[time-nuts] Re: FS740 Thoughts?
Hi This is one of those classic examples of what marketing writes and what (once the beer is poured) engineering actually knows. If you dive into the grubby details of the filtering. Been there / done that with a number of these guys. It’s very much *not* set up for GPS. The hanging bridge stuff PHK mentions in a previous post is one example. The need to adapt the filter to the monitored results is another weak area. Just how and when this or that pops up varies, but it will pop up. Bob > On Nov 18, 2021, at 2:02 PM, Dana Whitlow wrote: > > Bob, I seem to remember reading in the manual on the PRS-10 where they > expounded > at length about how they took great pains to make the unit work well with > typical noisy > GPS PPS input. > > I'm using a CNS Clock II as my primary GPSDO, checking phase slippage with > respect > to my PRS-10 to make decisions about when (or when not) to manually tweak > the PRS-10. > The CNS seems awfully noisy to me, even though the mfr says that it has > strong hanging > bridge correction. I run the two 10 MHz signals through a quadrature > demodulator whose > I & Q outputs go into a 2-channel DSO running in extremely low "roll > mode". I take a glance > at it several times per day, with each glance showing me the most recent > ~4-hour history. > When I do tweak, I strive to set the PRS-10 about 1E-11 low in frequency, > which yields a > couple of months of hands-off operation before it drifts to 1E-11 on the > high side. For > now, that has to suffice. But I'd rather not have to pay so much attention > to it, which is > why I dream about buying a new PRS-10 with PPS locking capability. > > Dana > > > On Thu, Nov 18, 2021 at 8:47 AM Poul-Henning Kamp > wrote: > >> >> Bob kb8tq writes: >> >>> The 'filter' in the PRS-10 is not really set up for a GPS sort of signal. >> >> It is particularly bad at handling GPS's because of the "hanging bridge" >> phenomena, and the better the GPS, the worse the result... >> >> I tried injecting the "negative sawtooth" via the serial port to my >> PRS10 but firmware features/bugs prevented that. >> >> The offset could only be changed permanently in the saved configuration, >> you could not change the running value on a second to second basis. >> >> -- >> 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@lists.febo.com -- To unsubscribe send >> an email to time-nuts-le...@lists.febo.com >> To unsubscribe, go to and follow the instructions there. >> > ___ > time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an > email to time-nuts-le...@lists.febo.com > To unsubscribe, go to and follow the instructions there. ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Bob, I seem to remember reading in the manual on the PRS-10 where they expounded at length about how they took great pains to make the unit work well with typical noisy GPS PPS input. I'm using a CNS Clock II as my primary GPSDO, checking phase slippage with respect to my PRS-10 to make decisions about when (or when not) to manually tweak the PRS-10. The CNS seems awfully noisy to me, even though the mfr says that it has strong hanging bridge correction. I run the two 10 MHz signals through a quadrature demodulator whose I & Q outputs go into a 2-channel DSO running in extremely low "roll mode". I take a glance at it several times per day, with each glance showing me the most recent ~4-hour history. When I do tweak, I strive to set the PRS-10 about 1E-11 low in frequency, which yields a couple of months of hands-off operation before it drifts to 1E-11 on the high side. For now, that has to suffice. But I'd rather not have to pay so much attention to it, which is why I dream about buying a new PRS-10 with PPS locking capability. Dana On Thu, Nov 18, 2021 at 8:47 AM Poul-Henning Kamp wrote: > > Bob kb8tq writes: > > > The 'filter' in the PRS-10 is not really set up for a GPS sort of signal. > > It is particularly bad at handling GPS's because of the "hanging bridge" > phenomena, and the better the GPS, the worse the result... > > I tried injecting the "negative sawtooth" via the serial port to my > PRS10 but firmware features/bugs prevented that. > > The offset could only be changed permanently in the saved configuration, > you could not change the running value on a second to second basis. > > -- > 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@lists.febo.com -- To unsubscribe send > an email to time-nuts-le...@lists.febo.com > To unsubscribe, go to and follow the instructions there. > ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Bob kb8tq writes: > The 'filter' in the PRS-10 is not really set up for a GPS sort of signal. It is particularly bad at handling GPS's because of the "hanging bridge" phenomena, and the better the GPS, the worse the result... I tried injecting the "negative sawtooth" via the serial port to my PRS10 but firmware features/bugs prevented that. The offset could only be changed permanently in the saved configuration, you could not change the running value on a second to second basis. -- 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@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Hi The 5 Hz (or whatever) switching and the hump definitely both originate in the same “feature” of the design. Does disabling the switching nuke the hump? If one has a PRS-10 I’d say it’s well worth trying. The “filter” in the PRS-10 is not really set up for a GPS sort of signal. It’s designed to handle a nice clean lab generated 1 pps. It’s there to lock the device up as part of a calibration routine. The 1 pps sync setups on other Rb’s are designed with the same target in mind. Bob > On Nov 18, 2021, at 7:50 AM, Dana Whitlow wrote: > > Poul, > > You did not really answer the basic question: "Does the 2.5 Hz > (or 5 Hz if applicable) C-field reversal cause the 2-sec bump in > the ADEV plot for the PRS-10?". > > BTW, some radio hams are involved with microwave and even > MMW communications between mobile stations, and I'd bet they > benefit quite a bit from any scheme that reduces magnetic field > sensitivity of their frequency references. Therefore, the C-field > reversal thing cannot be *all* bad. > > DanaK8YUM > > > Dana > > > On Thu, Nov 18, 2021 at 2:41 AM Poul-Henning Kamp > wrote: > >> >> Matt Huszagh writes: >>> "Poul-Henning Kamp" writes: >>> The PRS-10 switches the polarity of the C-field solenoid at 5Hz to cancel out varying external magnetic fields. If that is not a concern for you, for instance because you use it in a stationary application, it can be disabled with the "MS0" command. It's all in the manual. >>> >>> How does the 5Hz switching relate to the 2s hump in the ADEV plot? >> >> The manual says: >> >>"[...]the current in the coil is switched at a 5 Hz rate." >> >> You can either read that as: >> >>"There are five positive and five negative periods every second" >> >> or >>"The sign changes five times per second" >> >> It is not entirely obvious which reading is the correct one. >> >> When I experimented with it ages ago, I concluded the latter fit >> my data best, but that was a pretty early firmware version, with >> quite a number of variances from the manual. >> >> If your ambient magnetic field is stable, and it should be for >> time-nuts purposes, modulating the hyperfine transition is a bad >> idea, no matter the frequency. >> >> -- >> 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@lists.febo.com -- To unsubscribe send an > email to time-nuts-le...@lists.febo.com > To unsubscribe, go to and follow the instructions there. ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Dana Whitlow writes: > You did not really answer the basic question: "Does the 2.5 Hz > (or 5 Hz if applicable) C-field reversal cause the 2-sec bump in > the ADEV plot for the PRS-10?". I cannot answer that question as stated. My PRS10 died years ago, all I can say is that when I commissioned it, I disabled the MS because it improved things. > BTW, some radio hams are involved with microwave and even > MMW communications between mobile stations, [...] In any mobile application, or other application where the magnetic field is not nearly contant (for instance behind a steel door in a outdoor base-station cabinet, I would absolutely keep MS on. But for a time-nut installation, where the location is fixed, temperature controlled and, I would assume, due consideration is given to also keep the magnetic field from slamming around, I suggest MS be disabled. But it's a really trivial experiment to perform, so do it, measure and decide what is right for you. -- 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@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Poul, You did not really answer the basic question: "Does the 2.5 Hz (or 5 Hz if applicable) C-field reversal cause the 2-sec bump in the ADEV plot for the PRS-10?". BTW, some radio hams are involved with microwave and even MMW communications between mobile stations, and I'd bet they benefit quite a bit from any scheme that reduces magnetic field sensitivity of their frequency references. Therefore, the C-field reversal thing cannot be *all* bad. DanaK8YUM Dana On Thu, Nov 18, 2021 at 2:41 AM Poul-Henning Kamp wrote: > > Matt Huszagh writes: > > "Poul-Henning Kamp" writes: > > > > > The PRS-10 switches the polarity of the C-field solenoid at 5Hz to > > > cancel out varying external magnetic fields. > > > > > > If that is not a concern for you, for instance because you use it > > > in a stationary application, it can be disabled with the "MS0" command. > > > > > > It's all in the manual. > > > > How does the 5Hz switching relate to the 2s hump in the ADEV plot? > > The manual says: > > "[...]the current in the coil is switched at a 5 Hz rate." > > You can either read that as: > > "There are five positive and five negative periods every second" > > or > "The sign changes five times per second" > > It is not entirely obvious which reading is the correct one. > > When I experimented with it ages ago, I concluded the latter fit > my data best, but that was a pretty early firmware version, with > quite a number of variances from the manual. > > If your ambient magnetic field is stable, and it should be for > time-nuts purposes, modulating the hyperfine transition is a bad > idea, no matter the frequency. > > -- > 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@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Matt Huszagh writes: > "Poul-Henning Kamp" writes: > > > The PRS-10 switches the polarity of the C-field solenoid at 5Hz to > > cancel out varying external magnetic fields. > > > > If that is not a concern for you, for instance because you use it > > in a stationary application, it can be disabled with the "MS0" command. > > > > It's all in the manual. > > How does the 5Hz switching relate to the 2s hump in the ADEV plot? The manual says: "[...]the current in the coil is switched at a 5 Hz rate." You can either read that as: "There are five positive and five negative periods every second" or "The sign changes five times per second" It is not entirely obvious which reading is the correct one. When I experimented with it ages ago, I concluded the latter fit my data best, but that was a pretty early firmware version, with quite a number of variances from the manual. If your ambient magnetic field is stable, and it should be for time-nuts purposes, modulating the hyperfine transition is a bad idea, no matter the frequency. -- 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@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
"Poul-Henning Kamp" writes: > The PRS-10 switches the polarity of the C-field solenoid at 5Hz to > cancel out varying external magnetic fields. > > If that is not a concern for you, for instance because you use it > in a stationary application, it can be disabled with the "MS0" command. > > It's all in the manual. How does the 5Hz switching relate to the 2s hump in the ADEV plot? Matt ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Dana Whitlow writes: > But if you can stand the ADEV hump around 2 sec, the PRS-10 is delightful. The PRS-10 switches the polarity of the C-field solenoid at 5Hz to cancel out varying external magnetic fields. If that is not a concern for you, for instance because you use it in a stationary application, it can be disabled with the "MS0" command. It's all in the manual. -- 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@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Hi Dana: My PRS10 is a couple decades old and I use it in the 1PPS locking mode. But that may not have been the default. Maybe the surplus units just need the correct jumper change. https://prc68.com/I/PRS10.shtml -- Have Fun, Brooke Clarke https://www.PRC68.com axioms: 1. The extent to which you can fix or improve something will be limited by how well you understand how it works. 2. Everybody, with no exceptions, holds false beliefs. Original Message But if you can stand the ADEV hump around 2 sec, the PRS-10 is delightful. BTW, the ordinary factory new PRS-10 does provide for direct locking to 1 PPS, but many of the available surplus units had that feature removed. Tne new PRS-10's sold for years for $1595, but I see that it has gone up to $1695 since I had last checked. Dana On Wed, Nov 17, 2021 at 9:52 AM Bob kb8tq wrote: Hi In this case it’s not thermal. I’ve tested a ton of parts in the same environment ( maybe not 2,000 pounds but several hundred OCXO’s ). The PRS-10 gives you that hump at 2 seconds no matter what you do. It’s part of the basic design of the unit …. Bob On Nov 17, 2021, at 10:16 AM, Matt Huszagh wrote: Bob kb8tq writes: I have a 740 as do others on the list. The big gotcha is the OCXO in the device. It has various quirks that limit the performance of the device. The ADEV plot they show for the OCXO is a “sometimes / maybe / might” sort of thing. One of the “get to it eventually” projects here is to see if it can be replaced with something that has better ADEV performance. Good to know. Have you tried disciplining the PRS-10 in one of these units? Based on Dana's comments, it sounds like this may not exhibit the same ADEV susceptibility to the thermal environment. Matt ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there. ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there. ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Dana Whitlow writes: > But if you can stand the ADEV hump around 2 sec, the PRS-10 is delightful. > > BTW, the ordinary factory new PRS-10 does provide for direct locking to 1 > PPS, > but many of the available surplus units had that feature removed. Tne new > PRS-10's sold for years for $1595, but I see that it has gone up to $1695 > since > I had last checked. Have you had luck steering the PRS10 with GPS? One individual mentioned to me that despite repeated efforts, he was unable to discipline the PRS10 without significantly degrading the short-term stability and phase noise. However, if the performance as reported by SRS isn't just for some impossible, ideal environment than one solution would just be to replicate the circuitry they use for the loop filter and divider to generate the 10 MHz PPS signal. Or, to just purchase the FS740. Matt ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
But if you can stand the ADEV hump around 2 sec, the PRS-10 is delightful. BTW, the ordinary factory new PRS-10 does provide for direct locking to 1 PPS, but many of the available surplus units had that feature removed. Tne new PRS-10's sold for years for $1595, but I see that it has gone up to $1695 since I had last checked. Dana On Wed, Nov 17, 2021 at 9:52 AM Bob kb8tq wrote: > Hi > > In this case it’s not thermal. I’ve tested a ton of parts in the same > environment ( maybe > not 2,000 pounds but several hundred OCXO’s ). > > The PRS-10 gives you that hump at 2 seconds no matter what you do. It’s > part of the > basic design of the unit …. > > Bob > > > On Nov 17, 2021, at 10:16 AM, Matt Huszagh > wrote: > > > > Bob kb8tq writes: > > > >> I have a 740 as do others on the list. The big gotcha is the OCXO in > the device. > >> It has various quirks that limit the performance of the device. The > ADEV plot they > >> show for the OCXO is a “sometimes / maybe / might” sort of thing. One > of the > >> “get to it eventually” projects here is to see if it can be replaced > with something > >> that has better ADEV performance. > > > > Good to know. Have you tried disciplining the PRS-10 in one of these > > units? Based on Dana's comments, it sounds like this may not exhibit the > > same ADEV susceptibility to the thermal environment. > > > > Matt > ___ > time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send > an email to time-nuts-le...@lists.febo.com > To unsubscribe, go to and follow the instructions there. ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Bob kb8tq writes: > I have a 740 as do others on the list. The big gotcha is the OCXO in the > device. > It has various quirks that limit the performance of the device. The ADEV plot > they > show for the OCXO is a “sometimes / maybe / might” sort of thing. One of the > “get to it eventually” projects here is to see if it can be replaced with > something > that has better ADEV performance. Good to know. Have you tried disciplining the PRS-10 in one of these units? Based on Dana's comments, it sounds like this may not exhibit the same ADEV susceptibility to the thermal environment. Matt ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Hi In this case it’s not thermal. I’ve tested a ton of parts in the same environment ( maybe not 2,000 pounds but several hundred OCXO’s ). The PRS-10 gives you that hump at 2 seconds no matter what you do. It’s part of the basic design of the unit …. Bob > On Nov 17, 2021, at 10:16 AM, Matt Huszagh wrote: > > Bob kb8tq writes: > >> I have a 740 as do others on the list. The big gotcha is the OCXO in the >> device. >> It has various quirks that limit the performance of the device. The ADEV >> plot they >> show for the OCXO is a “sometimes / maybe / might” sort of thing. One of the >> “get to it eventually” projects here is to see if it can be replaced with >> something >> that has better ADEV performance. > > Good to know. Have you tried disciplining the PRS-10 in one of these > units? Based on Dana's comments, it sounds like this may not exhibit the > same ADEV susceptibility to the thermal environment. > > Matt ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
[time-nuts] Re: FS740 Thoughts?
Hi I have a 740 as do others on the list. The big gotcha is the OCXO in the device. It has various quirks that limit the performance of the device. The ADEV plot they show for the OCXO is a “sometimes / maybe / might” sort of thing. One of the “get to it eventually” projects here is to see if it can be replaced with something that has better ADEV performance. Be wary of folks showing up in dirt parking lots in the middle of Montana with one in the back of their car …. :) Doubly so if the car shows signs of having hauled a dozen Cs standards up various mountains….. If they are named Tom, simply give up and accept that once you see one you will eventually buy one… The process goes sort of like: Gee, that’s a lot of money. Later on, … I need a few more standard lines at this or that frequency … Then it goes to If only I could come up with *that* signal … hmmm … 740 would do that … h…. off you go to get one. It’s a handy box to do a variety of things with. The only issue I’ve seen is the OCXO performance. Bob > On Nov 17, 2021, at 12:50 AM, Matt Huszagh wrote: > > Hi, > > I'm curious to hear people's thoughts on the SRS FS740 > (https://www.thinksrs.com/products/fs740.html) as a lab frequency > standard. The phase noise performance of the OCXO and Rb options seem > quite good (the same specs apparently apply during GPS disciplining too) > > | offset | TCXO | OCXO/Rb | > | (Hz) | (dBc/Hz) | (dBc/Hz) | > |+--+--| > | 1 | -90| -108 | > | 10 | -104 | -130 | > | 100| -125 | -144 | > | 1k | -142 | -143 | > | 10k| -147 | -148 | > | 100k | -149 | -149 | > | 1M | -153 | -153 | > > as does the Allan deviation (during GPS disciplining) > > | time interval (s) | TCXO | OCXO | Rb | > |---+---+---+---| > | 0.01 | 1E-10 | 2E-11 | 2E-11 | > | 0.1 | 4E-11 | 4E-12 | 4E-12 | > | 1 | 3E-11 | 2E-12 | 1E-11 | > | 10| 1E-10 | 2E-12 | 6E-12 | > | 100 | 6E-11 | 2E-12 | 1E-12 | > | 1k| 6E-12 | 3E-12 | 1E-12 | > | 10k | 6E-13 | 6E-13 | 3E-13 | > | 100k | 6E-14 | 6E-14 | 3E-14 | > | 1M| 5E-15 | 5E-15 | 3E-15 | > > (I used the typical performance plots for the phase noise and Allan > deviation data) > > The OCXO does not exhibit the the hump around 2s seen by the rubidium > oscillator option (PRS10) and therefore could be seen as better in > applications not requiring the better holdover performance by the Rb. > > It also provides a pulse output with jitter < 50 ps RMS that could be > used as a PPS signal. It can also provide a low noise 100 MHz output > signal and can act as a frequency counter and DDS with high resolution > and accuracy. And it comes with the schematics! But it's definitely > not cheap. Other thoughts/opinions? > > Matt > ___ > time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an > email to time-nuts-le...@lists.febo.com > To unsubscribe, go to and follow the instructions there. ___ time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-le...@lists.febo.com To unsubscribe, go to and follow the instructions there.
