[time-nuts] Advantages & Disadvantages of the TPLL Method

Steve Rooke sar10538 at gmail.com
Mon Jun 21 08:34:41 UTC 2010


Bob,

On 21 June 2010 02:46, Robert Benward <rbenward at verizon.net> wrote:
> Steve,
> I am a professional engineer, but in this arena I am an amateur.  That is
> why I'm asking the questions, not to put down, but to understand some of the
> claims made.  And as I said in one of my previous emails, I've seen amateurs
> run circles around the professionals, and those professional admitting utter
> astonishment at those amateur accomplishments (this is in the area of
> amateur astrophotography).

You have not understood the point I was trying to make as you are
still trying to approach this from a professional engineer angle. Yes,
we are all trying to understand this implementation of the TPLL method
but as you have just seen, I just chose my words very carefully there.
I used the word "implementation" as it's not a complete new design,
this TPLL method has been well known and accepted for decades, only a
small part of it has been implemented in a slightly different way. To
put it simply, the VFC and Counter has been replaced by a system which
integrates the EFC voltage over the Tau0 measurement time by
oversampling. In this way, the oversampling performs the same function
as the VCF and the Counter but in an easier way. Perhaps you do not
understand how the original design of the NIST TPLL works and you
should refress yourself with that. Look especially closely as to why
they chose to use the EFC and the counter in there as this is very
important.

> What I have heard throughout this thread is a lot of bashing of those asking
> the questions, surfacing as derogatory and berating comments on other's
> understanding.  I have also heard much claims to a certain procedure without
> one iota of numerical mumbo-jumbo to back it up.

I think you will find that this is quite historical now as the
original reaction to this proposal was made in a very sceptical and
negative manner thereby showing the naysayers as very arrogant. This
went on for a long time and it's hardly surprising that this whole
affair has denigrated into this but I know that Warren is not to blame
for this, it started elsewheres.

> The issue here is an inability to describe a simple claim.  Pete has
> attempted to put things in simple numbers, and I see where he is going, and
> I concur with some of his calculations.  If one can not describe what
> appears to be a simple procedure, then I must question the basic
> understanding behind the explanation.  If you make a wild claim, and then
> you can't even get the bullet on the paper, then I must question the
> shooter's understanding.

Well, lets take the bullet and paper analogy first shall we. Warren
has shown that he can get the bullet right in the middle of the paper
even though he may be unable to describe it in a way that you require.
By demanding that this be first described in full scientific detail
before you will even look at it, smacks at throwing the baby out with
the bath water. I have only limited college education in electrical
engineering and I'm not able to put the math around this but I do
believe that I understand it logically and see how/why it works. I
have tried to express this previously in the only way I can that that
was plain old English but it seems that for some of you this is not
acceptable. Warren has been pushed up against the wall by some of you
and pressed to justify his implementation and when he tried to do that
in the only language he has that makes sense to him all you do is to
pick up on the semantics instead of trying to understand what he
means. Maybe he makes wild claims to the pedantic in you professional
engineers but why don't we try to find the language that is what
Warren really means.

> I guess I am not comfortable with the use of femtoseconds to describe
> frequency accuracy.  Technically, a locked PLL is at the exact frequency as
> the reference, as measured in the long term.  The phase between the two may
> not be at zero, that depends on the type of phase detector and the DC
> offsets in the system.  On the short term, phase noise of the reference will
> cause the loop to generate error terms which will change the phase of the
> DUT.  Oscillators are also specified using phase noise, e.g. 135dB down @
> 100Hz.  That specifies how much energy is not in the bandwidth of the
> carrier.  It also implies the phase is constantly changing!  If the phase is
> changing, the error term is changing, and so forth and so on.....Your
> measurement can only be as good as your reference oscillator.  A DVM can
> only average this error, it can't give you the instantaneous value of the
> peak deviation of the error signal, which is what you would need to claim fs
> cycle to cycle timing.  Fs units are appropriate for cycle to cycle
> variation, not long term or multicycle assements.  Even the best HP DVM is
> only good to 3ppm on the 100mV scale and the shortest reading is 167us.
> That's 10 orders of magnitude greater that the deviation you are trying to
> measure.  If you average the mixer output, you can no longer claim fs
> timing.  What you can claim is a long term frequency stability in ppm.

OK, so your smart, well done, congratulations! Let's forget about
femtoseconds and shove them firmly in the bin, so no one talk about
them again, OK. Now, if you look at Warren's block diagram you will
see that the reference (device given) and DUT oscillators are feed via
pads (values given) into a mixer. The output of this mixer is fed via
a 100kHz LPF to remove the sum product and then via a simple op-amp
with 100 gain. The rest of the loop is just a way to add the necessary
DC offset such that when the reference and DUT oscillators match, IE.
the average DC out of the mixer is 0V, the EFC on the reference
oscillator is biased correctly. So you can look up the output of a
10811 ocxo and also see the specs to work out the EFC F/V and I'm sure
this information is probably very easily available via the resources
in this list. To be honest, I expect that the EFC voltage deviation is
going to be small relative to the full range. You know that there is a
100kHz R/C filter in the loop so perhaps you can scribble some numbers
down on a piece of paper with your pencil. If you can't, what more do
you need to be able to do this, thanks for your help.

> This is my simple understanding of phase detectors and mixers.  You might
> get there by dividing down a bunch of numbers but I don't think the method
> supports the claim (of fs timing).

Let's kill the fs thing please!

Steve

> Bob
>
>
>
> ----- Original Message -----
>  From: Steve Rooke
>  To: Discussion of precise time and frequency measurement
>  Sent: Sunday, June 20, 2010 2:00 AM
>  Subject: Re: [time-nuts] Advantages & Disadvantages of the TPLL Method
>
>
>  Bob,
>
>  Can I answer this one.
>
>  On 20 June 2010 04:36, Robert Benward <rbenward at verizon.net> wrote:
>  > Warren,
>  > I was responding to ke5fx comment "using a 12-bit, 480-Hz serial DAQ in
>  > place of the voltage-to-frequency converter in the diagram above". A DAQ
>  > is a multifaceted data acquisition system, where as in your annotated
>  > diagram you showed an ADC.
>
>  The DAQ that Warren is referring to to has a 12bit ADC input capable
>  of performing up to 480 samples per second.
>
>  > I understand it's analog, but you said: "Say you have a nice logic gate
> with
>  > 1 ns delay" . So back to the analog loop, do you have an analysis that
> gets
>  > you from EFC to femtosecond stability? PLLs are notorious for phase
> noise,
>  > the phase noise actually representing the error term that brings the
> loop
>  > back into lock.
>
>  I personally think the 1fs issue has become way out of hand and people
>  are now focussing on that instead of the big picture. Whilst I
>  understand that the professional engineers on this list wish to pounce
>  on every t that is not crossed, every i that is not dotted, and
>  requiring a complete mathematical breakdown of everything, it is not
>  going to happen here. If those professional engineers would like to
>  assist with the process of understanding and documenting this idea in
>  a way that pulls their chain, that would be great, but if it's down to
>  pointing the finger at the amateur engineers and laughing, then
>  perhaps they need more education in etiquette. Remember the golden
>  rule, do unto others as you would wish to be done.
>
>  Sure, some of us do not have the correct technical engineering banter,
>  so when we call the World a sort of round ball shape, please don't
>  play deaf until we say it's an oblate spheroid. Try to help us
>  communicate with you, we are trying to describe things in the best way
>  we can and we have something useful to contribute, IE. just take
>  Warren's TPLL implementation which seems to be producing good results.
>  So why don't we try to understand exactly how it is doing this instead
>  of ripping it apart and saying you shouldn't do it that way, you have
>  to do it this way. Remember that geezer who invented the lightbulb, he
>  didn't work it all out mathematically on paper before he chose
>  tungsten, no he did it experimentally and everyone seems to think
>  highly of him.
>
>  Steve
>  >
>  >
>  > For your second email:
>  >
>  >>You are now averaging the "repeatable" jitter? YES
>  > I was not questioning the procedure, I was questioning the conclusion;
>  >
>  >>Are you using a digital phase detector or a mixer as shown? Analog
>  >>Phase detector
>  > Why the digital analogy if it's all analog?
>  >
>  >>Do you have an analysis of the loop sensitivity/resolution? No
>  >>analysis, No limit it is analog
>  > I don't agree with you about the limit, and without an analysis or even
> a
>  > simple calculation, how do arrive at femtosecond lock? if there is no
>  > limit, why not a hundred times less?
>  >
>  >>Why do you say the results are repeatable in the short term vs the long
>  >>term? Long term includes other factors such as non random drift, not
>  >>just "random Noise"
>  > Maybe so, but using the "short term" , is not a license to better jitter
>  > figures by a factor of 100. Since you are not using digital, I don't
> know
>  > where this example came from or why it is relevant.
>  >
>  >> Is there not a lower limit to how much you can average? Depends or
>  >> everything, but not up to > 1 sec of averaging when the conditions are
>  >> made right
>  > I don't understand how you arrive at this conclusion
>  >
>  >
>  > For your last email:
>  > What attracted me to the TPLL question now was that you comment that you
> are
>  > maintaining a femtosecond lock. Please don't dumb it down for me. I may
> not
>  > understand all the statistical stuff, but I can understand an analysis.
>  >
>  >
>  > Bob
>  >
>  >
>  >
>  > ----- Original Message -----
>  > From: "WarrenS" <warrensjmail-one at yahoo.com>
>  > To: "Discussion of precise time and frequency measurement"
>  > <time-nuts at febo.com>
>  > Sent: Saturday, June 19, 2010 3:27 AM
>  > Subject: Re: [time-nuts] Advantages & Disadvantages of the TPLL Method
>  >
>  >
>  >> Bob
>  >>
>  >>>> Don't know if I can explain it to you, I'm not so good at explaining,
>  >>>> I'll give it *ONE* try.
>  >>>> Example with some random picked numbers (JUST TO SHOW THE MAIN
> POINTS).
>  >>
>  >> I tried,
>  >> All information and test that are available on the TPLL is on JOHN'S
>  >> KE5FX
>  >> site or in past postings.
>  >> http://www.thegleam.com/ke5fx/tpll.htm
>  >>
>  >> One other thing I may not of made clear, The analog averaging thing
> does
>  >> not
>  >> help at low freq like at 1 PPS
>  >> The TPLL works great because it is at a high freq like 5 or 10 MHz.
>  >> DAQ == DataQ == ADC
>  >>
>  >>> I don't think 10ps is achievable under any dynamic conditions IMHO
>  >> OK, I don't really care, use whatever number you want, you'll still end
> up
>  >> below the Ref osc noise.
>  >> but
>  >> You may be surprised then by what the single shot "Aperture
> uncertainty"
>  >> specs are for the kind of devices that really care about this sort of
>  >> thing.
>  >> But then none of that really maters AT ALL,
>  >> because there is NO Digital anything in the simple TPLL before the ADC
>  >> where
>  >> a 10 Hz device would work fine for most.
>  >> I just gave you an example to try and answer your question on digital
>  >> logic
>  >> which was:
>  >>> How do you do fs when most digital logic has jitter several of orders
> of
>  >>> magnitude greater?
>  >>
>  >> ws
>  >>
>  >> ***************************
>  >> [time-nuts] Advantages & Disadvantages of the TPLL Method
>  >> Robert Benward rbenward at verizon.net
>  >> Sat Jun 19 03:18:05 UTC 2010
>  >>
>  >> Warren,
>  >> Is there not a lower limit to how much you can average? Yes, it's the
>  >> sqrt
>  >> of the number of samples, but doesn't noise,
>  >> hardware, and other perturbations limit the usefulness of this method?
>  >>
>  >>> Then one can get repeatable results say 100 times better from cycle to
>  >>> cycle in the short term.
>  >>> so down to 10ps repeatable.
>  >>
>  >> Why do you say the results are repeatable in the short term vs the long
>  >> term? Isn't what you defined above
>  >> (repeatability) the opposite of jitter? Jitter I thought was cycle to
>  >> cycle
>  >> variation in prop delay. On 1ns prop
>  >> devices, I don't think 50-100ps jitter is unreasonable under the most
>  >> optimum conditions, the most careful circuit
>  >> layout, and constant repeatable inputs. I don't think 10ps is
> achievable
>  >> under any dynamic conditions IMHO.
>  >>
>  >>> One can average 1,000,000 readings of the 10 ps jitter
>  >>> If they are truly random, that can give you a 1e-3 improvement (square
>  >>> root of number of samples averaged)
>  >>
>  >> You are now averaging the "repeatable" jitter.
>  >>
>  >> KE5FX's website shows a diagram and a link to your diagram as well. Are
>  >> you
>  >> using a digital phase detector or a mixer
>  >> as shown? BTW, KE5FX refers to DAQ as your update to the design, where
> I
>  >> believe he meant an ADC.
>  >>
>  >> You have my curiosity peaked. Do you have an analysis of the loop
>  >> sensitivity/resolution?
>  >>
>  >> Bob
>  >>
>  >>
>  >> ----- Original Message -----
>  >> From: "WarrenS" <warrensjmail-one at yahoo.com>
>  >> To: "Discussion of precise time and frequency measurement" <time-nuts
> at
>  >> febo.com>
>  >> Sent: Friday, June 18, 2010 6:49 PM
>  >> Subject: Re: [time-nuts] Advantages & Disadvantages of the TPLL Method
>  >>
>  >>
>  >>> Bob posted
>  >>>>can you explain it to me?
>  >>>
>  >>> Don't know, I'll give it ONE try.
>  >>> I'm not so good at explaining, but it is pretty basic if one does not
>  >>> start assuming that it can not be done at the
>  >>> start.
>  >>> It is mostly about averaging lots of those transitions, and the real
>  >>> trick
>  >>> is that it is not Digital.
>  >>> Analog has no lower limits except manly for Johnson noise type effects
>  >>> (mostly).
>  >>>
>  >>> Example with some random picked numbers.
>  >>> and assuming all analog that has no digital steps in it to limit
>  >>> resolution or add noise.
>  >>>
>  >>> Say you have a nice logic gate with 1 ns delay
>  >>> If you make it all nice and clean, and repeatable such as constant PS,
>  >>> rise time etc.
>  >>> Then one can get repeatable results say 100 times better from cycle to
>  >>> cycle in the short term.
>  >>> so down to 10ps repeatable.
>  >>> Now make things even more clean with no variations and assuming random
>  >>> noise.
>  >>> Now if one is doing this at 10 MHz and only cares about the average
> over
>  >>> 0.1 sec (10 Hz)
>  >>> One can average 1,000,000 readings of the 10 ps jitter
>  >>> If they are truly random, that can give you a 1e-3 improvement (square
>  >>> root of number of samples averaged)
>  >>> so now down to 10 fs of average jitter at 10 Hz for a 10 MHZ gate
>  >>> starting
>  >>> with a 1ns initial delay.
>  >>>
>  >>> OF course if Anything changes at all, it will drift much more than
> that,
>  >>> which may or may not mater much depending on
>  >>> what one is doing.
>  >>> If you only really care about the difference between any two
> consecutive
>  >>> 100 ms reading that are next to each other,
>  >>> as is (mostly) the case in ADEV, then not a big deal.
>  >>>
>  >>> IF it does matter or you want to do better, the next step is to do it
> all
>  >>> differential, so you are looking at only the
>  >>> different of two separate independent but equal circuits. Differential
>  >>> can
>  >>> give, say a 1000 to one or better
>  >>> improvement in drift due to common things such as temperature etc.
>  >>>
>  >>> If that helps explain the basics, good, if not you need to ask others
> to
>  >>> explain it better.
>  >>>
>  >>> And yes there all kinds of things that can & do go wrong and many ways
> to
>  >>> screw it up.
>  >>> so as easy as it sounds, it does take a bit of skill and art to do it.
>  >>> Especially when one realizes that you are measuring things << 0.001 in
> of
>  >>> distance change will have major effects on
>  >>> because of the speed of light.
>  >>> (approx 1ft /ns, 0.01 in/ps, 1 micron/4fs)
>  >>>
>  >>>
>  >>> Now if one starts out, not with a gate but a phase detector that is
> made
>  >>> for such things, and averages enough (but not
>  >>> to long) and is real careful,
>  >>> 1fs resolution is possible in the 100 Hz range with 10 MHz
>  >>>
>  >>> 10 MHz & 1fs at 100 Hz gives 1e-13 freq variation resolution at tau
> 10ms
>  >>> The simple BB TPLL is only getting about a tenth of that, (as shown on
>  >>> John's test plots) so it can be made much
>  >>> better with enough care, if anyone has a ref osc that needs it.
>  >>> But as I am always so quick to point out, the BB tester was not
> optimized
>  >>> for any one thing, It's performance was
>  >>> selected as a compromise for 'KISS' reasons. (KISS = Keep It Simple so
>  >>> the experts can understand.)
>  >>>
>  >>> please let me know on or off line if I'm wasting my time trying to
>  >>> explain
>  >>> this to the non "nut experts" without the
>  >>> help of the fancy math papers.
>  >>>
>  >>> ws
>  >>>
>  >>> *********************
>  >>> [time-nuts] Advantages & Disadvantages of the TPLL Method
>  >>> Robert Benward rbenward at verizon.net
>  >>> Fri Jun 18 20:23:40 UTC 2010
>  >>> Previous message: [time-nuts] Advantages & Disadvantages of the TPLL
>  >>> Method
>  >>> Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
>  >>>
>  >>> Warren,
>  >>> I'm a newbie, so can you explain it to me? Femto anything is something
>  >>> mostly reserved for a well equipped lab. How do you do it when most
>  >>> digital
>  >>> logic has jitter several of orders of magnitude greater?
>  >>>
>  >>> Bob
>  >>>
>  >>> *************************
>  >>
>  >>
>  >> _______________________________________________
>  >> time-nuts mailing list -- time-nuts at 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 at febo.com
>  > To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>  > and follow the instructions there.
>  >
>
>
>
>  --
>  Steve Rooke - ZL3TUV & G8KVD
>  The only reason for time is so that everything doesn't happen at once.
>  - Einstein
>
>  _______________________________________________
>  time-nuts mailing list -- time-nuts at 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 at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>



-- 
Steve Rooke - ZL3TUV & G8KVD
The only reason for time is so that everything doesn't happen at once.
- Einstein



More information about the time-nuts mailing list