.  Making microwave oscillators that can be used open loop
(especially inexpensive ones) is definitely a lost art.  It died with
the HP8640 sig gen.


In a homodyne radar, you've got to deal with all those things. 
Actually, magnetic fields aren't nearly as big a problem as the 120 Hz 
signal you get from all the fluorescent lights, which are basically 
radar reflectors that turn on and off every half cycle.


If you're doing a "door opener motion detector" at 10.525 GHz (ISM Part 
15) the doppler is 60 Hz/ m/s (roughly).  Walking speed is 1-2 m/s








Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] Close in phase noise of microwave VCOs

[Jim Lux]

I'm looking for some representative data for inexpensive microwave VCOs 
(in the 2.5-6 GHz range, in general).  Not in a locked loop situation, 
but just bare: with a DC voltage on the tuning input. I'm particularly 
interested in data closer than 100 Hz.


Most of the data sheets (e.g. from Minicircuits, ROS-3710; crystek 
CVCO33 series) show noise from 1 kHz or 10kHz out, because most of these 
parts are intended for use in a PLL, and the "close in" will be 
determined by the loop.


Before I go out and hook one of these up and measure it, I figured I'd 
ask if someone out there has done it, or if there's a data sheet.


I'm not looking for any particular part or frequency or even exact 
numbers: more "representative, typical" kind of performance one might 
get from one of the plethora of $20-50 VCOs out there.


Something like the ROS-3710 looks like it's about -30dB/decade trending 
to 20 dB/decade.

(-70 @ 1 kHz, -96@10kHz, -118@100kHz, -138@1MHz)

A paper I found on 77 GHz sources cite a 30dB/decade (actually they give 
it as f^-3.05).





background:

I've got a homodyne radar at work we use for detecting heartbeats of 
buried earthquake victims. I've also got a variety of gunn oscillator 
"doppler radars" of one sort or another.
There's 10GHz homodyne radars available for $5 from China (the selling 
prices range from $1 to $20, with corresponding inverse costs in 
shipping.. ).  They're designed for intrusion detectors and automatic 
door openers.


There's all kinds of cheap 2.45 GHz sources around: one might be able to 
repurpose an old 802.11b WiFi interface, for instance, although I think 
those are all synthesized PLL designs.


And my car has a 77 GHz radar in it for adaptive cruise 
control/automatic braking.


There's also Greg Charvat's "build a SAR with coffee cans and a laptop" 
mini-class/dissertation project.


 RF wise these radars are simple device, and I was asked to give a 
presentation to the JPL Amateur Radio Club on the principles and 
limitations on performance.  Most of the members of the JPLARC (like me) 
actually know quite a lot about RF design, so they'll be asking about 
"what about the phase noise of the Tx".  We all know, qualitatively, 
that gunns are noisy close in (to the bane of hams who want to do narrow 
band stuff with the old MaCom gunnplexers), although, like with the 
minicircuits VCOs, there's no published data on their 1-100 Hz phase noise.


So I'm writing up a set of notes on the various factors, and the self 
noise of the oscillator is particularly important when looking for low 
frequency modulations (like heartbeats at 1 Hz, or people walking).


I've got empirical "as measured in the system" data from my 3GHz 
homodyne radars, but I was looking for some component data as an example.



Actually, if someone has some close in data from a 10.525 GHz Gunn (or 
the newer motion detectors), I'd love to see that too.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] PHK quoted in Slate online article about leap second

[Jim Lux]

http://www.slate.com/articles/health_and_science/science/2015/06/next_leap_second_june_30_dangers_to_software_military_gps_banking_air_traffic.2.html
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] USB problems and solutions - Some what Off Topic --> USB-C

[Jim Lux]

On 6/2/15 4:07 AM, Bob Camp wrote:

Hi

The one thing I would be a bit careful about is the power levels.

Consider an 8 port hub:

5V 3A from each would be 24A total. That’s pretty unusual. Most hubs
give you one or two high current outputs.


There are very few 8 port hubs and lots of 7 port hubs (at least for USB 
1 and 2).  A 7 port hub is two 4 port hubs daisy chained.  For 
commercial products, most (if not all, I've not checked for sure) have 
only 2 of the high power jacks.


The power switching and control protocol is non-trivial, especially if 
you are connecting and disconnecting devices. On windows (and other 
OSes, too, I imagine) there's a whole thing where the OS tries to keep 
track of the state of the whole tree of USB devices, so that it doesn't 
try to send a "power up" message to a device that downstream of a 
"powered down" hub until it's sent the "hub power up" message.


There's also some weird (and entirely within spec, apparently) behavior 
of a hub where it, say, has 1.2 Amp total capability, so the first two 
0.5 Amp devices that are plugged in get the full allocation, and the 
rest do not get the "high power" acknowledgement.  Plugging in a 
combination of high and low power devices (or, equivalently, enabling 
and disabling them) can lead to things sometimes working and sometimes 
not.  (the Knapsack problem, which this sort of is, is NP hard, after all)


I've also found devices/hubs that seem to use some sort of ad-hoc power 
allocation scheme (actually measuring the power drawn, as opposed to 
just saying "high power (500mA)" and "low power(100mA)" devices when 
querying the device and/or looking at the pullup/pulldown  on D+/D-


One reference says "All USB devices enumerate as low-power devices at 
first. After enumeration, the host examines the bMaxPower field of the 
configuration descriptor for the device. If bMaxPower indicates that the 
device is high-power, and the power is available, the host allows the 
device to transition to high-power"


the whole "if power is available" might be done in real time.

And this causes real issues if you have a USB powered device that has 
multiple power modes (I have a bunch of radar modules that started out 
being USB powered, and have a low power "idle" mode and a high power 
"transmitter and receiver on" mode)



There's some complexity also with "bus powered" vs "self powered" hubs. 
 A bus powered only gets 500mA from upstream, so cannot really support 
any downstream devices at 500mA: therefore, 100mA for each of the 4 
downstream devices, and 100mA for the hub itself (if needed).



In any case, USB power management (and hub and device state management) 
is substantially more complex than one might think, and lame software 
drivers in the host can make it more complex; particularly if, as in 
most modern systems, there's lots of power management going on for 
hibernation and sleep modes.



There's a whole bunch of command line commands for Windows to manage 
this explicitly (if you don't want to use "device manager").  If you're 
doing a lot of USB stuff on windows, you NEED the devcon command, which 
gives a lot more visibility into the enumeration and hierarchy.


USB power management http://support.microsoft.com/kb/817900

devcon command:  http://support.microsoft.com/kb/311272

googling "Windows USB power management" will turn up a lot of info, too.


devcon can also be used to deal with USB COM ports that move around or 
disappear and reappear.


I have a system that has 5 Teensy 3.1 microcontrollers hooked to it via 
USB (emulating a very fast serial port).  The problem is that when the 
microcontroller changes USB device types depending on whether it's in 
"bootloader" mode or "running an Arduino program" mode.


I think devcon might also be a good way to suppress the notorious "GPS 
masquerading as a Microsoft Serial Mouse" problem which is quite annoying.











20V 5A (100W ea) on 8 ports would be 800W. That’s not going to be cheap.

Yes, the 20V is an “optional” part of the whole thing. The 3A does not appear
to be quite so easy to ignore. We’ll see what actually happens …..

Bob


On May 30, 2015, at 11:28 AM, Neil Schroeder  wrote:

USB-C will offer a number of things that I believe will be of benefit to
time nuts everywhere:

1) High current 5V up to 3A on every port
2) High voltage/current up to 20V/5A optionally on every port - sufficient
to power some rubidium oscillators natively, and a small boost to get the
rest of them.
3) a native UART channel - no more freaky USB interrupts/polling to get
your pulses
4) single omnipurpose connector ends with no insertion dependencies
5) Better, simpler device enumeration - while I haven't seen how it
addresses this personally, the stuff i have read is very promising.

Due to the switched controller nature of the interface, you should have
less nonstandard crap that may cause your computer to hang or other issues
related to drivers. The controller arbitrates a lot

Re: [time-nuts] iGPS?

[Jim Lux]

On 5/18/15 7:59 PM, Chris Albertson wrote:

Yes GPS can do better than 50M but we are talking about a single fix
from a cell phone in a moving car not a survey receiver.  .  The
displayed location is better because the phone applies a filter to the
location data.  Some thing like a Kalman filter.   I doubt the iPhone
can get within 10M from a moving car.

As for Iridium being an expensive for pay service.  But that is
because most users SEND data.  This new service is broadcast and costs
do not depend on the number of users.   Apple has sold 130,000,000
phones already just in this half of 2015.  A one time payment of about
$1 per phone might cover the costs.


Why iridium?  Why not Sirius or XM or DBS. Unless you want something 
world wide, as opposed to "populated areas served by broadcast radio and 
TV"


Heck, you could probably buy transponder time on a C-band satellite and 
radiate a GPS assist signal (that's what WAAS is, after all)



hm……I do believe the marketing boys have been playing with the
numbers. You would have to start
from a >50 M error to get them to make much sense based on what they are
doing.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] iGPS?

[Jim Lux]

On 5/18/15 11:06 AM, Brooke Clarke wrote:

Hi Bob:

In the link in the message from Brian it explains that iGPS is for
military users of the Iridium system.
The key feature is to allow a moving vehicle to lock on the GPS signal
while being jammed.  They do that and also get a more accurate fix by
using signals from the Iridium satellites.

I see a potential problem in that the Iridium signals are close in
frequency to GPS and a broad band jammer might cause a problem for both
of them.




Any system that is concerned about jamming is probably going to be 
immune to broadband jamming: after all, GPS signals are already below 
the noise floor.  Broadband jamming is a pretty ineffective use of RF power.





___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Time in a cave

[Jim Lux]

On 5/12/15 4:00 PM, Tucek, Joseph wrote:

I'm looking for information on non-GPS time sources.

For background, I need to provide PTP to a cluster where we don't have line of sight to 
the sky, and are unlikely to get roof-rights without a fight.  There are CDMA solutions 
that would work (e.g. Endrun Technologies), but I was wondering if there were any other 
options.  I either need an indoor capable PTP, or an indoor capable PPS.  Microsemi 
claims to have an indoor capable "GNSS" system, but I've yet to find a sales 
rep to talk about it; if anyone has a link to one who can, I'd love to find out the 
problems^W^W^W^W talk to them about it.



But the entire cluster is interconnected?
Or do you need some sort of wireless distribution.


For an example of something that almost but doesn't quite work, Beagle Software 
has a CDMA NTP server, but they do neither PTP nor PPS in the CDMA version.  
Similarly, Meinberg will sell a PTP unit that freeruns (if you override the 
config), but they have no solution to discipline via CDMA.

I'm also curious if anyone has any idea about non-GPS time sync after CDMA gets 
turned off (can I get time from 4G?).



What about an off the shelf Rb that puts out 1pps or NTP or PTP.
SRS (http://www.thinksrs.com/products/fs725.htm) does 1pps
MicroSemi (aka Symmetricom,HP,Datum, etc.) has all manner of equipment 
that has quartz or Rb references and probably any interface you care to 
name (for a price).


If you need sync to outside sources periodically, most of these could be 
"carried" to somewhere you have sky visibility to get a GPS 1pps, wait 
long enough to synchronize it up, then carry it back down and go on 
holdover.





My endgame worst case is to just do PPS from a stratum 2 NTP (or even a 
freerunning oscillator) and lie to my PTP server; hard sync to UTC is a 
secondary concern so long as the cluster agrees with itself.  Endrun is looking 
pretty good, but I'd really like to have a second option to compare against.

-Joe
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] EMI and CE certification

[Jim Lux]

On 5/11/15 12:52 PM, Adrian Godwin wrote:

Is it driven as  an inductive loop? That might put it under different
regulations.
On 11 May 2015 17:47, "Chuck Harris"  wrote:


Yes, but in the case of the lawnmower fence, and the
invisible dog fence, the transmitter drives the fence
as an antenna.

In the US, the antenna size for "free bands" is seriously
limited.  As an example, the so called "Lowfer" band at
136KHz is limited to antennas no larger than 15m in length.

And, that is one of the larger limitations.

15m would encircle only a very small lawn.

OBTW, I realized on reading my post below, that I was very
unclear on what could "be foiled."  I meant that the
operating permission for the lawnmower system could probably
be foiled by looking into the maximum antenna lengths for
unlicensed services of this sort, in this frequency range.

I would quite imagine that any certification they may have
is for the transmitter and receiver, without an antenna.





The US FCC Part 15 limit is probably 2400/f(kHz) uV/m field strength at 
300m distance.


It's probably pretty easy to meet that.


https://transition.fcc.gov/Bureaus/Engineering_Technology/Documents/bulletins/oet63/oet63rev.pdf

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] SVG Re: lawnmower robots may be the end of VLF timekeeping

[Jim Lux]

On 5/10/15 11:40 PM, Poul-Henning Kamp wrote:


In message <45C7C6B09BC548C19241E4E0673E9E9F@system072>, "Bill Hawkins" writes:


Did the pictures have to be in SVG format?

Is this only a problem for those who routinely use SVG?


A problem how ?

I *like* SVG since you can zoom without pixellation effects,
and spent an afternoon writing code to screen-dump the HP8568
into SVG format for the very same reason.





I like SVG for the same reason..

Not all browsers provide the same flexibility in viewing the data, or 
are as "smart" when plotting a very dense set of lines: say you've got 
30,000 datapoints which are individual vectors... if you were displaying 
this in a tiny window that is a few hundred pixels, a smart rendering 
engine would do some "collapsing" of the vectors, which would make 
rendering faster.


There are smart and less smart renderers of .eps and .pdf too..


I wish that some of the more popular tools (Matlab, Octave) would 
directly export their plots as svg.  For all I know the latest version 
might do that, I should check.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] lawnmower robots may be the end of VLF timekeeping

[Jim Lux]

On 5/9/15 5:15 AM, Poul-Henning Kamp wrote:

I spent some time capturing some data today.

The measurements is from my $20 loop-antenna in the attic, which is
something like 8 meters up and 10 meters besides the lawn-mower loop:

http://phk.freebsd.dk/time/20150509.html



Clearly, you need a better nav system for the robot based on precision 
time of flight measurements from a network of transmitters around the 
property linked to your hydrogen maser.


More practically, what about some sort of canceller.. This is low 
frequency, so if you put a "pickup loop" near the wire, you can collect 
a sample of the transmitted signal, and then adjust the mag and phase to 
cancel at your timing receiver antenna.  I suspect the variation in mag 
and phase will be quite small over time/temperature/weather/volume 
occupancy.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Important parameters for a GPS/GNSS antenna

[Jim Lux]

On 5/8/15 11:37 PM, Poul-Henning Kamp wrote:


In message , Bob Camp writes:


The “put the antenna up and rotate it to see what happens” experiment
has indeed been done. The objective was not correcting the antenna’s
issues, but validating that their model of the antenna’s phase
center was correct. They were trying to see if anechoic chamber
data really gave correct answers in free space.


So this could be a realistic way for us to calibrate the phase-center
of an antenna ?




Yes.. actually, the best way in the long run would be to collect many 
hours of GPS satellite data (carrier phase)  with the antenna in one 
position.  Then rotate the antenna to a new position, collect a bunch 
more data, repeat, etc.



Then, you post process using the known position of the satellite, which 
gives you a direction of arrival relative to the antenna.


You probably don't need so have a real precise position for the antenna: 
the apparent motion of the phase center as a function of az/el is 
probably fairly slow.


Isn't that how they collected phase center data for all those antennas 
on the UNAVCO site:

http://facility.unavco.org/kb/questions/458/UNAVCO+Resources%3A+GNSS+Antennas

one of the reports has this interesting statement:
Antenna rotation tests work well to identify inconsistencies in mean 
phase center offsets. By occupying a short baseline (less than 10 
meters) and rotating the antenna orientation 180 degrees it is possible 
to see changes in the baseline length caused by the antenna phase 
center. For antennas of the same type, the rotation tests will highlight 
variations of an individual antenna relative to the pool of antennas.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Important parameters for a GPS/GNSS antenna

[Jim Lux]

On 5/7/15 7:23 AM, Attila Kinali wrote:

On Sun, 03 May 2015 07:29:30 +
"Poul-Henning Kamp"  wrote:


When you post-process raw GPS data you get to include antenna phase
center / gain / az/el corrections for free.


Speaking of which...

I wonder if anybody ever made a rotating GPS antenna to average out the
X-Y phase-center offset ?


There is a severe mechanical problem with that. Moving contacts
are very hard to keep electrically stable. It works for simple
power and digital signal wires, as there the only considerations
are resistance over the joint and sparks. If you need to transmit
analog signals you generally convert them into a form that does
not depend on the amplitude of the signal (either going digital
or doing a voltage to frequency conversion).

Of course, this does not really work with a gps antenna, unless you
put the whole receiver onto the rotary table. But then you shift
the problem onto the PPS output (note: amplitude noise translates
into phase noise).

The best system i know about, for such rotating contacts with analog
(possibly high frequency) signals is a small pot with mercury. But
i guess you can see the problems that causes.

The second best, but which only works with high frequency signals,
is to use a hollow waveguide. There you need "only" to ensure that
the waveguide walls are properly connected and you have a large
area to use for that. If you can ensure that there is only one mode,
you can make it such, that there is no current flowing over the gap.



They make rotary coax joints as well as waveguide. The coax is an air 
dielectric type.


There is, inevitably, some wow and flutter in S21 (and S11)

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Lucent KS-24361/Thunderbolt antenna advice

[Jim Lux]

On 5/6/15 3:09 PM, Bob Camp wrote:



GPS helix antennas were a really big deal in about 1982. Once people started to 
get experience with GPS and a variety of designs, they became less of a big 
deal. I do not know of any modern
precision antennas that use a helix.



Most precision antennas I've seen recently use some form of a crossed 
dipole, with drooping elements with a weird shape (to get the match 
decent at all the frequencies, and to get the relative phase shift right).



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Lucent KS-24361/Thunderbolt antenna advice

[Jim Lux]

On 5/6/15 12:53 AM, John Marsden wrote:


Ok, I only ask becuse there seemed to be a big thing about LHCP quad helix 
antennas - even to the point of seein an article showing how to 'unwrap. a RHCP 
Q-H, and rewrap it 'inside-out' to change the polarisation to LHCP.
I'm seriously considering making an active Q H if I can't find anything that 
looks promising - pending the answer to the question above, of course - I don't 
want to spend $100 making a '$40' one ;)




Helical antennas are really non-critical in terms of design.  I'm not 
even sure that a quadrature helix is actually what you want: they tend 
to have more response at the horizon (for a short helix) than straight 
up, but I would think that for timing applications, you'd rather get 
strong signals from overhead, rather than low angle signals subject to 
multipath, etc.


(remember that the GPS satellite's transmit antenna pattern is bigger at 
the edges than in the middle, so the incident flux on the ground is 
about the same regardless of elevation angle: this is different than the 
typical case for, say, LEO amateur radio, where the satellite is 
essentially omni, and you want an antenna with more gain at the horizon)


A quad helix was popular for handheld GPS units because it's got a 
reasonably good pattern that's almost omnidirectional (even if the axial 
ratio isn't so hot in all directions), so it works well in any orientation.


A regular helix with a gain of 10 dBi or so  (4 turns) will have a 3dB 
beamwidth of 52 degrees, and will be about 6 cm in diameter and 20 cm 
long.  3 turns is 60 degrees bw (3db)


The folks at JPL wind their helices on things like appropriately sized 
plastic cups and put them on one of those bowl shaped things that fit 
under a stove burner as a ground plane (hence the "helibowl" moniker). 
You get a decent match by adjusting the distance of the end of the 
bottom turn from the ground plane (making a sort of tapered transformer 
from the 100-150 ohm impedance of the helix to the 50 or 75 ohms of the 
feed line)



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Tuning a Trimble Thunderbolt

[Jim Lux]

On 4/20/15 7:25 PM, Charles Steinmetz wrote:



Unfortunately, you are unlikely to do any better than this with the
antenna location you described.  Time to buy a house, with no tall trees
nearby.  (You may already have heard that time-nuttiness can be
expensive  ;-)




Actually, what you want is an isolated mountain top of solid granite 
connected to a big pluton below, with a nice drilled anchor.  Your 
laboratory/lair can be in a cave below the mountain top.


Being on top of a tall mountain also reduces the effect of atmospheric 
effects, but will increase the solid earth tides (and, of course, you 
are closer to the moon and sun).


I wonder if being on top of a dormant volcano (e.g. Mt. Waialeale on 
Kauai) would be stable.  An active volcano (Mauna Kea/Mauna Loa) is 
going to be moving around a lot.  Tahiti, perhaps?  Nice pleasant 
weather.  Mt. Roriama in Venezuela appears to be well above the 
surroundings, but is quite flat on top, so you'd need a structure to get 
your choke ring antenna up high enough.  Close to the equator too.


As the previous poster pointed out, serious time-nuttery can get expensive.

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.