Another way to look at this is to say a real VCXO has an analog input, subject to noise, pushing and pullin, potentially finite DAC resolution, and a not insignificant Q associated with the crystal. If you were to average a crystal oscillator frequency and sampled say every 100 seconds, you’d find the tuning steps approximate a nearly continuous adjust range. A digital NCO at best is one count ahead or one count behind. The shorter the time interval the worse it looks. If you divided a higher frequency NCO to obtain a lower frequency output, you’d improve by the division ratio given fixed sample interval. Good frequency planning can probably help in rendering a system amenable to analysis, but ultimately even with good clock planning you're going to have to deal with loop bandwidth and that pesky +-1 count per interval, and it's transformation from time domain to frequency domain. ________________________________ From: rbenward--- via time-nuts <time-nuts@lists.febo.com> Sent: Thursday, January 16, 2025 10:56:04 AM To: 'Discussion of precise time and frequency measurement' <time-nuts@lists.febo.com> Cc: rbenward@verizon.net <rbenward@verizon.net> Subject: [time-nuts] Re: GPSDO "breathing" behavior Hi Bob, I agree with you, you don't "tune" the oscillator. To expand, the NCO has two parts, the oscillator, and the numerical control. The NCO is a very long digital divider, with single or multiple loops. The oscillator can be at any high frequency, and the NCO "tuning" is what gives you the span of 10Hz to 10MHz. The oscillator will drift, but the processor will adjust the output by "tuning" the NCO. The tuning is simply writing new digital values to the NCO. I assume, whether the oscillator is built in to the NCO or separate, it's a tiny little oscillator, probably not even a TCXO. So, it drifts, again, I assume for the most part, in one direction. The processor will then periodically (1PPS) adjust the NCO before the phase rolls over. The oscillator, NCO, and the processor form what looks very much like a frequency synthesizer. You mentioned "disciplining". Isn't that more of a processor disciplining? The processor learns the drift, and when in hold over, continues to adjust the EFC to zero the anticipated drift? Is the prediction function active when the 1PPS is available, or only when in holdover? Maybe it's all semantics! Bob B -----Original Message----- From: Bob Camp via time-nuts <time-nuts@lists.febo.com> Sent: Wednesday, January 15, 2025 5:41 PM To: Discussion of precise time and frequency measurement <time-nuts@lists.febo.com> Cc: Bob Camp <kb8tq@n1k.org> Subject: [time-nuts] Re: GPSDO "breathing" behavior Importance: High Hi What the modules do both on the PPS and on their other signals is to “dither” an internal oscillator. The oscillator is *not* tuned to frequency. It free runs at some independent frequency (and phase). The edges of the output are shifted from time to time to drop or add one period of the internal oscillator. This is not the same thing as tuning the oscillator on frequency. That’s why the “disciplining” term is important in “GPSDO”. Bob > On Jan 15, 2025, at 4:06 PM, Peter McCollum via time-nuts <time-nuts@lists.febo.com> wrote: > > Thanks for the comments so far folks! I am definitely NOT questioning > anyone's expertise, but I remain somewhat confused about the assertion > that the CW12-TIM is "not a GPSDO". If not, then what is it? It > obviously "disciplines an oscillator" in some way (see data below). > > To clarify the "breathing" behavior: > Attached is a pic of the Lissajous, comparing the Sulzer vs the > CW12-TIM outputs, both at 1 Mhz. I have adjusted the fine-tuning on > the Sulzer such that the ellipse gradually changes shape over a period of many minutes. > BUT, over a period of perhaps 3 or 4 seconds, the ellipse goes back > and forth between the two versions shown in the picture. As I > mentioned before, it is NOT a sudden change, but rather looks like a > lung expanding and contracting somewhat. > I interpret that to mean that there is a *temporary* change in freq, > lasting less than 0.5 seconds or so, then the freq does the same thing > in the opposite direction, and the ellipse is back to it's original > shape/size. In the time frame of (say) a minute or two, there is NO > overall change to the ellipse. > This behavior is only seen with the CW12-TIM output - if I compare the > Sulzer against something else that is "pretty stable" (such as an HP > 5345A), I don't see the "breathing". > > Another data point: > If I power up the CW12-TIM with its antenna shielded to limit its > access to GPS signals, the Lissajous "rotates" at a rate of perhaps > 1.5-2 Hz (out of > 1 Mhz). Then, as soon as it starts to acquire sats, the Lissajous > stops moving... The datasheet says that the freq output is reliable > when it has a 2-D solution. > > Also attached is two pages from the CW12-TIM docs. > Is that NOT describing a "true GPSDO" ? > I see that the doc says "full PCR MTIE performance", although I > certainly don't understand the implications of that statement... > > Pete > > > > On Wed, Jan 15, 2025 at 1:06 PM Bob Camp via time-nuts < > time-nuts@lists.febo.com> wrote: > >> Hi >> >> GPSDO = GPS Disciplined Oscilator >> >> They have: >> >> 1) A GPS module of some sort >> >> 2) A control loop that also acts as a noise filter >> >> 3) An independent oscillator that the control loop manages. >> >> With just the module you are missing 2 and 3. >> >> GPS modules have provided “chopped up” output signals for at least >> two decades and closer to three. >> What they provide is not the same thing as what a GPSDO delivers. >> >> Bob >> >>> On Jan 15, 2025, at 9:19 AM, Peter McCollum via time-nuts < >> time-nuts@lists.febo.com> wrote: >>> >>> Azelio wrote: >>>> The CW12-TIM is not a GPSDO >>> >>> From the CW12-TIM manual: >>> "The graph below demonstrates the ability of the CW12-TIM GPS to >>> continue to provide a GPS disciplined output frequency with the GPS >>> aerial located completely inside a building..." >>> >>> Is it just a matter of semantics, or is there a firm definition of >>> "disciplined"? >>> >>> Pete >>> >>> >>> >>> On Wed, Jan 15, 2025 at 3:01 AM Azelio Boriani via time-nuts < >>> time-nuts@lists.febo.com> wrote: >>> >>>> The CW12-TIM is not a GPSDO, the frequency output is not >>>> disciplined, it is, like the PPS, derived from the division of the >>>> free running internal clock, driving the divisor (someway: >>>> added/skipped cycles) to align it >> with >>>> the GPS. To mitigate the "breathing" there should be a sort of >>>> sawtooth correction also for the frequency output (not possible for that speed). >>>> >>>> On Wed, Jan 15, 2025 at 8:32 AM Peter McCollum via time-nuts < >>>> time-nuts@lists.febo.com> wrote: >>>> >>>>> Just curious if this is normal behavior. >>>>> I use an X-Y display to show a Lissajous of my CW12-TIM module vs >>>>> a >>>> fairly >>>>> stable oscillator (Sulzer 5A, or Spectracom 8161), all at 1 MHz. >>>>> The CW12's output gives an (apparently) very solid output averaged >>>>> over many seconds or longer. >>>>> But in the short term, the Lissajous does a "breathing" behavior - >> every >>>> 3 >>>>> or 4 seconds (it's variable), the pattern expands a few degrees in >>>>> one direction, then goes back in the other direction. It's NOT a >>>>> sudden >> jump, >>>>> rather it is very much like watching someone's lungs taking a >>>>> breath >> then >>>>> letting it out. >>>>> >>>>> Is it normal, or does it imply that the "discipline" in the GPSDO >>>>> is repeatedly needing a correction? >>>>> >>>>> Thanks for any comments! >>>>> >>>>> Pete >>>>> _______________________________________________ >>>>> time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe >>>>> send an email to time-nuts-leave@lists.febo.com >>>>> >>>> _______________________________________________ >>>> time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe >>>> send an email to time-nuts-leave@lists.febo.com >>> _______________________________________________ >>> time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe >>> send an email to time-nuts-leave@lists.febo.com >> _______________________________________________ >> time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe >> send an email to time-nuts-leave@lists.febo.com > <https://urldefense.proofpoint.com/v2/url?u=http-3A__Lissajous.JPG&d=DwIGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=JsDsKeR7cZC8wbZhIlxxBQ&m=yVh8m8tmSsEJ7hVfAOlEU8Yu0BRbEo6imagx9T1UmSjyXgVofche0UTjoChwhtPC&s=XIJgAKCNHYiFSQbTueQahZaV1baoT4TDR-5QIb2uv7k&e=><https://urldefense.proofpoint.com/v2/url?u=http-3A__p2.jpg&d=DwIGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=JsDsKeR7cZC8wbZhIlxxBQ&m=yVh8m8tmSsEJ7hVfAOlEU8Yu0BRbEo6imagx9T1UmSjyXgVofche0UTjoChwhtPC&s=2tKTcEL1FZSJkIl5Z1GuTTPRoo3NGvUrn1M78-FpW7Q&e=><https://urldefense.proofpoint.com/v2/url?u=http-3A__p1.jpg&d=DwIGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=JsDsKeR7cZC8wbZhIlxxBQ&m=yVh8m8tmSsEJ7hVfAOlEU8Yu0BRbEo6imagx9T1UmSjyXgVofche0UTjoChwhtPC&s=CWqDfQKJ3oIz8lF6ks5e50WakcdqoEn8WtAJdyCiEQI&e=>_______________________________________ > ________ time-nuts mailing list -- time-nuts@lists.febo.com To > unsubscribe send an email to time-nuts-leave@lists.febo.com _______________________________________________ time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe send an email to time-nuts-leave@lists.febo.com _______________________________________________ time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe send an email to time-nuts-leave@lists.febo.com

Hi The purpose of the control loop in a GPSDO is to eliminate noise, not to add it. A GPSDO has a clean output spectrum rather than a total mess. That’s part of why it is what it is. The processing in the GPS module does no noise reduction at all. You get all the trash the GPS puts out based on it’s second to second estimates. Bob > On Jan 16, 2025, at 11:56 AM, rbenward--- via time-nuts <time-nuts@lists.febo.com> wrote: > > Hi Bob, > I agree with you, you don't "tune" the oscillator. To expand, the NCO has two parts, the oscillator, and the numerical control. The NCO is a very long digital divider, with single or multiple loops. The oscillator can be at any high frequency, and the NCO "tuning" is what gives you the span of 10Hz to 10MHz. > > The oscillator will drift, but the processor will adjust the output by "tuning" the NCO. The tuning is simply writing new digital values to the NCO. I assume, whether the oscillator is built in to the NCO or separate, it's a tiny little oscillator, probably not even a TCXO. So, it drifts, again, I assume for the most part, in one direction. The processor will then periodically (1PPS) adjust the NCO before the phase rolls over. The oscillator, NCO, and the processor form what looks very much like a frequency synthesizer. > > You mentioned "disciplining". Isn't that more of a processor disciplining? The processor learns the drift, and when in hold over, continues to adjust the EFC to zero the anticipated drift? Is the prediction function active when the 1PPS is available, or only when in holdover? Maybe it's all semantics! > > Bob B > > -----Original Message----- > From: Bob Camp via time-nuts <time-nuts@lists.febo.com> > Sent: Wednesday, January 15, 2025 5:41 PM > To: Discussion of precise time and frequency measurement <time-nuts@lists.febo.com> > Cc: Bob Camp <kb8tq@n1k.org> > Subject: [time-nuts] Re: GPSDO "breathing" behavior > Importance: High > > Hi > > What the modules do both on the PPS and on their other signals is to “dither” an internal oscillator. > The oscillator is *not* tuned to frequency. It free runs at some independent frequency (and phase). > The edges of the output are shifted from time to time to drop or add one period of the internal oscillator. > > This is not the same thing as tuning the oscillator on frequency. That’s why the “disciplining” term is important in “GPSDO”. > > Bob > >> On Jan 15, 2025, at 4:06 PM, Peter McCollum via time-nuts <time-nuts@lists.febo.com> wrote: >> >> Thanks for the comments so far folks! I am definitely NOT questioning >> anyone's expertise, but I remain somewhat confused about the assertion >> that the CW12-TIM is "not a GPSDO". If not, then what is it? It >> obviously "disciplines an oscillator" in some way (see data below). >> >> To clarify the "breathing" behavior: >> Attached is a pic of the Lissajous, comparing the Sulzer vs the >> CW12-TIM outputs, both at 1 Mhz. I have adjusted the fine-tuning on >> the Sulzer such that the ellipse gradually changes shape over a period of many minutes. >> BUT, over a period of perhaps 3 or 4 seconds, the ellipse goes back >> and forth between the two versions shown in the picture. As I >> mentioned before, it is NOT a sudden change, but rather looks like a >> lung expanding and contracting somewhat. >> I interpret that to mean that there is a *temporary* change in freq, >> lasting less than 0.5 seconds or so, then the freq does the same thing >> in the opposite direction, and the ellipse is back to it's original >> shape/size. In the time frame of (say) a minute or two, there is NO >> overall change to the ellipse. >> This behavior is only seen with the CW12-TIM output - if I compare the >> Sulzer against something else that is "pretty stable" (such as an HP >> 5345A), I don't see the "breathing". >> >> Another data point: >> If I power up the CW12-TIM with its antenna shielded to limit its >> access to GPS signals, the Lissajous "rotates" at a rate of perhaps >> 1.5-2 Hz (out of >> 1 Mhz). Then, as soon as it starts to acquire sats, the Lissajous >> stops moving... The datasheet says that the freq output is reliable >> when it has a 2-D solution. >> >> Also attached is two pages from the CW12-TIM docs. >> Is that NOT describing a "true GPSDO" ? >> I see that the doc says "full PCR MTIE performance", although I >> certainly don't understand the implications of that statement... >> >> Pete >> >> >> >> On Wed, Jan 15, 2025 at 1:06 PM Bob Camp via time-nuts < >> time-nuts@lists.febo.com> wrote: >> >>> Hi >>> >>> GPSDO = GPS Disciplined Oscilator >>> >>> They have: >>> >>> 1) A GPS module of some sort >>> >>> 2) A control loop that also acts as a noise filter >>> >>> 3) An independent oscillator that the control loop manages. >>> >>> With just the module you are missing 2 and 3. >>> >>> GPS modules have provided “chopped up” output signals for at least >>> two decades and closer to three. >>> What they provide is not the same thing as what a GPSDO delivers. >>> >>> Bob >>> >>>> On Jan 15, 2025, at 9:19 AM, Peter McCollum via time-nuts < >>> time-nuts@lists.febo.com> wrote: >>>> >>>> Azelio wrote: >>>>> The CW12-TIM is not a GPSDO >>>> >>>> From the CW12-TIM manual: >>>> "The graph below demonstrates the ability of the CW12-TIM GPS to >>>> continue to provide a GPS disciplined output frequency with the GPS >>>> aerial located completely inside a building..." >>>> >>>> Is it just a matter of semantics, or is there a firm definition of >>>> "disciplined"? >>>> >>>> Pete >>>> >>>> >>>> >>>> On Wed, Jan 15, 2025 at 3:01 AM Azelio Boriani via time-nuts < >>>> time-nuts@lists.febo.com> wrote: >>>> >>>>> The CW12-TIM is not a GPSDO, the frequency output is not >>>>> disciplined, it is, like the PPS, derived from the division of the >>>>> free running internal clock, driving the divisor (someway: >>>>> added/skipped cycles) to align it >>> with >>>>> the GPS. To mitigate the "breathing" there should be a sort of >>>>> sawtooth correction also for the frequency output (not possible for that speed). >>>>> >>>>> On Wed, Jan 15, 2025 at 8:32 AM Peter McCollum via time-nuts < >>>>> time-nuts@lists.febo.com> wrote: >>>>> >>>>>> Just curious if this is normal behavior. >>>>>> I use an X-Y display to show a Lissajous of my CW12-TIM module vs >>>>>> a >>>>> fairly >>>>>> stable oscillator (Sulzer 5A, or Spectracom 8161), all at 1 MHz. >>>>>> The CW12's output gives an (apparently) very solid output averaged >>>>>> over many seconds or longer. >>>>>> But in the short term, the Lissajous does a "breathing" behavior - >>> every >>>>> 3 >>>>>> or 4 seconds (it's variable), the pattern expands a few degrees in >>>>>> one direction, then goes back in the other direction. It's NOT a >>>>>> sudden >>> jump, >>>>>> rather it is very much like watching someone's lungs taking a >>>>>> breath >>> then >>>>>> letting it out. >>>>>> >>>>>> Is it normal, or does it imply that the "discipline" in the GPSDO >>>>>> is repeatedly needing a correction? >>>>>> >>>>>> Thanks for any comments! >>>>>> >>>>>> Pete >>>>>> _______________________________________________ >>>>>> time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe >>>>>> send an email to time-nuts-leave@lists.febo.com >>>>>> >>>>> _______________________________________________ >>>>> time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe >>>>> send an email to time-nuts-leave@lists.febo.com >>>> _______________________________________________ >>>> time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe >>>> send an email to time-nuts-leave@lists.febo.com >>> _______________________________________________ >>> time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe >>> send an email to time-nuts-leave@lists.febo.com >> <Lissajous.JPG><p2.jpg><p1.jpg>_______________________________________ >> ________ time-nuts mailing list -- time-nuts@lists.febo.com To >> unsubscribe send an email to time-nuts-leave@lists.febo.com > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe send an email to time-nuts-leave@lists.febo.com > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe send an email to time-nuts-leave@lists.febo.com

Thomas Valerio

On 2025-01-15 22:38, Thomas Valerio via time-nuts wrote: > I am probably one of the least qualified members on this list to > comment. > HOWEVER, I think that the the fact that the CW12-TIM is meant to be a > ((an) almost, maybe exact or better) drop in replacement for a Motorola > M12+, which by design has NEVER been or ASSERTED to be ANYTHING other > than > only the GPS module for a GPSDO or frequency standard should be a MAJOR > clue. Just my $.02 > This explains and demonstrates the M12 sawtooth pretty well. http://www.leapsecond.com/pages/m12/sawtooth.htm Marshall Eubanks > Thomas Valerio > > >> Thanks for the comments so far folks! I am definitely NOT questioning >> anyone's expertise, but I remain somewhat confused about the assertion >> that >> the CW12-TIM is "not a GPSDO". If not, then what is it? It obviously >> "disciplines an oscillator" in some way (see data below). >> >> To clarify the "breathing" behavior: >> Attached is a pic of the Lissajous, comparing the Sulzer vs the >> CW12-TIM >> outputs, both at 1 Mhz. I have adjusted the fine-tuning on the Sulzer >> such >> that the ellipse gradually changes shape over a period of many >> minutes. >> BUT, over a period of perhaps 3 or 4 seconds, the ellipse goes back >> and >> forth between the two versions shown in the picture. As I mentioned >> before, >> it is NOT a sudden change, but rather looks like a lung expanding and >> contracting somewhat. >> I interpret that to mean that there is a *temporary* change in freq, >> lasting less than 0.5 seconds or so, then the freq does the same thing >> in >> the opposite direction, and the ellipse is back to it's original >> shape/size. In the time frame of (say) a minute or two, there is NO >> overall >> change to the ellipse. >> This behavior is only seen with the CW12-TIM output - if I compare the >> Sulzer against something else that is "pretty stable" (such as an HP >> 5345A), I don't see the "breathing". >> >> Another data point: >> If I power up the CW12-TIM with its antenna shielded to limit its >> access >> to >> GPS signals, the Lissajous "rotates" at a rate of perhaps 1.5-2 Hz >> (out of >> 1 Mhz). Then, as soon as it starts to acquire sats, the Lissajous >> stops >> moving... The datasheet says that the freq output is reliable when it >> has >> a >> 2-D solution. >> >> Also attached is two pages from the CW12-TIM docs. >> Is that NOT describing a "true GPSDO" ? >> I see that the doc says "full PCR MTIE performance", although I >> certainly >> don't understand the implications of that statement... >> >> Pete >> >> >> >> On Wed, Jan 15, 2025 at 1:06 PM Bob Camp via time-nuts < >> time-nuts@lists.febo.com> wrote: >> >>> Hi >>> >>> GPSDO = GPS Disciplined Oscilator >>> >>> They have: >>> >>> 1) A GPS module of some sort >>> >>> 2) A control loop that also acts as a noise filter >>> >>> 3) An independent oscillator that the control loop manages. >>> >>> With just the module you are missing 2 and 3. >>> >>> GPS modules have provided “chopped up” output signals for at >>> least >>> two >>> decades and closer to three. >>> What they provide is not the same thing as what a GPSDO delivers. >>> >>> Bob >>> >>> > On Jan 15, 2025, at 9:19 AM, Peter McCollum via time-nuts < >>> time-nuts@lists.febo.com> wrote: >>> > >>> > Azelio wrote: >>> >> The CW12-TIM is not a GPSDO >>> > >>> > From the CW12-TIM manual: >>> > "The graph below demonstrates the ability of the CW12-TIM GPS to >>> continue >>> > to provide a GPS disciplined output >>> > frequency with the GPS aerial located completely inside a building..." >>> > >>> > Is it just a matter of semantics, or is there a firm definition of >>> > "disciplined"? >>> > >>> > Pete >>> > >>> > >>> > >>> > On Wed, Jan 15, 2025 at 3:01 AM Azelio Boriani via time-nuts < >>> > time-nuts@lists.febo.com> wrote: >>> > >>> >> The CW12-TIM is not a GPSDO, the frequency output is not disciplined, >>> it >>> >> is, like the PPS, derived from the division of the free running >>> internal >>> >> clock, driving the divisor (someway: added/skipped cycles) to align >>> it >>> with >>> >> the GPS. To mitigate the "breathing" there should be a sort of >>> sawtooth >>> >> correction also for the frequency output (not possible for that >>> speed). >>> >> >>> >> On Wed, Jan 15, 2025 at 8:32 AM Peter McCollum via time-nuts < >>> >> time-nuts@lists.febo.com> wrote: >>> >> >>> >>> Just curious if this is normal behavior. >>> >>> I use an X-Y display to show a Lissajous of my CW12-TIM module vs a >>> >> fairly >>> >>> stable oscillator (Sulzer 5A, or Spectracom 8161), all at 1 MHz. >>> >>> The CW12's output gives an (apparently) very solid output averaged >>> over >>> >>> many seconds or longer. >>> >>> But in the short term, the Lissajous does a "breathing" behavior - >>> every >>> >> 3 >>> >>> or 4 seconds (it's variable), the pattern expands a few degrees in >>> one >>> >>> direction, then goes back in the other direction. It's NOT a sudden >>> jump, >>> >>> rather it is very much like watching someone's lungs taking a breath >>> then >>> >>> letting it out. >>> >>> >>> >>> Is it normal, or does it imply that the "discipline" in the GPSDO is >>> >>> repeatedly needing a correction? >>> >>> >>> >>> Thanks for any comments! >>> >>> >>> >>> Pete >>> >>> _______________________________________________ >>> >>> time-nuts mailing list -- time-nuts@lists.febo.com >>> >>> To unsubscribe send an email to time-nuts-leave@lists.febo.com >>> >>> >>> >> _______________________________________________ >>> >> time-nuts mailing list -- time-nuts@lists.febo.com >>> >> To unsubscribe send an email to time-nuts-leave@lists.febo.com >>> > _______________________________________________ >>> > time-nuts mailing list -- time-nuts@lists.febo.com >>> > To unsubscribe send an email to time-nuts-leave@lists.febo.com >>> _______________________________________________ >>> time-nuts mailing list -- time-nuts@lists.febo.com >>> To unsubscribe send an email to time-nuts-leave@lists.febo.com >> _______________________________________________ >> time-nuts mailing list -- time-nuts@lists.febo.com >> To unsubscribe send an email to time-nuts-leave@lists.febo.com > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe send an email to time-nuts-leave@lists.febo.com