Recently some Datum Startloc II GPSDO's appeared on Ebay (originally around
$50, now $100). The main seller appears to be in Canada. These are a telco
type GPSDO that appears to be an attempt at a replacement for Trimble
Thunderbolts. I recently got one in and did some playing with it
If OpenPnp generates a gcode file for a particular CNC controller (say Linux
CNC) my program can easily translate it into something pretty much any other
CNC machine can read (including some machines that have not existed for over 40
years). It would not be hard to add the ability to read a
Trimble can operate in a 3D positioning operation mode that works that way.
But their "overdetermined clock" mode is what most people call "position hold"
mode. Set the receiver to "overdetermined clock", and the lat/lon/alt
coordinates it sends never change.
As far
I do mostly digital stuff and 80% of my resistor usage is 10K, maybe 10% 1K,
a couple of other values are 9%, random values make up the remaining 1%.
For caps I use 0.47uF 0805's most of the time and 10uF (can't remember the
size). Why 0.47 uF you may ask? Well, I thought I was buying
It is basically what most people call "position hold mode" where the receiver
has known (and accurate) fixed coordinates. With lat/lon/alt known, it can
solve for time with just one satellite visible. Or, if more than one sat is
visible, its time solutions can be more accurate because noise
I may have mis-spoken on one aspect of these devices... it is now seems to be
doing 4-hour power-on surveys... Or perhaps it has some measurement threshold
that allows it to end a survey before the full 14,400 samples have been taken.
I did a power cycle, went to dinner, and three hours
I've done a over one hundred boards with a Kapton stencil without any problems
(0603 size parts) and 50 boards with a stencil cut out of vellum paper (a
plasticized paper available at craft stores, red cuts very well with 405nm
lasers) by my home-made laser cutter head for a desktop CNC
I have seen pick and place systems built around CNC machines (same applies to
3D printers). The reel strips are fed through a slotted guide. The pickup
head has a finger (or some use the pickup nozzle... a flat tipped hypodermic
needle) that is used to advance the reel. It drops down into
All timing receivers (including Trimble) calculate an overdetermined clock
solution if more than 4 sats are being tracked (and the signal meets their
TRAIM quality level). But, again, in Trimble-speak, "overdetermined clock
mode" is what most manufacturers call "position hold" mode.
> The dynam al capacitance is a couple of fF, and thus R1 is between 10k to 80k
> depending on the size
A friend of mine has been fighting a 32 kHz clock problem on an ARM processor
(used to wake up from periodic sleeps). In discussions with the CPU and
crystal makers it looks like some of
I'm partial to the Youyue 858d. Not because it is particularly good, but
because there is some nice open-source firmware for it. It's also rather
inexpensive (around $40). Note that there are a lot makers of -858d
stations. Not all use the AVR chip that the open source firmware runs
> When is some organization going to explain what happened in February for
> almost two hours starting at 00:16 GMT? That subject has gone silent. Rob,
> NC0B
I heard back from NAVCEN. They said it was a Trimble issue and that Trimble
would contact me (they didn't). But that does not jive
Also, when messing with a 10544A be aware that that the frequency tuning
adjustment must be made with an insulated tool.
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
Not to mention that a lot of pots that have been sitting at the same setting
for years tend to distort the resistance element where the wiper has been
making contact. Then when you make a small adjustment you cannot reach the
value that you need. Ahhh, the subtle wonders of aligning old
Actually not hard to do... lay out circuit board (free version of Eagle),
have boards fab'd at Oshpark.com or your favorite Chinese proto shop (I like
gojgo.com). Have solder paste stencil made at oshstencils.com. Squeege
solder paste down with a credit card. Place components by hand.
Yes, all satellites dropped out from 00:16:51 to 01:54:02 UTC in Dallas,
Texas. Was tracking 5 sats then, poof, none.Global warming? Gravity
waves? More bogus GPS control segment uploads? Somebody gots some 'splain to
do...
I filed an outage report with NAVCEN via their web site and got an email back
asking for more info (which I supplied). I got the impression that they were
not aware of what was going on. Googling around also does not show any
official reports of problems. Rather surprising considering the
Attached is a Lady Heather plot of the "event". Note that there were a couple
of brief places during the outage where one satellite was tracked.
___
time-nuts mailing list -- time-nuts@febo.com
To
NIST is working on a quantum thermometer. Apparently the current version is
not very accurate, but they are working on it. If it ever comes to be,
should be interesting... particularly when paired with their photonic
thermometers.
I believe that it was mentioned here that the BG7TBL GPSDO disciplined
oscillator controls the OCXO using a FLL (frequency locked loop) instead of a
PLL (phase locked loop). Their implementation causes a slight error in the
output frequency. Search the time-nuts archives (best to use Google)
Mark's Law of Rubidium Goodness... the bigger the box, the better it is. The
HP5065A is one of the best units ever made. It can rival a cesium beam unit.
The X72 is a horrid little creature.
I would also go with the M100 / FRK units. The LPRO an FE56xx units were
designed for telecom use
When an old electronic gizmo starts failing across multiple units, the culprit
is almost always failing electrolytic capacitors. Check your GPS board for
electrolytic caps and replace them. There could be some other component that
is susceptible to an age-related failure mode, but caps are
In many ways the 5065A is the probably the most repairable of all the units
(closely followed by the FRK family and the M100). They all use parts that
are mostly still available and the circuitry is accessible. You can assume
that the lamp (and maybe some of the microwave parts) in any Rb
Yes, they do. They seem to use different orbital, gravity, geoid, TAI, etc
models. Trying to merge fixes from different systems leads to all sorts of
problem, particularly at nuttery levels. I've seen reports that GPS and
GLONASS fixes can be off well over 50 meters when referenced to the
And that can be very interesting... a while back I read some stories on how
the NSA, police, etc could find out where an audio recording was made by
correlating AC hum in the recording with logs that they had of the power grids.
Apparently logging AC mains is rather popular among the spooky /
Another issue is that FTDI is back at it again with making their drivers
incompatible with clone FTDI chips. Their previous attempt actually bricked
the clone USB interface chips. Their latest driver just sends garbage out the
serial port. If you are using a cheap USB to serial converter
Apparently the PLCC version of the chip is a lot easier to find than the DIP
chip. One could make an adapter board to use the PLCC in a DIP socket. The
PLCC will probably not fit between the rows of legs of the DIP, so the adapter
would need to be bigger than the DIP footprint. Or if
Try iy with a pull-up. I have seen LPROs with an open collector type of
output.
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and
Heather can log the binary packets that it sees.Use keyboard command "oj"
to toggle packet logging mode. Then use the "wlw" command to open the log
file. Let it run for a couple of minutes. Close the log file with "wls".
> Mil-spec parts would be somewhat more reliable than commercial parts.
Actually, that is seldom true. The main difference between mil-spec parts and
commercial parts tends to be in the post-packaging device testing (e.g..
extended temperature / voltage range). They usually have the same guts
On the subject of RS-232 converter chips... I have had problems running
MAX3232's at 3.6V. Sometimes they work, sometimes they don't. Sometimes they
get freaky hot. The general symptom seems to be no -6V output.
I've never had problems running MAX232A's at 3.6V, even though they are
Well, if you think the 5V ref is bad, you could pull the ref chip and jumper
the logic 5V to its output pad (through a resistor... maybe 1K) to do a quick
and dirty test. If the resistor output is dragged down, you know something is
loading down the ref voltage.
> This is great news. Will the code be available for inspection /
> modification?Yes. It is the same source code as the current version. Just
> compile under Linux and voila... The code currently has a no-commerical
> copyright on it, but I am going to GPL it (if I can ever figure out what
John Miles is now hosting the latest Lady Heather 4.0 files. His server is a
little less skeezy than tinyupload.com (which on Windows machines apparently
offers to update your Java with about every click). I've used tinyupload for
years on a Mac, and never got such a "generous" offer. No
I have uploaded a ZIP file to tinyupload.com with the source and a Windoze .EXE
file for the beta test version of Lady Heather. For Windows users, copy the
heather.exe to your current Lady Heather directory (might want to first rename
or copy your old heather.exe as a backup.
For Linux users,
She lives! She sings! She dances!Turned out it was not the serial port that was
blocking... it was the X11 event handler. XNextEvent blocks if no event is
pending. There is a X11 call that got around it...
So far everything seems to be working well. I still need to do a lot of detail
work for
>Would there be enough horsepower for a Pi 3 to run Lady Heather and act as a
>stratum 1 NTP server?
I suspect so, the PI3 has quad core 64-bit capable 1.2GHz processor. The PI3
seems to be about 50% faster than the PI2. It also runs about code about as
fast as a 2 GHz Pentium 4. But the
Ars Technica just published a piece on leapseconds. There is some interesting
info on what makes the earth a poor clock and our abilities to measure it
(want to smooth it out a bit... cut down all the trees).
The question came up about Lady Heather's system usage on a PI3. Turn's out
she's quite the little resource hog. CPU usage hovers between a whooping 2 and
3 percent ;-) Running on a PI2 showed about the same.
Slinging a window violently around on the srceen and it peaks to around 20%.
So
There is what looks like a decent carrier board for Ublox modules on
OSHPARK.COM's shared project library. It has a voltage regulator and RS-232
interface on it... would be nicer if it had a prototype area and swoopty PPS
driver, but I'm too lazy to lay out a better one. Three boards will
I have added some code to Lady Heather to talk to SCPI GPSDOs (also Motorola,
Ublox, and NMEA) and am looking for a guinea pig to test it... I have a Z3801A
and want to see if it is also 3815/3816/etc compatible. Please contact me off
list.
I might be able to send a WIndows .EXE , but a lot
Lady Heathe does have sidereal time support. Set the time zone name to LMST or
GMST or LAST or GAST and you get sidereal times...
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
Alas, it has become very hard to email an EXE file. The ISPs are on to all
the tricks... zipping them, renaming them, etc no longer work. The net has
become a real nanny about protecting people from potential malware, etc.
Things like EXE files disguised as PDF file were/are a real threat
I think the EEPROM in AVR chips is rated at 100,000 writes. Several people
have tested them and most found they would survive at least 5 million writes...
YMMV, of course.
One write every 24 hours is still under
10K writes in a reasonable lifespan for a GPSDO. It’s a
If you want to get fancy you could modify Lady Heather to do the deed. Heather
has a "singing clock" mode and a cuckoo/chime clock mode for playing sound
files at various times. It also has routines for controlling the serial port
DTR and RTS lines (currently used for PWMing a fan if
> See www.leapsecond.com/tmp/5071a-12-run8-5d-10d.gif for a plot of a bunch of
> 5071A Cs clocks.
Interesting... what happened on day 4 to send them all bonkers?
___
time-nuts mailing list --
Also, a lot of laser diodes don't like to be "cold-started". Your modulation
scheme needs the laser to always be on at some minimum level above Ith. Just
crudely switching from off to on can quickly kill the diode.
There are "eye-safe" wavelengths that some laser diodes can operate at
(generally greater than 1300 nm). These are much less prone to damage eyes.
Basically your eyeball juice blocks the wavelength. Still, there is some
potential for cornea and lens damage at higher powers.
You could try Lady Heather's precision 48 hour survey. It basically does a
weighted average of median filtered fixes over 1 hour time intervals. The
median filtering automatically throws outliers out of the data.
> Sorry, no Thunderbolts. TVB emailed me to ask if I had one, to make
> the same suggestion.
Lady Heather also works with things like the Trimble Resolution-T and
Resoultion SMT timing receivers (latest version includes the RES360). You can
get a Resolution SMT on Ebay for $25. I would go with
Well, Lady Heather finally got off her ample ass and dug into the tao of X11
and all things Linuxy. The program is currently working well enough to
display log files, etc. Serial port initialization for non-blocking
asynchronous I/O needs to be completed. I have it working on a couple of
> Why not use something that is 100% respectable and well known, like github
Yes, it will put up on Github (and with a GPL type license), but the Linux
version is a bit too much of an early work-in-progress hack right now.
I would greatly appreciate and comments / experiences from anybody
Here are the results of measuring the difference between the time code in a GPS
receiver time message and the arrival time of the last byte of the message.
Negative values mean that the receiver sends the timing message after the 1PPS
pulse that it describes. The table also shows the standard
And somebody paid well over 10 times the market rate to lease server space that
was one building closer to the NYSE computers. My idea to put an end to this
bogo-trading nonsense is to add a random delay to all the trades.
> Someone recently built a new fiber route
Heather will take either time format, but requests the receiver to send T2
format.
I originally thought the SCPI receivers would be right on time due to my
original measurements of their message jitter, but when I started measureing
the actual message arrival times... surprise, surprise,
I have replaced the fans in all my 5370's with quiet 12V (150-250 mA)
brushless DC fans running off of the 10V rail. I placed themocouples around
the inside of the unit and compared before and after temperatures and found no
significant differences. I have used several different models of
I got in a Jupiter-T Pico timing receiver (from the good folks at RDR
Electronics). These are pulls from something made by Lucent. It is a tiny 12
channel timing receiver with an even tinier (0.5mm pitch) 2x10 pin connector.
RDR ships them with a mating connector and an antenna pigtail with
I saw no problems with Ublox Neo 6M, LEA-5T, LEA-6T, Ublox 7, or Ublox 8
receivers.
A Synergy M12+ did not report the leap pending in the @@Bj message, but did in
the @@Gj message.
-
Anyone know if the M12T or Lea Bloc GPS chips have the same problem as the
I bit bang serial all the time... it is much easier if to do if you are only
banging serial output. Reading serial streams at higher bit rates while doing
other useful stuff can be challenging... particularly if you don't have a timer
handy.
I once coded up a PIC to output a time code /
The closest thing to an official Lady Heather site if John Mile's ke5fx.com
I also want to put it it up on Github when the code settles down some.
The good Lady Heather doesn't accept donations of the monetary kind.
Receivers not currently hoarded in her dungeon are appreciated.
The Skylab is effectively useless for sub-second timing. The message arrival
time periodically jumps around, up to +/- 300 msecs. There are a couple of
values that it seems to prefer, but any value can be seen.
NMEA works a lot better on most receivers that I expected. They send several
Lady Heather keeps the date/time time in two sets of variables. One is the
receiver time message values in UTC (or GPS) time. The other is local time
(UTC/GPS adjusted for time zone or time scale) that is used for the clock
displays. There are integer year,month,day, hours,minutes,seconds
Lady Heather now works with Venus receivers in mixed binary / NMEA mode. So
now it can display the sky view data (via NMEA sentences) while running in
Venus binary mode. I also added support for parsing the $PSTI NMEA sentence
that contains the sawtooth correction info. Why somebody
Navspark will sell you one in on a board for around $80. They claim 6 ns
timing accuracy. Also 0.01 ppb on the 10 MHz output. I sent Navspark a couple
of questions about their USB interface and they never responded...
http://navspark.mybigcommerce.com/ns-t-precision-timing-gps-receiver/
I have implemented support for all of the relevant Venus binary messages.
It's just that I don't have a Venus timing receiver to test the timing receiver
specific messages (which are only three messages that control the self-survey
/ position hold modes). Oh, and besides the lack of a
I'd get the $80 assembled unit from Navspark first and see if it works for you.
If not, you won't have wasted your time and money on doing a PC board.
Lady Heather can now talk to Venus receivers in binary mode (also handles
generic NMEA receivers). I don't have a Venus timing receiver so
I am the creator author of Lady Heather. The code in Lady Heather started out
from a program that I wrote to control Magellan GPS receivers (like the
OEM-5000) back in days of the first Gulf War... it ran under DOS as a text mode
only program.
When TAPR did their group buy for Trimble
I have added some code to my message time offset measurement routine to
calculate a histogram of the values (along with the average and standard
deviation). I'm now using the peak histogram bin(s) to determine the message
offset time. The histogram technique has the advantage of ignoring
Easy, I have a 4 value array. It keeps count of the number of times each
0.1 msec step between -2000 .. 2000 msecs was seen. If two or more bins wind
up with the same max count, I report the average of those bin times, otherwise
it's the bin with the highest count. I also dump the
Yes, I find it confusing also. I've been reporting the value that Lady
Heather uses to do most of her evil internal message off machinations with...
For those, the negative offset value is the "natural" polarity. I'm probably
going to change it around to something humans (including me)
Well, the whole point of the exercise is to see how well you can do if you
DON'T have an internet connection, a 1PPS signal, or a stratum 1 time server
available... only the humble messages coming from a 10 dollar GPS receiver.
Try getting a net connection in the middle of the Gobi desert
I have a Z3812A that I added back in the GPS receiver and modified it (moving
6 zero ohm resistors) to work as a standalone GPS. Basically it's now a
stand-alone Z3811. I have Lady Heather working with it now. The hack of
using the RS-422 output to directly drive a RS-232 serial port does
Apple's new file system timestamps files with nanosecond resolution. A lot of
Linux file systems also do that now. The nanosecond ain't what it used to
be... I can imagine people wanting picosecond timestamps in the near future.
Who knows, maybe we'll have something like NTP compensating
Attached is a plot of the timing message arrival times of the Trimble
Resolution-T receiver running in TSIP mode. Like the Adafruit Ultimate, its
firmware seems to do little to get the timing message out at a fixed reference
point.
As Tom pointed out, this is only an issue if you are
A couple of people have asked about the poor message arrival time performance
of the popular Adafruit Ultimate GPS receiver. I modified Lady Heather to
analyze the message arrival times using a histogram instead of a simple
average. When I looked at the histogram data (.01 msec resolution),
It is only tracking the GPS sats, running at 1Hz rate. Almost every GPS
receiver that can run faster than 1Hz has warnings that the 1PPS output is only
valid/stable if the device is configured for 1Hz output.
My Ublox 8 receiver tracks GPS/Glonass/Beidou/SBAS sats. I have seen it
As mentioned in the post the times reported are the time stamp in the receiver
packet minus the system clock time when it was received... negative value
indicate the message arrives after the PPS. The polarity of the reported value
is consistent with how Lady Heather makes use of the
I found what was causing the apparent ramps in the message offset time for the
Motorola mode receivers and the Z38xx receivers. Here is the updated and
corrected info. Note that a couple of receivers do have some minor ramp
sawtooths in their message timing. They are less than +/- 10 msecs
And it does NOT say they DON'T use the '838. No where do they mention an '822.
They do mention an '838 and the data-sheet they link to is for the '838.
Their listing seems to meet the legal definitions of deceptive advertising and
bait-and-swith sales. If they were in the US, they would
Pretty much all the telecom GPSDO's are driven by an on-board, isolated DC-DC
converter. Some are 24V nominal input, some are 48V (a few are even 18-72V).
The '+' / '-' designation is pretty much irrelevant since the converter inputs
are isolated and are not referenced to chassis ground...
Lady Heather has the ability to set the system clock from any just about any
GPS receiver. It can set the clock every day, hour, minute, or when system
time and receiver time diverge by more than "x" milliseconds. Most stock
Windows systems have a clock granularity of +/- 15.6 milliseconds,
The NVS NV08C GPS receiver module is a rather nice little 32 channel
GPS/GLONASS/SBAS receiver that has an output message that gives you the raw
satellite bit streams. They can be had for around $60 bare module ($120 on a
breakout board from sylphase.com). They also support carrier phase
The NTBW50AA (and most other telecom GPSDO's) will run on 48V. This will
solve your problems for dirt cheap. They are very high quality
industrial/medical supplies (not cheap junk wall-warts) that currently sell
from distributors for over $80.
Lady Heather can now work with GPSD... it is one of the simpler "receivers"
that it works with. It took about 500 lines of code (about the same as NMEA)
to interface GPSD to Lady Heather. Most receivers with binary messages take
around 1000 lines, but they usually support more features.
A friend on mine once worked on projects to build very sensitive magnetometers
(submarine detection and space probes). Their test lab was in the middle of a
square mile of land selected for its low magnetic residue properties. The
building was made completely (of carefully selected) wood...
A lot of the newer GPS modules can accept "dead reckoning" data from external
sensors and integrate that into their location solutions. The receiver does
all the Kalman filtering foo for you. Also you can get receivers with up to
50Hz position updates (you have to run the com link at very
I have seen some reports of some Skylab GPS boards that had some rather sketchy
GPS code. A friend had one that was off around 0.5 degrees in longitude...
but only if your location was between +/- 90 degrees longitude. He was on the
east coast and had the problem. He sent it to me (96
Most definitely... if I turn on Lady Hather's display filter (which does a
sliding average of "n" readings) the standard deviation and jitter values drop
dramatically. With a 60 second filter the deviation was down to around .15
msecs and the peak-peak jitter below a millisecond. Average
I got in a Z3812A (aka Lucent KS-24361 REF-0) and installed a Motorola UT+ in
it following the excellent guide by Peter Garde (google is your friend) and
have it working as a standalone GPSDO...basically it now a REF1 unit). The
mod mostly involves moving half a dozen 0 ohm resistors (I
I added the ability of Lady Heather to calculate the time offset of the timing
message from "wall clock" time. It calculates the difference between the
system clock time to the time that the (end of) the timing message arrived.
The result is only as good as your system clock, so the system
For a clock that is observed by eye, the NMEA message on most receivers comes
out close enough to the second to not need the 1PPS. I think there is even a
section in the NMEA spec that says how close to the second the message should
come out. Receivers that output binary messages are even
I added the ability for Lady Heather to measure a plot the difference of the
arrival times of each timing message (actually the time when Lady Heather
receives the last byte of the timing message from the operating system). The
end-of-message arrival time is time stamped to nominally
Ooops, that tboltjit.gif plot was actually from a Trimble Resolution-T SMT
timing receiver running with an indoor antenna, not a Thunderbolt. The
Thunderbolt with an outdoor antenna is about twice as good.
I just added the ability to feed the timing jitter value to Lady Heather's real
time
Heather's gotta work with XP (and maybe Win98)... too many people (including
me) run it on old trashy laptops, so no fancy pants new fangled Windoze calls
allowed...
In the past I've avoided the use of QueryPerformanceCounter due to potential
issues with AMD processors, multi-core processors
That plot was from a 12 hour run. I've done 48+ hour runs and did not see
anything strange. I'm not currently measuring the offset of the message from
the 1PPS, just the stability/jitter in the timings of the last byte of the
timing message. If the timing message offset time from 1PPS
Very few people have the ability to build Lady Heather for Windoze from
source... pretty much everybodyl uses the binary. Having people build from
source is not an option (except on Linux, it's relatively painless and the
tools are easily available and free, plus Linux users tend to know what
Or use the sawtooth compensation value to control an external variable delay
line circuit to move around the PPS signal from the receiver. This can get
interesting to implement if the receiver can output negative values for the
sawtooth compensation (hint: add a bias to the sawtooth value to
My LEA-6T board came out of one of those cheap Chinese drone pucks. It has a
ceramic patch antenna on board. I have it sitting on the floor of the lower
floor of a two-story, stucco over wire mesh house (not close to any windows).
Same for some Sirf, Venus, and standard Ublox receivers.
Yes, turning on the display filter only gives an indication that one could
gain some some benefit if they were to use the message arrival timing to
implement some sort of NTP-ish algorithm to their 1PPS-less clock.
I have added some options to Lady Heather calculate the adevs of the message
I think they are all pretty much the same unit. I don't know if any of the
oscillator models is any better. I assume they are all equivalent and actual
performance (like any OCXO) is luck of the draw.
Those boards run on 6VDC. The Trimble units use a 5V oscillator (which lots of
places
801 - 900 of 1485 matches
Mail list logo