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

[Richard (Rick) Karlquist]

It's been 20 years since I presented that paper in San Francisco
at FCS and I had just about forgotten about it.  It is
flattering to realize that people are still reading it now.

It might be useful for the discussion here if I explained why I
wrote the paper.  We had recently completed the 5071A, which,
unlike previous cesium standards, did not use the C field for fine
tuning.  This strategy turned out to be a real winner, provided
that we had a solution to offsetting the frequency.  The
5071A architecture provides something of a free lunch in this
respect because the synthesizer doesn't have to use the
10811 as a 10 MHz reference and then generate "almost" 10
MHz.  That is problematic for reasons I will explain below
that are probably familiar to many time nuts.  In the 5071A,
we multiply the 10811 to 9280 and then offset it about 87
MHz using a phase locked crystal oscillator that is driven
by the DDS.  This avoids multiplying up the DDS and the
XO cleans up DDs spurs, except very close in ones.  By
using a custom DDS, we could play our cards right to avoid
close in spurs.  The mix from 9280 to 9192.63277 is done
using a sideband, so the "mixer" is "free".  The CBT is
a "free" filter to remove the "carrier" and opposite sideband.
If you want more details, you can read my 1992 paper about
it.

None of this gives us a leg up on a bolt-on circuit that
makes almost 10 MHz from exactly 10 MHz.  That is the worst
thing you can try to do with a DDS in terms of spurs.  The
often seen paradigm of cleaning up a DDS with a phase locked
crystal oscillator is limited because (at least in 1995) DDS
chip sets use RF DACs limited to 10 or 12 bits.  Even now,
RF DAC's are limited to an advertised 16 bits, but still only 10
or 12 effective bits.  So if you want really low spurs, you
can't get them with a DDS.  The XO clean up doesn't work for
close in spurs, inside the PLL bandwidth.  And if and stuck
with an almost but not quite unity synthesis ratio, you are
painted into a corner.

I had always been fascinated with the 5100 direct synthesizer.
In theory, this is extensible to arbitrarily fine steps and
arbitrarily low spurs.  An interesting feature of the 5100
was that the finest resolution "decade" was actually a
continuously tunable variable frequency oscillator.  This
was feasible because the oscillator was divided by my a
huge divisor and could provide a useful output with "infinite"
resolution.  What I did was de-construct the 5100 architecture
and reduce it to a skeleton architecture that took advantage
of the near unity ratio to simplify the block diagram.  The
continuously tunable oscillator was replaced by a DDS.  It
took advantage of the availability of cheap 10.7 MHz filters.
Even if I could have chosen any center frequency I wanted,
these turned out to be about right, assuming you are working
at 10 MHz.  If you wanted to make a 5 MHz version, or a 100 MHz
version, you would have to get custom made filters.

About that time, HP started getting into the "smart clock"
business.  Although I was able to add varactor tuning to the
E1938A oscillator without degrading its stability, it
created a new problem of how to generate the DC tuning voltage.
To help with this, I put a 2.5V voltage reference inside
the E1938A.  The initial reference I chose turned out to be
noisy and I had to replace it with a lower noise chip.
Both the tuning voltage and the reference were brought
out of the E1938A "hockey puck" with dedicated return
lines, to avoid ground loops.

This still left the problem of coming up with a very stable
and accurate DC DAC.  IMHO, the smart clock designers never
really solved this problem.  They just got by with a marginal
system.  They didn't have the space to build an offset
synthesizer like the one in my 1995 paper.

A few years ago, Agilent introduced an arbitrary waveform
synthesizer using a very advanced RF DAC code-named "Griffin"
that was also proposed to be used in various other instruments.
This chip could be used to make an interesting offset
synthesizer, but the cost and complexity would make the
architecture in my 1995 paper look really attractive.
The Griffin can generate any frequency, but if you only
need a very narrow band of frequencies, all that capability
goes to waste.

Rick Karlquist N6RK
(now retired from Agilent/Keysight)


On 12/8/2015 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 

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 12/8/15 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?

[Poul-Henning Kamp]

In message <56675da0.4050...@rubidium.dyndns.org>, Magnus Danielson writes:

>> So, what did I miss? Why do people use DAC-EFC control instead of
>> the DDS scheme?
>
>The main reason I would say is habbits, people have habbits and stick to 
>them.

And I think right after that comes the problem of spurs.

Controlling DDS spurs is mostly a solved problem, when you get to choose
your input and output frequencies for that purpose.

But as "afterburner" for an OCXO, there is no restriction on the two
frequencies, and they are almost guaranteed to be close-in multiples,
so coping with the spurs becomes non-trivial.

That basically means that you need 50% higher DDS and DAC resolution
than everybody else in the market an suddenly it isn't cheap any more.


-- 
Poul-Henning Kamp   | UNIX since Zilog Zeus 3.20
p...@freebsd.org | TCP/IP since RFC 956
FreeBSD committer   | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
___
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?

[Attila Kinali]

God eftermiddag,

On Tue, 8 Dec 2015 23:45:52 +0100
Magnus Danielson  wrote:

 
> If you would setup essentially a micro-stepper design, such as those 
> being used for cesium and hydrogen masers, but maybe adapted to a 
> hobbyist needs and with straight-forward way of building and tune-up, 
> then we could alter the design pattern. The phase-noise and long term 
> stability issues is clear.

It doesn't look too difficult to crank something out within a rainy
weekend or two. But I am most likely underestimating the amount of work :-)
 
> Doing control loop using a phase-stepper is a little bit different, and 
> has a few minor design-challenges, but once mastered is essentially the 
> same. EFC or C-field control then becomes more an initial setup.

What makes the control loop different (beside that you control phase
and not frequency, and thus have to integrate)?


> An alternative approach divider wise is to use re-generative dividers.
> For Rick's approach there would be a number of these at the same 
> frequency (nominally), so the same design-pattern would apply. However, 
> that would only be meaningful if you need really need to keep the noise 
> down.

Yes, I thought about that as well, the problem here is that the low
noise mixers designs use transformers, which make everything bulky
and expensive (the usual suspects cost 2USD/piece and use about 1cm^2).
The one design that comes to mind that doesn't need transformers is
the tripple Gilbert-Cell design, but that might be higher in noise.
(Heck, i should just sit down and do some noise calculations)
Additionally, there is a need for relative steep filters for 667kHz


Attila Kinali


-- 
It is upon moral qualities that a society is ultimately founded. All 
the prosperity and technological sophistication in the world is of no 
use without that foundation.
 -- Miss Matheson, The Diamond Age, Neil Stephenson
___
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?

[Bob Camp]

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. 

> 
> 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 …. 

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.

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?

[Don Latham]

A friend and I have been messing with a DDS replacement for the VFO in older
radios. The odds runs between 5 and 5.5 MHz. There are some mixers that
generate the final LO frequency. We found many many birdies (caused by spurs
for the non-hams) over the tuning ranges.  We had to put in a lo-pass filter,
7stage commercial type to get rid of the birdies. But as I recall, there have
been several cautions on this list about filters causing temperature
dependence.  I haven't read the whole of this thread, so it may have already
been mentioned.
Merry Christmas,
Don

Magnus Danielson
> God kväll,
>
> On 12/09/2015 11:47 AM, Attila Kinali wrote:
>> God eftermiddag,
>>
>> On Tue, 8 Dec 2015 23:45:52 +0100
>> Magnus Danielson  wrote:
>>
>>
>>> If you would setup essentially a micro-stepper design, such as those
>>> being used for cesium and hydrogen masers, but maybe adapted to a
>>> hobbyist needs and with straight-forward way of building and tune-up,
>>> then we could alter the design pattern. The phase-noise and long term
>>> stability issues is clear.
>>
>> It doesn't look too difficult to crank something out within a rainy
>> weekend or two. But I am most likely underestimating the amount of work :-)
>
> Indeed. As any engineering time estimate, you need to multiply with pi.
> At work, we engineers divide our estimates with pi before giving it to
> the project managers, as they will multiply with pi before putting it
> into their time-plan. :)
>
>>> Doing control loop using a phase-stepper is a little bit different, and
>>> has a few minor design-challenges, but once mastered is essentially the
>>> same. EFC or C-field control then becomes more an initial setup.
>>
>> What makes the control loop different (beside that you control phase
>> and not frequency, and thus have to integrate)?
>
> Well, that is a little bit different right there. Depending on your
> setup, you might have to consider how phase-wrapping and similar
> saturations that happens over a long time. If you think about it, it's
> manageable.
> One useful trick is to let the phase-wrapping be that of the numeric
> wrapping, and then handle that case for time-stamps, so that the
> numerical extension becomes trivial. If you don't, you can get some very
> interesting problems.
>
>>> An alternative approach divider wise is to use re-generative dividers.
>>> For Rick's approach there would be a number of these at the same
>>> frequency (nominally), so the same design-pattern would apply. However,
>>> that would only be meaningful if you need really need to keep the noise
>>> down.
>>
>> Yes, I thought about that as well, the problem here is that the low
>> noise mixers designs use transformers, which make everything bulky
>> and expensive (the usual suspects cost 2USD/piece and use about 1cm^2).
>> The one design that comes to mind that doesn't need transformers is
>> the tripple Gilbert-Cell design, but that might be higher in noise.
>> (Heck, i should just sit down and do some noise calculations)
>> Additionally, there is a need for relative steep filters for 667kHz
>
> Indeed. For most uses, re-generative dividers will not be needed.
>
> I should do more experiments on that stuff.
>
> Cheers,
> Magnus
> ___
> 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.
>
>


-- 
Felix qui potuit rerum cognoscere causas.
Lucky is he who has been able to understand the causes of things.
Virgil
---
"Noli sinere nothos te opprimere"

Dr. Don Latham, AJ7LL
Six Mile Systems LLC, 17850 Six Mile Road
Huson, MT, 59846
mailing address:  POBox 404
Frenchtown MT 59834-0404

VOX 406-626-4304
CEL 406-241-5093
Skype: buffler2
www.lightningforensics.com
www.sixmilesystems.com


___
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?

[Magnus Danielson]

Don,

DDSes used to be very crude in terms of spurs. However, things have 
improved significantly and some of the chips you can get now is pretty 
impressive. It has even inspired some significantly different radio 
designs in the ham world.


As for filters, yes, they can vary delay with temperature, but the first 
rule of thumb is to keep filters low-Q (that is, both resonances/poles 
and zeros) because delay is proportional to Q. Then, well, temperature 
sensitive components, so choose those which is stable. Dielectras with 
low relative dielectric constants tend to be more stable, just as 
low-permability magnetics.


73 de SA0MAD Magnus

On 12/09/2015 10:44 PM, Don Latham wrote:

A friend and I have been messing with a DDS replacement for the VFO in older
radios. The odds runs between 5 and 5.5 MHz. There are some mixers that
generate the final LO frequency. We found many many birdies (caused by spurs
for the non-hams) over the tuning ranges.  We had to put in a lo-pass filter,
7stage commercial type to get rid of the birdies. But as I recall, there have
been several cautions on this list about filters causing temperature
dependence.  I haven't read the whole of this thread, so it may have already
been mentioned.
Merry Christmas,
Don

Magnus Danielson

God kväll,

On 12/09/2015 11:47 AM, Attila Kinali wrote:

God eftermiddag,

On Tue, 8 Dec 2015 23:45:52 +0100
Magnus Danielson  wrote:



If you would setup essentially a micro-stepper design, such as those
being used for cesium and hydrogen masers, but maybe adapted to a
hobbyist needs and with straight-forward way of building and tune-up,
then we could alter the design pattern. The phase-noise and long term
stability issues is clear.


It doesn't look too difficult to crank something out within a rainy
weekend or two. But I am most likely underestimating the amount of work :-)


Indeed. As any engineering time estimate, you need to multiply with pi.
At work, we engineers divide our estimates with pi before giving it to
the project managers, as they will multiply with pi before putting it
into their time-plan. :)


Doing control loop using a phase-stepper is a little bit different, and
has a few minor design-challenges, but once mastered is essentially the
same. EFC or C-field control then becomes more an initial setup.


What makes the control loop different (beside that you control phase
and not frequency, and thus have to integrate)?


Well, that is a little bit different right there. Depending on your
setup, you might have to consider how phase-wrapping and similar
saturations that happens over a long time. If you think about it, it's
manageable.
One useful trick is to let the phase-wrapping be that of the numeric
wrapping, and then handle that case for time-stamps, so that the
numerical extension becomes trivial. If you don't, you can get some very
interesting problems.


An alternative approach divider wise is to use re-generative dividers.
For Rick's approach there would be a number of these at the same
frequency (nominally), so the same design-pattern would apply. However,
that would only be meaningful if you need really need to keep the noise
down.


Yes, I thought about that as well, the problem here is that the low
noise mixers designs use transformers, which make everything bulky
and expensive (the usual suspects cost 2USD/piece and use about 1cm^2).
The one design that comes to mind that doesn't need transformers is
the tripple Gilbert-Cell design, but that might be higher in noise.
(Heck, i should just sit down and do some noise calculations)
Additionally, there is a need for relative steep filters for 667kHz


Indeed. For most uses, re-generative dividers will not be needed.

I should do more experiments on that stuff.

Cheers,
Magnus
___
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?

[Poul-Henning Kamp]

In message <566883d1.6050...@rubidium.dyndns.org>, Magnus Danielson writes:

>Indeed. As any engineering time estimate, you need to multiply with pi. 
>At work, we engineers divide our estimates with pi before giving it to 
>the project managers, as they will multiply with pi before putting it 
>into their time-plan. :)

Interesting.

In computing we make a random guess at the duration.

The manager is supposed to multiply the numeric part by two and
shift the unit to the next higher one.

-- 
Poul-Henning Kamp   | UNIX since Zilog Zeus 3.20
p...@freebsd.org | TCP/IP since RFC 956
FreeBSD committer   | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
___
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?

[Bob Camp]

Hi

> On Dec 9, 2015, at 9:34 AM, Jim Lux  wrote:
> 
> 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.

Regardless of the oven technology, the Q range on the crystal is pretty much 
fixed by the blank diameter and the blank design. The USO crystals are no 
higher or lower Q than those used in other oscillators. The first order impact 
of the tuning circuit is adding a bit of resistance in series with the crystal. 
The
oscillator circuit does this to a much greater extent. The active components in 
the oscillator stage have 1/F noise issues to a much greater extent than a 
narrow 
bandwidth tuning circuit. Can you crank the tuning range up far enough that a 
noisy enough reference becomes a problem? Sure you can. Can you tune the 
oscillator off far enough with a mechanical 
capacitor that it gets into trouble - yup. 

What works to your advantage with a fancy oscillator is that it has a very low 
aging rate. You also can afford to hand select parts. The *required* tuning 
range for an OCXO typically 
scales with it’s performance level. A low cost TCXO may indeed need a +/-10 ppm 
range to tune for 20 years. A high performance OCXO may be equally “happy” with 
a range 
below +/- 0.05 ppm.  

> 
> 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)

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



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)). 

Bob

> 
> 
> 
> 
>> 
>> Bob
>> 
>>> 
>>> 

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

[Magnus Danielson]

God kväll,

On 12/09/2015 11:47 AM, Attila Kinali wrote:

God eftermiddag,

On Tue, 8 Dec 2015 23:45:52 +0100
Magnus Danielson  wrote:



If you would setup essentially a micro-stepper design, such as those
being used for cesium and hydrogen masers, but maybe adapted to a
hobbyist needs and with straight-forward way of building and tune-up,
then we could alter the design pattern. The phase-noise and long term
stability issues is clear.


It doesn't look too difficult to crank something out within a rainy
weekend or two. But I am most likely underestimating the amount of work :-)


Indeed. As any engineering time estimate, you need to multiply with pi. 
At work, we engineers divide our estimates with pi before giving it to 
the project managers, as they will multiply with pi before putting it 
into their time-plan. :)



Doing control loop using a phase-stepper is a little bit different, and
has a few minor design-challenges, but once mastered is essentially the
same. EFC or C-field control then becomes more an initial setup.


What makes the control loop different (beside that you control phase
and not frequency, and thus have to integrate)?


Well, that is a little bit different right there. Depending on your 
setup, you might have to consider how phase-wrapping and similar 
saturations that happens over a long time. If you think about it, it's 
manageable.
One useful trick is to let the phase-wrapping be that of the numeric 
wrapping, and then handle that case for time-stamps, so that the 
numerical extension becomes trivial. If you don't, you can get some very 
interesting problems.



An alternative approach divider wise is to use re-generative dividers.
For Rick's approach there would be a number of these at the same
frequency (nominally), so the same design-pattern would apply. However,
that would only be meaningful if you need really need to keep the noise
down.


Yes, I thought about that as well, the problem here is that the low
noise mixers designs use transformers, which make everything bulky
and expensive (the usual suspects cost 2USD/piece and use about 1cm^2).
The one design that comes to mind that doesn't need transformers is
the tripple Gilbert-Cell design, but that might be higher in noise.
(Heck, i should just sit down and do some noise calculations)
Additionally, there is a need for relative steep filters for 667kHz


Indeed. For most uses, re-generative dividers will not be needed.

I should do more experiments on that stuff.

Cheers,
Magnus
___
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?

[Magnus Danielson]

Poul-Henning,

On 12/09/2015 09:24 AM, Poul-Henning Kamp wrote:


In message <56675da0.4050...@rubidium.dyndns.org>, Magnus Danielson writes:


So, what did I miss? Why do people use DAC-EFC control instead of
the DDS scheme?


The main reason I would say is habbits, people have habbits and stick to
them.


And I think right after that comes the problem of spurs.

Controlling DDS spurs is mostly a solved problem, when you get to choose
your input and output frequencies for that purpose.

But as "afterburner" for an OCXO, there is no restriction on the two
frequencies, and they are almost guaranteed to be close-in multiples,
so coping with the spurs becomes non-trivial.

That basically means that you need 50% higher DDS and DAC resolution
than everybody else in the market an suddenly it isn't cheap any more.




Well, consider that using a circuit similar to the one Rick has in his 
article, then frequency deviation and spurs will be significantly scaled 
down. For a good crystal, you will maybe have a handfull of ppms, let's 
say +/- 10 ppm range. A scale-down of 1000-10 would be worth 
considering. That helps with DDS spurs.


Cheers,
Magnus
___
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?

[Bert Kehren via time-nuts]

Some of the suggestions offered here are incorporated in the FE  
205/405/505. We have done extensive work using a 405 with very good results. 
Tom  did 
some tests on the 406 and we traced issues back to the temperature  control. 
We disabled it and continue to have an excellent GPSDO. Tests are on  hold 
do to my move and the realization that we need to improve our measuring  
capabilities, an ongoing process nearing completion.
This concept allows FE to use a wide range of OCXO's with no selection or  
tuning resulting in lower XTAL cost but also lower phase and AV.
Bert Kehren
 
 
 
In a message dated 12/9/2015 10:00:19 A.M. Eastern Standard Time,  
jim...@earthlink.net writes:

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.
___
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] GPSDO and oscillator steering - EFC vs DDS schemes?

[Attila Kinali]

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?

Given that Archita Hati et al. were getting very low noise numbers on
their RF signal generation scheme using dividers [2], I don't think that
the noise of the mixers would be the limiting factor here, but rather
that the phase noise should be still dominated by the 10MHz oscillator.

My guestimate is that something like this would cost approximately 5USD
per divider stage, plus 20 USD for the DDS plus initial mixer. The only
problem would be to get a narrow band 10.0MHz filter (I couldn't find
one within 5 minutes of googling). 5 stages should cost around 50-70USD)
and will give a resolution better than 5uHz (100MHz DDS with 24bit)
down to 20pHz range (100MHz DDS with 32bit), which are 1:5e-13 
and 1:2e-15 respectively. 

Compared to an EFC system that costs somewhere in the range of 10-50USD
and gives a resolution of something between 1:5e-12 (0.3ppm tuning range,
16bit DAC) and 1:1e-13 (10^-7 tuning range and 20bit DAC). Especially the
20bit DAC version gives a lot of electrical problems, starting from the
stability of the reference, leakage current trough various components and the 
PCB etc pp, while the DDS scheme, as a "digital" scheme is virtually free
of those.

So, the DDS scheme is easier to reproduce, more stable over time and
costs only slightly more (unless you try to use an LTZ1000 as reference,
then the reference alone costs more then the whole DDS scheme).

So, what did I miss? Why do people use DAC-EFC control instead of
the DDS scheme?

Attila Kinali

[1] "A narrow band high-resolution synthesizer using a direct digital
synthesiser followed by repeated dividing and mixing", Richard Karlquist, 1995
http://www.karlquist.com/FCS95.pdf

[2] "State-of-the-Art RF Signal Generation From Optical Frequency Division".
by Hati, Nelson, Barnes, Lirette, Fortier, Quinlan, DeSalvo, Ludlow, Diddams,
Howe, 2013
http://tf.boulder.nist.gov/general/pdf/2646.pdf

-- 
It is upon moral qualities that a society is ultimately founded. All 
the prosperity and technological sophistication in the world is of no 
use without that foundation.
 -- Miss Matheson, The Diamond Age, Neil Stephenson
___
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?

[Azelio Boriani]

Something like good_100MHz_OCXO+DDS => same as a BVA?

On Tue, Dec 8, 2015 at 5:32 PM, 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?
>
> Given that Archita Hati et al. were getting very low noise numbers on
> their RF signal generation scheme using dividers [2], I don't think that
> the noise of the mixers would be the limiting factor here, but rather
> that the phase noise should be still dominated by the 10MHz oscillator.
>
> My guestimate is that something like this would cost approximately 5USD
> per divider stage, plus 20 USD for the DDS plus initial mixer. The only
> problem would be to get a narrow band 10.0MHz filter (I couldn't find
> one within 5 minutes of googling). 5 stages should cost around 50-70USD)
> and will give a resolution better than 5uHz (100MHz DDS with 24bit)
> down to 20pHz range (100MHz DDS with 32bit), which are 1:5e-13
> and 1:2e-15 respectively.
>
> Compared to an EFC system that costs somewhere in the range of 10-50USD
> and gives a resolution of something between 1:5e-12 (0.3ppm tuning range,
> 16bit DAC) and 1:1e-13 (10^-7 tuning range and 20bit DAC). Especially the
> 20bit DAC version gives a lot of electrical problems, starting from the
> stability of the reference, leakage current trough various components and the 
> PCB etc pp, while the DDS scheme, as a "digital" scheme is virtually free
> of those.
>
> So, the DDS scheme is easier to reproduce, more stable over time and
> costs only slightly more (unless you try to use an LTZ1000 as reference,
> then the reference alone costs more then the whole DDS scheme).
>
> So, what did I miss? Why do people use DAC-EFC control instead of
> the DDS scheme?
>
> Attila Kinali
>
> [1] "A narrow band high-resolution synthesizer using a direct digital
> synthesiser followed by repeated dividing and mixing", Richard Karlquist, 1995
> http://www.karlquist.com/FCS95.pdf
>
> [2] "State-of-the-Art RF Signal Generation From Optical Frequency Division".
> by Hati, Nelson, Barnes, Lirette, Fortier, Quinlan, DeSalvo, Ludlow, Diddams,
> Howe, 2013
> http://tf.boulder.nist.gov/general/pdf/2646.pdf
>
> --
> It is upon moral qualities that a society is ultimately founded. All
> the prosperity and technological sophistication in the world is of no
> use without that foundation.
>  -- Miss Matheson, The Diamond Age, Neil Stephenson
> ___
> 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?

[Azelio Boriani]

Given that until now good (maximum stability) OCXO are much less than
100MHz, from the OCXO we exploit its high stability and we impose
accuracy from a coordinated source: the OCXO+EFC method uses the
built-in stability and disciplines the accuracy.
The DDS method virtually can start from any oscillator, apply a
suitable correction function giving the same result, transferring the
hardware characteristic of a BVA (for example) into the driving
function.
Can a DDS be driven with the speed necessary to correct the output so
that it results in the same stability as a BVA, starting from a given
unstable oscillator?
Or, how much unstable can be the 100MHz starting oscillator so that I
can obtain after the DDS+suitable_driving_function  the same final
stability as an ordinary 10MHz OCXO?

On Tue, Dec 8, 2015 at 6:44 PM, Azelio Boriani  wrote:
> Something like good_100MHz_OCXO+DDS => same as a BVA?
>
> On Tue, Dec 8, 2015 at 5:32 PM, 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?
>>
>> Given that Archita Hati et al. were getting very low noise numbers on
>> their RF signal generation scheme using dividers [2], I don't think that
>> the noise of the mixers would be the limiting factor here, but rather
>> that the phase noise should be still dominated by the 10MHz oscillator.
>>
>> My guestimate is that something like this would cost approximately 5USD
>> per divider stage, plus 20 USD for the DDS plus initial mixer. The only
>> problem would be to get a narrow band 10.0MHz filter (I couldn't find
>> one within 5 minutes of googling). 5 stages should cost around 50-70USD)
>> and will give a resolution better than 5uHz (100MHz DDS with 24bit)
>> down to 20pHz range (100MHz DDS with 32bit), which are 1:5e-13
>> and 1:2e-15 respectively.
>>
>> Compared to an EFC system that costs somewhere in the range of 10-50USD
>> and gives a resolution of something between 1:5e-12 (0.3ppm tuning range,
>> 16bit DAC) and 1:1e-13 (10^-7 tuning range and 20bit DAC). Especially the
>> 20bit DAC version gives a lot of electrical problems, starting from the
>> stability of the reference, leakage current trough various components and 
>> the PCB etc pp, while the DDS scheme, as a "digital" scheme is virtually free
>> of those.
>>
>> So, the DDS scheme is easier to reproduce, more stable over time and
>> costs only slightly more (unless you try to use an LTZ1000 as reference,
>> then the reference alone costs more then the whole DDS scheme).
>>
>> So, what did I miss? Why do people use DAC-EFC control instead of
>> the DDS scheme?
>>
>> Attila Kinali
>>
>> [1] "A narrow band high-resolution synthesizer using a direct digital
>> synthesiser followed by repeated dividing and mixing", Richard Karlquist, 
>> 1995
>> http://www.karlquist.com/FCS95.pdf
>>
>> [2] "State-of-the-Art RF Signal Generation From Optical Frequency Division".
>> by Hati, Nelson, Barnes, Lirette, Fortier, Quinlan, DeSalvo, Ludlow, Diddams,
>> Howe, 2013
>> http://tf.boulder.nist.gov/general/pdf/2646.pdf
>>
>> --
>> It is upon moral qualities that a society is ultimately founded. All
>> the prosperity and technological sophistication in the world is of no
>> use without that foundation.
>>  -- Miss Matheson, The Diamond Age, Neil Stephenson
>> ___
>> 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 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] GPSDO and oscillator steering - EFC vs DDS schemes?

[Attila Kinali]

On Tue, 8 Dec 2015 17:32:17 +0100
Attila Kinali  wrote:

> The only
> problem would be to get a narrow band 10.0MHz filter (I couldn't find
> one within 5 minutes of googling). 5 stages should cost around 50-70USD)

Correction: I should have looked at "Resonators" instead of "Filters"
10MHz ceramic resonators are readily available and have Q factors in
the range of 900 to 4000 which results in a BW between 11kHz and 2.5kHz.

Attila Kinali
-- 
It is upon moral qualities that a society is ultimately founded. All 
the prosperity and technological sophistication in the world is of no 
use without that foundation.
 -- Miss Matheson, The Diamond Age, Neil Stephenson
___
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?

[Magnus Danielson]

Jim,

On 12/08/2015 07:09 PM, Jim Lux wrote:

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.


Depends on what you do. I think the first reaction would be "why make it 
complex?" as the DAC and EFC is relatively simple and "understood".
Second arguments would be component count and board space. The uncounted 
case would be in the herritage department, which would prohibit the 
thought to even be considered in the first place.


However, the modern DDS chips makes it more and more useful for many cases.


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)


Consider that their use here would mean that such effects would be mixed 
down considerably. Doing a good DDS steering algorithm with plenty of 
updates would help to keep things smooth.


Personally, I use a combination of the techniques, you can play 
interesting tricks with DDSes if you build them yourself. I then lock up 
oscillators when that is the best approach, for say phase-noise.


Cheers,
Magnus
___
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?

[Magnus Danielson]

God kväll Attila,

On 12/08/2015 05:32 PM, 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?

Given that Archita Hati et al. were getting very low noise numbers on
their RF signal generation scheme using dividers [2], I don't think that
the noise of the mixers would be the limiting factor here, but rather
that the phase noise should be still dominated by the 10MHz oscillator.

My guestimate is that something like this would cost approximately 5USD
per divider stage, plus 20 USD for the DDS plus initial mixer. The only
problem would be to get a narrow band 10.0MHz filter (I couldn't find
one within 5 minutes of googling). 5 stages should cost around 50-70USD)
and will give a resolution better than 5uHz (100MHz DDS with 24bit)
down to 20pHz range (100MHz DDS with 32bit), which are 1:5e-13
and 1:2e-15 respectively.

Compared to an EFC system that costs somewhere in the range of 10-50USD
and gives a resolution of something between 1:5e-12 (0.3ppm tuning range,
16bit DAC) and 1:1e-13 (10^-7 tuning range and 20bit DAC). Especially the
20bit DAC version gives a lot of electrical problems, starting from the
stability of the reference, leakage current trough various components and the PCB etc pp, 
while the DDS scheme, as a "digital" scheme is virtually free
of those.

So, the DDS scheme is easier to reproduce, more stable over time and
costs only slightly more (unless you try to use an LTZ1000 as reference,
then the reference alone costs more then the whole DDS scheme).

So, what did I miss? Why do people use DAC-EFC control instead of
the DDS scheme?


The main reason I would say is habbits, people have habbits and stick to 
them. Many follow the design patterns of others, often as found in 
books, as tought in university, as inherited within a company, as found 
on Internet, as design by fellow hams, whatever.


If you would setup essentially a micro-stepper design, such as those 
being used for cesium and hydrogen masers, but maybe adapted to a 
hobbyist needs and with straight-forward way of building and tune-up, 
then we could alter the design pattern. The phase-noise and long term 
stability issues is clear.


Doing control loop using a phase-stepper is a little bit different, and 
has a few minor design-challenges, but once mastered is essentially the 
same. EFC or C-field control then becomes more an initial setup.


An alternative approach divider wise is to use re-generative dividers.
For Rick's approach there would be a number of these at the same 
frequency (nominally), so the same design-pattern would apply. However, 
that would only be meaningful if you need really need to keep the noise 
down.


Cheers,
Magnus
___
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?

[Attila Kinali]

Ciao Azelio,

On Tue, 8 Dec 2015 19:36:37 +0100
Azelio Boriani  wrote:

> Given that until now good (maximum stability) OCXO are much less than
> 100MHz, from the OCXO we exploit its high stability and we impose
> accuracy from a coordinated source: the OCXO+EFC method uses the
> built-in stability and disciplines the accuracy.
> The DDS method virtually can start from any oscillator, apply a
> suitable correction function giving the same result, transferring the
> hardware characteristic of a BVA (for example) into the driving
> function.
> Can a DDS be driven with the speed necessary to correct the output so
> that it results in the same stability as a BVA, starting from a given
> unstable oscillator?
> Or, how much unstable can be the 100MHz starting oscillator so that I
> can obtain after the DDS+suitable_driving_function  the same final
> stability as an ordinary 10MHz OCXO?

No, no matter what kind of steering you use, you will not make the
clock more stable. The BVA is a stable oscillator by and in itself.
Even if you do not lock it to anything, it rivals several comercial
atomic frequency standards in stability up to several dozen to hundreds
of seconds. If you start with an "normal" OCXO, you will never get that
stability, no matter how hard you try to compensate for all environmental
factors. But that's not the point of my question.

The point of my question was, why people use EFC control with all its
problems and its "low" resolution, when there is better methods known.
Or, whether I missed some important point that is not obvious from reading
an otherwise good paper.


Attila Kinali
-- 
Reading can seriously damage your ignorance.
-- unknown
___
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?

[Magnus Danielson]

Azelio,

You don't get BVA performance easily out of an oscillator being 
significantly less stable than a BVA. Some environmental aspects you can 
dampen, some you can compensate, but then as you hit the fundamental 
noise processes of the oscillator. Knowing how systematics affects the 
oscillator without comparing to a much stabler source, does help you 
only to some degree when you loose that source and need to free-wheel 
without the correction input. If you have a more stable source, use that 
directly. There are oscillators which may put BVAs to the test for noise 
and systematics.


Rather than trying to do that, I think the alternative approach should 
be viewed as an interesting concept where you overcome several of the 
problems with the DAC and EFC approach, such as the DAC resolution (yes, 
I've fought that one), the DAC reference stability (as you open the 
loop, it becomes an issue), the EFC steering curve (it's not very 
linear) and range, the EFC temperature sensitivity (which becomes 
uncompensated as you go open-loop, but also shows up as subtle 
state-changes in closed-loop).


Another benefit of not steering the oscillator directly is when you 
build an ensamble of them, then you can more precisely predict their 
futute behavior when you don't steer them, as you reduce those unknown 
non-linear errors.


So, in short, I think you would fool yourself in believing you can 
significantly alter the noise process performance of your oscillator, 
see it in an improvement in how you can linearly compensate the offset 
of your oscillator, and with that possibly focus more on trimming some 
of the systematic corrections better. The noise of the core oscillator 
will keep dominate, it's just the reducing of the syntesis chain we keep 
down.


Cheers,
Magnus

On 12/08/2015 07:36 PM, Azelio Boriani wrote:

Given that until now good (maximum stability) OCXO are much less than
100MHz, from the OCXO we exploit its high stability and we impose
accuracy from a coordinated source: the OCXO+EFC method uses the
built-in stability and disciplines the accuracy.
The DDS method virtually can start from any oscillator, apply a
suitable correction function giving the same result, transferring the
hardware characteristic of a BVA (for example) into the driving
function.
Can a DDS be driven with the speed necessary to correct the output so
that it results in the same stability as a BVA, starting from a given
unstable oscillator?
Or, how much unstable can be the 100MHz starting oscillator so that I
can obtain after the DDS+suitable_driving_function  the same final
stability as an ordinary 10MHz OCXO?

On Tue, Dec 8, 2015 at 6:44 PM, Azelio Boriani  wrote:

Something like good_100MHz_OCXO+DDS => same as a BVA?

On Tue, Dec 8, 2015 at 5:32 PM, 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?

Given that Archita Hati et al. were getting very low noise numbers on
their RF signal generation scheme using dividers [2], I don't think that
the noise of the mixers would be the limiting factor here, but rather
that the phase noise should be still dominated by the 10MHz oscillator.

My guestimate is that something like this would cost approximately 5USD
per divider stage, plus 20 USD for the DDS plus initial mixer. The only
problem would be to get a narrow band 10.0MHz filter (I couldn't find
one within 5 minutes of googling). 5 stages should cost around 50-70USD)
and will give a resolution better than 5uHz (100MHz DDS with 24bit)
down to 20pHz range (100MHz DDS with 32bit), which are 1:5e-13
and 1:2e-15 respectively.

Compared to an EFC system that costs somewhere in the range of 10-50USD
and gives a resolution of something between 1:5e-12 (0.3ppm tuning range,
16bit DAC) and 1:1e-13 (10^-7 tuning range and 20bit DAC). Especially the
20bit DAC version gives a lot of electrical problems, starting from the
stability of the reference, leakage current trough various components and the PCB etc pp, 
while the DDS scheme, as a "digital" scheme is virtually free
of those.

So, the DDS scheme is easier to reproduce, more stable over time and
costs only slightly more (unless you try to use an LTZ1000 as reference,
then the reference alone costs more then the whole DDS scheme).

So, what did I miss? Why do people use 

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

[Bob Camp]

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

> On Dec 8, 2015, at 11:32 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?
> 
> Given that Archita Hati et al. were getting very low noise numbers on
> their RF signal generation scheme using dividers [2], I don't think that
> the noise of the mixers would be the limiting factor here, but rather
> that the phase noise should be still dominated by the 10MHz oscillator.
> 
> My guestimate is that something like this would cost approximately 5USD
> per divider stage, plus 20 USD for the DDS plus initial mixer. The only
> problem would be to get a narrow band 10.0MHz filter (I couldn't find
> one within 5 minutes of googling). 5 stages should cost around 50-70USD)
> and will give a resolution better than 5uHz (100MHz DDS with 24bit)
> down to 20pHz range (100MHz DDS with 32bit), which are 1:5e-13 
> and 1:2e-15 respectively. 
> 
> Compared to an EFC system that costs somewhere in the range of 10-50USD
> and gives a resolution of something between 1:5e-12 (0.3ppm tuning range,
> 16bit DAC) and 1:1e-13 (10^-7 tuning range and 20bit DAC). Especially the
> 20bit DAC version gives a lot of electrical problems, starting from the
> stability of the reference, leakage current trough various components and the 
> PCB etc pp, while the DDS scheme, as a "digital" scheme is virtually free
> of those.
> 
> So, the DDS scheme is easier to reproduce, more stable over time and
> costs only slightly more (unless you try to use an LTZ1000 as reference,
> then the reference alone costs more then the whole DDS scheme).
> 
> So, what did I miss? Why do people use DAC-EFC control instead of
> the DDS scheme?
> 
>   Attila Kinali
> 
> [1] "A narrow band high-resolution synthesizer using a direct digital
> synthesiser followed by repeated dividing and mixing", Richard Karlquist, 1995
> http://www.karlquist.com/FCS95.pdf
> 
> [2] "State-of-the-Art RF Signal Generation From Optical Frequency Division".
> by Hati, Nelson, Barnes, Lirette, Fortier, Quinlan, DeSalvo, Ludlow, Diddams,
> Howe, 2013
> http://tf.boulder.nist.gov/general/pdf/2646.pdf
> 
> -- 
> It is upon moral qualities that a society is ultimately founded. All 
> the prosperity and technological sophistication in the world is of no 
> use without that foundation.
> -- Miss Matheson, The Diamond Age, Neil Stephenson
> ___
> 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?

[Attila Kinali]

On Tue, 8 Dec 2015 10:09:44 -0800
Jim Lux  wrote:

> > 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.

That might be an explanation.

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

Well, Rick's paper is already 20 years old. There has been lots of
discussion of DDS in the radio community and I think that community
has a lot of knowledge of the idiosyncracies of DDS. I'm not sure
how it works for the  time and frequency control community, though.

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

Compared to the 1-5W steady state of an OCXO?
But yes, the circuit would probably need something in the range of 2-3W.
 
> 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 might be naive, but I would have used an small FPGA with EEPROM cells
(like the Lattice ICE or the Altera MAX) and build an DDS by hand that
is tailored for exactly that purpose (to get minimal spurs), add an
10bit DAC, combine everything with an 60-100MHz VCXO that is PLL-locked to
the 10MHz input. Well, actually the VCXO might not even be necessary,
if the FPGA internal PLL is not too noisy, there is some heavy "damping"
of the noise through the divider stages anyways.


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


Unless you are a hobbyist.


Attila Kinali

-- 
Reading can seriously damage your ignorance.
-- unknown
___
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?

[Alexander Pummer]

Bar Giora Goldberg wrote a book on DDS Digital Frequency Synthesis 
Demystified 
see 
here: 
http://www.amazon.com/Frequency-Synthesis-Demystified-Bar-Giora-Goldberg/dp/1878707477
[· In-depth coverage of modern digital implementations of frequency 
synthesis architectures

· Numerous design examples drawn from actual engineering projects
· The accompanying CD includes digital frequency synthesis design tools 
and an electronic version of the book]
Hi was the chef technologist of  Sciteq, that company made the first 
usable DDS devices.

73
KJ6UHN
Alex


On 12/8/2015 4:00 PM, Attila Kinali wrote:

On Tue, 8 Dec 2015 10:09:44 -0800
Jim Lux  wrote:


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.

That might be an explanation.


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

Well, Rick's paper is already 20 years old. There has been lots of
discussion of DDS in the radio community and I think that community
has a lot of knowledge of the idiosyncracies of DDS. I'm not sure
how it works for the  time and frequency control community, though.


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

Compared to the 1-5W steady state of an OCXO?
But yes, the circuit would probably need something in the range of 2-3W.
  

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 might be naive, but I would have used an small FPGA with EEPROM cells
(like the Lattice ICE or the Altera MAX) and build an DDS by hand that
is tailored for exactly that purpose (to get minimal spurs), add an
10bit DAC, combine everything with an 60-100MHz VCXO that is PLL-locked to
the 10MHz input. Well, actually the VCXO might not even be necessary,
if the FPGA internal PLL is not too noisy, there is some heavy "damping"
of the noise through the divider stages anyways.



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


Unless you are a hobbyist.


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.