Re: [time-nuts] Some results of PRS10 and Trimble Resolution
On Fri, June 30, 2006 4:34, [EMAIL PROTECTED] said: Hello Tom, I had a conversation with Sam S. from TSC the other day, and he said that it's probably not possible to get 10ns GPS accuracy anyways due to the multipath issues, Ionospheric issues, antenna survey issues, thermal issues etc. Why are dual freq. receivers not used in timing instruments? Ionospheric errors are almost removed. Surveyors in every small town on earth are routinely getting sub 5cm accuracy in real time. Are not time labs using it to transfer their cesium time to each other? -- Björn ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Why are dual freq. receivers not used in timing instruments? Ionospheric They are. But only in the high-end instruments. errors are almost removed. Surveyors in every small town on earth are routinely getting sub 5cm accuracy in real time. Are not time labs using it to transfer their cesium time to each other? True, but what's your guess on the price ratio between a 5 cm dual frequency survey-grade timing receiver and a M12 or Res-T? I'd guess 100x to 1000x. Anyone know for sure? /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
Hi Tom, The reason I investigated this is that I needed a very good (2E-012 @ 1s) short term ADEV reference to measure the performance of our GPSDO (which is around 2E-012 at 1 second as you know :) on the TSC5120A. I found that I could use the MTI 260 reference from TSC for this purpose up to about 50s intervalls, then switch to the Rb above 50s. Right. Depending on the UUT I sometimes do this too. They both have their strengths depending on the measurement intervall. Yes, and this is true in general. It is not uncommon to have to use two different references at two different times to cover the range of tau you want to measure. It would make sense to slave an MTI 260 OCXO to a PRS10 with a PLL lock time of about 20s. This would generate a very clean output - the Phase noise of the 260 is better than the PRS10, as well as the ADEV 20s would be improved. Well, you can try this, and I know what you're saying, but in my book it is more reliable and easier to make two runs; one with each pure clean reference standard rather than create (and debug, and trust) a noisy hybrid of the two references to make one run. For some tau in the middle the hybrid must always have more noise than one of the two references so I say just skip the hybrid and use each reference for the range of tau that it excels. /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
From: Tom Van Baak [EMAIL PROTECTED] Subject: Re: [time-nuts] Some results of PRS10 and Trimble Resolution Date: Thu, 29 Jun 2006 22:44:53 -0700 Message-ID: [EMAIL PROTECTED] Why are dual freq. receivers not used in timing instruments? Ionospheric They are. But only in the high-end instruments. The metrology labs use survey GPS receivers for this obvious reason. errors are almost removed. Surveyors in every small town on earth are routinely getting sub 5cm accuracy in real time. Are not time labs using it to transfer their cesium time to each other? True, but what's your guess on the price ratio between a 5 cm dual frequency survey-grade timing receiver and a M12 or Res-T? I'd guess 100x to 1000x. Anyone know for sure? Hmm, come to think of it, I have never really cared to figure it out. The surveying folks have fancy antennas too. They go around the world (south pole included) and measure all kinds of stuff, as the drift-rate of the polar ice in the Antarctic. The necessary components isn't directly high-volume compared to the normal L1 C/A receivers. Cheers, Magnus ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hello Tom, I had a conversation with Sam S. from TSC the other day, and he said that it's probably not possible to get 10ns GPS accuracy anyways due to the multipath issues, Ionospheric issues, antenna survey issues, thermal issues etc. Yes, with a standard GPS receiver or GPSDO, I very much agree with this. See also the links to GPS papers I posted earlier today which will give you a feel for what level of accuracy or stability you get from various GPS time transfer techniques. Note accuracy and stability are two different goals. Related to that, GPS-based frequency reference products are plentiful, cheap, and in widespread use while GPS-based time transfer products are few, very expensive, and have a small use base. I think most of us time-nuts use GPS as a source of precise time interval (GPS as a ~ 1e-13 stable frequency reference) rather than a source of absolute time (~ ns accurate UTC). This is why uncalibrated GPS receivers work for all of us. See Rick's papers on M12 calibration at USNO. http://www.gpstime.com/ http://tycho.usno.navy.mil/ptti/ptti2002/paper9.pdf Note NIST's use of regular GPS receivers (Oncore?) for frequency _stability_ (not time _accuracy_): http://www.tf.nist.gov/service/fms.htm So 10 ns _stability_ with GPS over a day is quite doable on the cheap. 10 ns _accuracy_ with GPS is quite another matter and requires a huge amount of work. On this list, DougH is probably the only one who's pulled it off: http://www.leapsecond.com/time-nuts.htm Our units typically average the GPS 1PPS over 30 minutes, so having less than 3.33ns error on the 1PPS capture may not improve things much because while the error stays always at +-3.33ns it get's averaged over the measurement intervall. 3.3ns / 30 m = 2e-12 and sets your lower bound. But it would depend on your choice of LO and PLL if this is a limiting factor or not. Here's something to try: deliberately degrade your 1PPS TIC resolution in software and see what effect it makes on the stability of your RF or 1PPS output. /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hi Björn: I don't think that's that case. I live in a small town and hired a local surveyor chosen because he's the guy that uses GPS. He was able to use two GPS receivers separated by say 100 feet and connected with a cable to determine an accurate relative bearing between his total station and the remote GPS. I had placed wood stakes at a few locations that were very close to Longitude and Latitude points ending exactly on the seconds, for example 123:09:50.0W by 39:11:24.0N. The idea was for him to locate the exact points and to locate the house GPS antenna. To do this he needed to go back to his office and post process the data. Since the points I wanted to know about were separated by exactly 1 second of Lat and 1 second of Lon he just placed his total station a few feet from the corner one and shot the other stakes. It's my understanding the the military has the crypto key needed to use the L2 channel to full effect and that's not available to surveyors. What surveyors do is record the carrier phase of both the L1 and L2 signals and resolve the ambiguity in post processing. There are a couple of flavors of post processing. In one only the data recorded in the field is used, in the other the filed data is combined with actual data taken at a nearby reference station from the satellites used in the field data. The second method corrects for errors in the ephemeris data broadcast by the satellites. I think what you may be referring to are the differential correction methods, either Low Frequency or satellite broadcasts that improve the position accuracy of GPS receivers. But I don't think these can be used to improve timing accuracy. Have Fun, Brooke Clarke -- w/Java http://www.PRC68.com w/o Java http://www.pacificsites.com/~brooke/PRC68COM.shtml http://www.precisionclock.com [EMAIL PROTECTED] wrote: . . . . Why are dual freq. receivers not used in timing instruments? Ionospheric errors are almost removed. Surveyors in every small town on earth are routinely getting sub 5cm accuracy in real time. Are not time labs using it to transfer their cesium time to each other? -- Björn ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
On Fri, June 30, 2006 20:49, Brooke Clarke said: Hi Björn: I don't think that's that case. I live in a small town and hired a local surveyor chosen because he's the guy that uses GPS. Maybe I was wrong with extrapolating the local situation to a global level. With the present dense network of CORS-like stations [1]. Surveyors are moving away from running their own RTK-base stations. With close to full coverage RTK-corrections is distributed via the cellphone network, which means the surveyor only need the rover side of the traditional equipment. antenna. To do this he needed to go back to his office and post process the data. That makes his GPS recievers cheaper, than the ones where the software options for RTK is enabled. He also does not need a real time radio connection. Precision is about the same or better, since he can download better ephemeris, and can tinker more with parameters in the software. It's my understanding the the military has the crypto key needed to use the L2 channel to full effect and that's not available to surveyors. That is true in theory. In reality Ashtech (and others) early found that the crypto code rate is much lower then the chipping rate. Details was posted here in the last 6 months? IIRC the known P-code will flip sign every 511 bits depending on the crypto code. This makes the highend receivers do much more than intended on L2. What surveyors do is record the carrier phase of both the L1 and L2 signals and resolve the ambiguity in post processing. There are a couple of flavors of post processing. And there is real time flavors as well, with the best going down below 5cm errors. I think what you may be referring to are the differential correction methods, either Low Frequency or satellite broadcasts that improve the position accuracy of GPS receivers. But I don't think these can be used to improve timing accuracy. These are the SBAS (WAAS, EGNOS, MSAT), commercial satellite based, coast guard DGPS, etc. Where the good do sub 1m and SBAS a little worse. Why would these not make absolute accuracy better? They do improve positioning. -- Björn [1] http://swepos.lmv.lm.se/natverksrtk/nat_postj_031012.gif ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
would you have a schematic of an interpolator with say 5ns resolution to share? Said, On interpolators: 1) Another good example is found in the op/svc manual for the SR620 -- which implements (in simple, non-ASIC, 1980's TTL/ECL technology) time interpolators that are good to 25 ps. Maybe overkill for you, but the design is worth understanding. 2) I recall this recent paper ($) also made good reading: Review of methods for time interval measurements with picosecond resolution http://www.iop.org/EJ/article/0026-1394/41/1/004/met4_1_004.pdf 3) Google for words like time interval interpolator and you will find quite a lot of sub-ns TIC ideas and products. Few seem to be cheap or available, though. 4) The wonderful US patent database also has design info on time interpolators (which is good for us home time-nuts but perhaps not so good for you if you plan to put any of this into your commercial GPSDO product). Since you're asking about time interval counters in relation to your GPSDO design, I want to tempt you to do some math first: 1) You know, or can measure, the ADEV of your GPS engine (M12M, or Res-T, etc.). 2) Same goes for the OCXO you have chosen. 3) Decide what GPSDO PLL time constant or PID parameters to use. 4) And then calculate what minimum TIC resolution is necessary for you to have to meet your product spec. For any tau, what limits GPSDO performance is a combination of GPS, OCXO, PLL, and TIC. You may in fact find your current 3.33 ns guess is fine. Or maybe you'll find that it's not near good enough and you are leaving good performance on the table. Let us know what you come up with. Perhaps it will shed light on why the classic HP GPSDO used a ~100 ps TIC. /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
In message [EMAIL PROTECTED], Tom Van Baak writes: My question exactly. I just talked to someone who uses SRS PRS10 Rb's, and apparently SRS admitted to him that the 10s ADEV is fairly bad on the PRS10 units (without any details). Part of the trouble is the 5 second polarity change of the C-field, but if you disable that, you ruin the long term numbers. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | 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 https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
From: Tom Van Baak [EMAIL PROTECTED] Subject: Re: [time-nuts] Some results of PRS10 and Trimble Resolution Date: Wed, 28 Jun 2006 23:01:42 -0700 Message-ID: [EMAIL PROTECTED] would you have a schematic of an interpolator with say 5ns resolution to share? Said, On interpolators: 1) Another good example is found in the op/svc manual for the SR620 -- which implements (in simple, non-ASIC, 1980's TTL/ECL technology) time interpolators that are good to 25 ps. Maybe overkill for you, but the design is worth understanding. Do you happend to have that (and other SR) service manual online somewhere? I haven't checked their line to closely. 2) I recall this recent paper ($) also made good reading: Review of methods for time interval measurements with picosecond resolution http://www.iop.org/EJ/article/0026-1394/41/1/004/met4_1_004.pdf It gives an overview of methods and seems to be in a little deeper on various forms of gate-based methods. 3) Google for words like time interval interpolator and you will find quite a lot of sub-ns TIC ideas and products. Few seem to be cheap or available, though. Indeed. It takes a bit of wading through tought. 4) The wonderful US patent database also has design info on time interpolators (which is good for us home time-nuts but perhaps not so good for you if you plan to put any of this into your commercial GPSDO product). There are several paths to finding what you want, one is to search for the manufactuers name (HP for instance). You may in fact find your current 3.33 ns guess is fine. Or maybe you'll find that it's not near good enough and you are leaving good performance on the table. Let us know what you come up with. Perhaps it will shed light on why the classic HP GPSDO used a ~100 ps TIC. BTW, I know it (the Z3801) has interpolators. I just plain forgot for a while. Cheers, Magnus ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
Thanks for the link to your report, Tom. I was wondering what this 2 sec bump in the ADEV was that everyone was talking about. The bump at 2 sec is just the loop tau of the main clock servo. The physics package is performing at the (fairly typical) level of about 2e-11/sqrt(tau). The OCXO is somewhat better than that. From Tom's plot it looks like the OCXO is good to about 3-4e-12 on tau=0.1 to 1 second. Because they've set the loop tau to about 1 seconds, the performance of the OCXO is degraded as it is steered to the atoms on that timescale. Never mind fancy RS232 commands, the real thing that sets SRS apart from the competition is that they use a good OCXO for the local oscillator, rather than a TCXO. This gives about 30 dB superior phase noise on the output and also contributes to the apparent bump described above. In more typical Rb standards, a TCXO is chosen with minimally adequate ADEV to support the 1 second stability goal, say 1-2e-11 on 0.1 to 1 second. Then, rather than a bump, you just see a flat line from 0.1 to 1 second which then turns downward onto the 1/sqrt(tau) line outside of the loop tau. Don't sweat the bump, you wouldn't see it if it weren't for the superior oscillator. Note that, given a superior OCXO, SRS could have chosen to set the loop tau longer. If the OCXO stays flat, they could set the loop tau way out at 10 seconds. Then the STS would look the OCXO, flat at 3-4e-12 from 0.1 to 10 seconds and then turn gracefully downward onto the 1/sqrt(tau) line somewhere around 10-20 seconds. The problem with this, of course, is that if something bad happens to the OCXO, such as a frequency hop, it takes 10-20 seconds to steer it out, rather than 1 second. Allowing a frequency hop, which could be as big as 1e-9, to persist for 10-20 seconds would destroy your medium and long-term stability records (and probably three or four other critical specifications). I don't recall whether SRS gives you access to the loop tau over RS232 but, if they do, you could try turning it up to 5-10 seconds just to see your 1 sec stability improve. Depending on the quality of the OCXO and the environment of your laboratory (temperature, vibration, power supply stability, etc.) you may be able to get away with running the rubidium at longer loop tau for months before you see an unpleasant frequency hop. Of course, the real culprit here is Timing Solutions and their damn fine TS5110. If it weren't for them, we wouldn't be having this conversation. -RL Robert Lutwak, Senior Scientist Symmetricom - Technology Realization Center 34 Tozer Rd. Beverly, MA 01915 (978) 232-1461 Voice [EMAIL PROTECTED] (Business) (978) 927-4099 FAX [EMAIL PROTECTED] (Personal) (339) 927-7896 Mobile - Original Message - From: Tom Van Baak [EMAIL PROTECTED] To: Discussion of precise time and frequency measurement time-nuts@febo.com Sent: Wednesday, June 28, 2006 11:16 PM Subject: Re: [time-nuts] Some results of PRS10 and Trimble Resolution T It's exactly peaked at 2s, it's got to be something digital like the DAC control of the OCXO etc. Any SRS designers among this group? BTW: while I cannot prove it anymore, I am pretty confident the PRS10 did the same thing while free-running (without 1PPS input). I'm thinking it looks exactly at 2s simply because your plot only has data points for only 1s, 2s, 3s, etc. A high ADEV bump near tau 2 s suggests a phase modulation around 4 s. If you look closely you can see this in the phase plots; it's somewhere between 4 and 5 seconds. It shows up in the power spectrum near 4.4 s. Many plots for you in this free-running PRS10 report: http://www.leapsecond.com/museum/prs10/ /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
Thanks for the link to your report, Tom. I was wondering what this 2 sec bump in the ADEV was that everyone was talking about. The bump at 2 sec is just the loop tau of the main clock servo. The physics package is performing at the (fairly typical) level of about 2e-11/sqrt(tau). The OCXO is somewhat better than that. From Tom's plot it looks like the OCXO is good to about 3-4e-12 on tau=0.1 to 1 second. Because they've set the loop tau to about 1 seconds, the performance of the OCXO is degraded as it is steered to the atoms on that timescale. ... Robert, Thanks for all your detailed comments. To support your points, here's an LPRO rubidium for contrast. http://www.leapsecond.com/museum/lpro/ True, there's no bump in the LPRO -- but when you compare the LPRO vs. PRS10 plots you see that the bump is evidence of a good thing for the PRS10. If nothing else everyone should view this one plot where 4 rubidiums are compared: http://www.leapsecond.com/images/4rb.gif /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
In message [EMAIL PROTECTED], Tom Van Baak writes: If nothing else everyone should view this one plot where 4 rubidiums are compared: http://www.leapsecond.com/images/4rb.gif Tom, can you try to measure the PRS10 with the magnetic switching turned off ? -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | 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 https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
Hello Robert, thanks so much for this insight! It all makes sense now. Cranking up the Tau to 10 or 20s via RS232 would be great, then I could use the PRS10 as a reference for the TSC meter to do ADEV testing on our GPSDO. As it stands now, I am measuring the PRS10 noise up to 10s or so not our GPSDO :) BTW: the two ADEV plots I sent to this mailing list were captured on the new TSC5120A (even damn finer machine than the TSC5110 :). I am glad that Tom's TSC5110 measurements are pretty much identical to mine in regards to the 1s-10s PRS10 ADEV. bye, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hello Tom, I had a conversation with Sam S. from TSC the other day, and he said that it's probably not possible to get 10ns GPS accuracy anyways due to the multipath issues, Ionospheric issues, antenna survey issues, thermal issues etc. I tend to agree with Sam; to get this kind of accuracy to UTC ([or better: GPS time] a documented 2ns within 300 hours as tested on the M12+ by Synergy/USNO) one needs to average over several hours, or even days as done by USNO. This is documented in many long term plots of GPS versus Cesium. Instantaneous drift in all of these plots seems to be +-5ns due to diurnal effects etc. Our units typically average the GPS 1PPS over 30 minutes, so having less than 3.33ns error on the 1PPS capture may not improve things much because while the error stays always at +-3.33ns it get's averaged over the measurement intervall. Of course this would be different if we use a Cesium 1PPS output to lock to. Also, one must ask the question: what is the correct time? Is the relative time between two GPS disciplined stations close to each other such as are used for CDMA base stations (some of the errors cancel out between the two stations' frequency differences) good enough? Certainly diurnal and ionospheric effects between the two stations will cancel if they see the same sat's. Sam had a point in that to get 1ns accuracy from your GPS, you must have a fantastic antenna not affected by environmental effects, and also know it's position to within 1 foot in all three dimensions! bye, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
Hi Tom, The reason I investigated this is that I needed a very good (2E-012 @ 1s) short term ADEV reference to measure the performance of our GPSDO (which is around 2E-012 at 1 second as you know :) on the TSC5120A. I found that I could use the MTI 260 reference from TSC for this purpose up to about 50s intervalls, then switch to the Rb above 50s. They both have their strengths depending on the measurement intervall. It would make sense to slave an MTI 260 OCXO to a PRS10 with a PLL lock time of about 20s. This would generate a very clean output - the Phase noise of the 260 is better than the PRS10, as well as the ADEV 20s would be improved. bye, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
the M12+'s pps jitter has a strong frequency component at a repetion rate of 2 s. This component has a peak to peak value of 2-4 ns and is almost unseen in the raw pps values. It is, however, clearly to be seen Where does that 2 seconds come from? Is that just one set of observations, or does it hold for most M12+s and most of the time? One of Said's graphs had a minor bump at roughly 2 seconds. I saw another graph with a small spike at 2 seconds, but I don't remember where. TVB's graphs show that the sawtooth period changes. I assume the sawtooth is the beat of the local OSC with the 1 PPS from GPS. I'm assuming the sawtooth graphs are derived from a greatly expanded temperature vs frequency graphs - similar to a Fresnel lens. The flat spots (suspension bridge) are not necessarily zero temperature vs frequency. They might be where the temperature is changing by an integral number of cycles per second per second. (It might be atmosphere or multipath that is changing rather than temperature.) So why does that cause a bump at 2 seconds? Why not 1 second or 5 seconds? It seems reasonable to filter out the 5 seconds, but I can't see how to filter out 1 second without also filtering out the 2 seconds. I think I'd expect a shoulder rather than a bump. It would fall off above 2 seconds and blend in below 2 seconds. -- The suespammers.org mail server is located in California. So are all my other mailboxes. Please do not send unsolicited bulk e-mail or unsolicited commercial e-mail to my suespammers.org address or any of my other addresses. These are my opinions, not necessarily my employer's. I hate spam. ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hmmm.. How fast could one update the internal cable delay value?
N9RKM
On Tue Jun 27 19:41 , 'Tom Van Baak' [EMAIL PROTECTED] sent:
Now, another way to handle this is to run the hardware
1 PPS output, with all its jitter, though a programmable
nanosecond-resolution digital delay generator. Each
second you use the firmware reported future negative
sawtooth as the programmed delay value for the next
1 PPS pulse. Tom Clark and I talked about this years
ago and Rick Hambly implemented this as an option
in his CNS Clock II ( [1]http://www.cnssys.com/ ).
Note this hardware trick only works for GPS OEM
boards that predict sawtooth error for the next 1PPS
(vs. reporting the sawtooth error of the previous 1PPS).
Motorola is OK here.
If you're tempted, see some programmable delay lines at:
[2]http://www.maxim-ic.com/products/timers/delay_lines.cfm
/tvb
[3]http://www.LeapSecond.com
___
time-nuts mailing list
[EMAIL PROTECTED]
[5]https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
References
1. file://localhost/tmp/parse.pl?redirect=http%3A%2F%2Fwww.cnssys.com%2F
2.
file://localhost/tmp/parse.pl?redirect=http%3A%2F%2Fwww.maxim-ic.com%2Fproducts%2Ftimers%2Fdelay_lines.cfm
3. file://localhost/tmp/parse.pl?redirect=http%3A%2F%2Fwww.LeapSecond.com
4. javascript:top.opencompose('time-nuts@febo.com','','','')
5.
file://localhost/tmp/parse.pl?redirect=https%3A%2F%2Fwww.febo.com%2Fcgi-bin%2Fmailman%2Flistinfo%2Ftime-nuts
___
time-nuts mailing list
time-nuts@febo.com
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
In message [EMAIL PROTECTED], Fred King writes: Hmmm.. How fast could one update the internal cable delay value? It doesn't help you, because the PPS pulse can only be generated at the same discreete places relative to the Oncore clock. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | 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 https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hello Tom, We approached the sawtooth correction on the software side: we sample the 1PPS with 6.66ns resolution in our new Fury GPSDO and apply software adjusted sawtooth correction (post-capture), and this yields an easily visible, very significant reduction in the 1PPS capture noise with the Motorola M12+ receivers. You can really see the difference second-to-second when the correction is turned on/off. This yields an average sampling quantization noise of +-3.33ns. Sounds good. Given that you still have many ns of jitter in the 1PPS output you might want to consider dividing your OCXO down to 1PPS and using that for your master 1PPS user output; you can reduce the output jitter 1000x that way. This is what HP did. The physical sawtooth on the 1PPS signal is actually helpfull in this setup as you mention since it dithers the LSB quantization noise, and thus actually improves the quantization resolution over time (with simple averaging low pass filtering of the captured data). This trick is called dithering in audio etc. Loosely related to that is DAC dithering which was mentioned here a while ago. The one difference, though, is that the dithering you get with a 1PPS output is not always random enough to guarantee that you get a clean mean after N seconds. Sometimes it works well, but you can also get very unlucky (e.g., the hanging bridge effect in that m12 web page). The longer you average the less this is a problem; 10 or 30 minutes makes it a non-issue. But with much shorter averaging times you can see how it might cause trouble. Using delay lines may be tricky and expensive, they usually are temperature sensitive, and only yield good results if the capture of the 1PPS is done in fast enough (10ns capture resolution). But if your capture is fast enough anyways, there is no need to use a delay line since the correction can be done in software. The 8-bit 1 ns resolution part from Dallas/Maxim looked pretty cheap to me. $5, I think. Yeah, you'd have to check the tempco but also weigh it against the tempco of many other key components of a GPSDO. Delay lines also have another disadvantage: the 1PPS correction from the GPS can be positive or negative in time, so in theory you would need a negative time delay I wouldn't call it a disadvantage; you just advance the GPS tick by the right amount to compensate. The same way you compensate for antenna delay, antenna cable delay, GPS engine delay, 1 PPS output cable and 74AC buffer delays, etc. (Einstein would be happy :). So for the delay line to work correctly, you have to set the 0ns delay tap equivalent to the most negative pulse, all other pulses will incur more than 0ns delay. You are now effectively delaying the average 1PPS by 1/2 the spread of the pulses. Thus the output of the delay line is always late (by about 30ns for commercial receivers) on average for standard GPS receivers. This error Right, actually any delay greater than half the sawtooth jitter would work. Using half the range of an 8-bit delay generator might be nice too; that allows you to give the user a fine adjustment feature in the range +/- 128 ns with 1 ns resolution. will have to be subtracted later. One trick when using Motorola GPS's is to set the cable delay to an additional 30ns to make the GPS receiver itself compensate for this delay by issuing the 1PPS output 30ns too fast. Yes, the antenna delay parameter is the way to do this. An interesting effect of sampling at +-3.33ns is that the GPS errors such as multipath, atmospheric, and GPS crystal temperature related issues become clearly visible with this kind of resolution... Note the 58503 and Z3801A receivers (now ten years old) have 100 ps resolution. But I'm guessing your 3.33 ns detector is cheaper than HP's! /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hi Said, A few questions regarding Delay line vs. Software correction: Before asking the question let me check if I understood you correctly. Noting that you have a resolution of 6.66ns, I presume you are running at a 150MHz clock speed. In other words you delay the 1PPS by some integer number of clock cycles (e.g. n times 6.66ns). This is done by rounding the sawtooth error provided by the GPS (typically something between -128ns and +128ns for the M12+T) to the nearest 6.66ns. Now the random sawtooth jitter of the 1PPS signal and the slight drift on your 150MHz clock adds out of phase (most of the time), or put otherwise - it is sometimes less, and sometimes more - which averages to a nice final resolution of 3.33ns. Now, the question: I'm primarily interested in relative time stability between two or more identical GPSDO devices. If I use a delay line it will add an absolute offset (+- a few ns) plus some jitter (the Maxim/Dallas datasheets doesn't specify the added jitter) to each GPSDO. The absolute offset doesn't really worry me since I'm only interested in relative stability (Maybe I'm to calibrate the +- few ns out first using the method you suggested e.g. cable delay function). It is the jitter that worries me. If I try correct for a 2ns sawtooth error with the delay adding jitter of that same order it really defies the purpose. On the other hand, if I delay it in software (which I had in mind in the first place) I have to go for a high frequency clock oscillator. Nothing wrong with that, I guess you could go as high as your processor can handle. However, the EMI generated (radiated EM, ground bounce, power supply noise etc.) by such a high speed clock worries me. For digital stuff it is not as bad I guess, but the DACs, VRefs and FS will suffer due to this. My gut tells me that if the EMI is going to be a problem I would rather go for the delay line option with a much lower clock speed say 64MHz (capture resolution of 15.625ns). However, no I'm back where I started since now I make a new sawtooth error of 15.625ns! So, the only option is to go for a higher clock speed. What are your experiences with the resulting EMI? Is it a problem? How do you combat it? Regards, Stephan Sandenbergh - Said wrote: Message: 10 Date: Tue, 27 Jun 2006 21:38:02 EDT From: [EMAIL PROTECTED] Subject: Re: [time-nuts] Some results of PRS10 and Trimble Resolution To: [EMAIL PROTECTED], time-nuts@febo.com Message-ID: [EMAIL PROTECTED] Content-Type: text/plain; charset=US-ASCII Hello Tom, We approached the sawtooth correction on the software side: we sample the 1PPS with 6.66ns resolution in our new Fury GPSDO and apply software adjusted sawtooth correction (post-capture), and this yields an easily visible, very significant reduction in the 1PPS capture noise with the Motorola M12+ receivers. You can really see the difference second-to-second when the correction is turned on/off. This yields an average sampling quantization noise of +-3.33ns. The physical sawtooth on the 1PPS signal is actually helpfull in this setup as you mention since it dithers the LSB quantization noise, and thus actually improves the quantization resolution over time (with simple averaging low pass filtering of the captured data). This trick is called dithering in audio etc. Using delay lines may be tricky and expensive, they usually are temperature sensitive, and only yield good results if the capture of the 1PPS is done in fast enough (10ns capture resolution). But if your capture is fast enough anyways, there is no need to use a delay line since the correction can be done in software. Delay lines also have another disadvantage: the 1PPS correction from the GPS can be positive or negative in time, so in theory you would need a negative time delay (Einstein would be happy :). So for the delay line to work correctly, you have to set the 0ns delay tap equivalent to the most negative pulse, all other pulses will incur more than 0ns delay. You are now effectively delaying the average 1PPS by 1/2 the spread of the pulses. Thus the output of the delay line is always late (by about 30ns for commercial receivers) on average for standard GPS receivers. This error will have to be subtracted later. One trick when using Motorola GPS's is to set the cable delay to an additional 30ns to make the GPS receiver itself compensate for this delay by issuing the 1PPS output 30ns too fast. An interesting effect of sampling at +-3.33ns is that the GPS errors such as multipath, atmospheric, and GPS crystal temperature related issues become clearly visible with this kind of resolution... Bye, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
From: Stephan Sandenbergh [EMAIL PROTECTED] Subject: Re: [time-nuts] Some results of PRS10 and Trimble Resolution Date: Wed, 28 Jun 2006 18:22:58 +0200 Message-ID: [EMAIL PROTECTED] Stephan, So, the only option is to go for a higher clock speed. What are your experiences with the resulting EMI? Is it a problem? How do you combat it? You interpolate using analogue interpolators. Take a look into the HP5335A counter for instance. It happilly acheives 1 ns single-shot resolution but it counts at a mear 10 MHz clock. The analog interpolators of the HP5335A acts like pulse-stretchers and will make the pulse 200 times longer. It will see a pulse which is 100-200 ns long, but will convert it into a pulse which is 20-40 us long, which turns into 200 to 400 counts long in its 10 MHz clock. These interpolator counters is actually in normal TTL in the HP5335A, and is actually only 8-bit, so they will wrap-around, but the software will de-wrap it since there is no ambiguity in reality. The cost for the interpolators themselfs is not very high, not compared to all the fancy stuff around them anyway. It will cost some additional logic for error-pulse generation and also for the additional counters (you usually have one start and one stop counter). Another cost is the longer conversiontime. However, if you work on PPS clocks, then a conversion-time of 40 us is very quick anyway and not much of a problem. If you scale this design a little, getting 200 ps or 100 ps should be reachable without too much of an effort. Infact, I suspect that this is exactly what HP did in the Z3801. Their measurement FPGA was really not up to any magic in speeds. Soo, if you want to improve your measurement resolution only, it is simple. In a very similar fashion, you can actually do the revese, in order to compensate for the time error of a signal in your clock. However, thus designs you can buy canned if you don't want to DIY. A DAC working as a programable current source (preferably through a current mirror) and a cap with a buffer and comparator aught to do it. Cheers, Magnus ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hmmm.. How fast could one update the internal cable delay value? N9RKM That won't help. The antenna delay is handled in software, and the receiver still has the same clock granularity for placing the pulse. -- ch ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hi Stephan, as CH said, the antenna delay is a static setting in the M12+, it's not designed to be updated every second. I think it is only usefull in cancelling out systemic delays in the setup such as TIA, delay-line, cabling etc. Yes, we do run our sampler at 150MHz, with a resulting +-3.33ns resolution uncertainty. I know of one company making Military GPSDO's, and they use 5MHz sampling, so they have a 200ns window of uncertainty, seems to be enough for their application. Using an interpolator with a slow basic clock as Magnus suggests is probably the easiest way to prevent EMI issues, but it has some disadvantages such as: * You need to design a fairly tricky high-linearity charge pump to make it work well, using high-quality components such as Polyester caps, low INL/DNL/Tempco ADC's and ADC reference, or high quality comparators and sample-and-holds, etc to make it insensitive to temperature changes etc. This is essentially an analog design, with voltages being captured by an ADC, then converted to time steps in software. * The interpolator needs to be calibrated to give good results, especially if the charge pump is not very linear. This is also needed due to temp changes, as well as the ADC's errors such as the INL/DNL etc. This could be done automatically, but does require a bit of circuitry and know how. The advantage of the interpolator is that it can have a very high resolution (ps) if designed properly. I have a Wavecrest interpolator board at home that has 10ps resolution, and it is bigger than one of the old full-size AT IBM motherboards... its tricky circuitry to design and make it work well. One interesting fact is that Wavecrest uses actual rigid coax cables wound up in loops to create delays on that board! They do suggest to calibrate the system every 24 hours, or if there is a 5 Deg C change in temp. Using the digital approach, you only need two parts: a fast PLD/FPGA, and an external clock source (10MHz for example if the PLD has a PLL). Depending on the PLD, you could get one that has an internal PLL running at 500MHz or even faster from your external 10MHz. These fast signals are only inside the FPLD/FPGA, so it's easy to shield the circuit to prevent EMI. NEC-Tokin makes ferrite shielding materials for exactly this purpose (self-adhesive, you can glue them onto the PLD). We have an option to place the entire PLD circuit inside a Faraday metal-shield on our Fury GPSDO, no issue with EMI in that case. bye, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
In message [EMAIL PROTECTED], [EMAIL PROTECTED] writes: Using an interpolator with a slow basic clock as Magnus suggests is probably the easiest way to prevent EMI issues, but it has some disadvantages such as: * You need to design a fairly tricky [...] I actually benchmarked the on in the PRS10 by feeding it a 1PPS generated from another Rb that was deliberately adjust approx 1e-10 low. I can't seem to find the results anymore, but if any of you want to reproduce it, it was a pretty trivial setup. As I remember it, the standard deviation was approx 3nsec, and I seeing some missing codes, Ie: readings that just didn't happen in practice. Testing any other such design is similarly trivial. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | 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 https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
From: [EMAIL PROTECTED] Subject: Re: [time-nuts] Some results of PRS10 and Trimble Resolution Date: Wed, 28 Jun 2006 16:19:04 EDT Message-ID: [EMAIL PROTECTED] Hi Said, Using an interpolator with a slow basic clock as Magnus suggests is probably the easiest way to prevent EMI issues, but it has some disadvantages such as: * You need to design a fairly tricky high-linearity charge pump to make it work well, using high-quality components such as Polyester caps, low INL/DNL/Tempco ADC's and ADC reference, or high quality comparators and sample-and-holds, etc to make it insensitive to temperature changes etc. This is essentially an analog design, with voltages being captured by an ADC, then converted to time steps in software. * The interpolator needs to be calibrated to give good results, especially if the charge pump is not very linear. This is also needed due to temp changes, as well as the ADC's errors such as the INL/DNL etc. This could be done automatically, but does require a bit of circuitry and know how. You run into more of these problems if you go for larger gains. If you go for a modest gain of 10-200 you have still a fairly easy design-effort and besides, running a clock around 100 MHz isn't as large problem as it used to be. With a x100 scaling the 10 ns resolution has become 100 ps. For such a interpolator you can get fairly good result just by feeding it a short-pulse/long-pulse training (1 or 2 cycles - i.e. 10 ns and 20 ns). You can then use that for either compensation or better yeat - trimming of the currents. The larger gain, the more important the linearity becomes. You can either fight it by increase linearity or you can fight it by trimming up the scale and linearize it through a look-up-table. This however requires a method to produce various forms of time intervals with known occurence, but it has been done. The advantage of the interpolator is that it can have a very high resolution (ps) if designed properly. I have a Wavecrest interpolator board at home that has 10ps resolution, and it is bigger than one of the old full-size AT IBM motherboards... its tricky circuitry to design and make it work well. One interesting fact is that Wavecrest uses actual rigid coax cables wound up in loops to create delays on that board! They do suggest to calibrate the system every 24 hours, or if there is a 5 Deg C change in temp. As we push digital up in clock, the timing resolution which used to be alot of black magic is becoming much more available. Look at the HP5335A, it used 10 MHz and a fairly simple interpolator design. Using the same basic design but running at 100 MHz (which comes cheap today) should give you 100 ps without too much of a head-ache. Infact, in the HP5335A they lost some precission in the way they treated data, so they only said 1 ns when they infact measured with 500 ps resolution. Using the digital approach, you only need two parts: a fast PLD/FPGA, and an external clock source (10MHz for example if the PLD has a PLL). Actually, today you use FPGA and interpolators together. You can infact get 100 ps single-shot resolution straight out of a FPGA, but you will have to spend some time on the EMI issues. Depending on the PLD, you could get one that has an internal PLL running at 500MHz or even faster from your external 10MHz. These fast signals are only inside the FPLD/FPGA, so it's easy to shield the circuit to prevent EMI. You still need to care about propper decoupling caps. Some (actually a certain) FPGA manufactuers don't have a real PLL, but have relied solely on DLLs, but those are not as timing-clean as a real PLL can give you, so the internal jitter can actually be quite high. Cheers, Magnus ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hi Magnus, would you have a schematic of an interpolator with say 5ns resolution to share? Thanks, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
HI Hmurray, you said: One of Said's graphs had a minor bump at roughly 2 seconds. I saw another graph with a small spike at 2 seconds, but I don't remember where. I thought these bumps were rather considerable, at about 1 to 2E-011, about 8 - 10x what our GPS-locked OCXO's do. So why does that cause a bump at 2 seconds? Why not 1 second or 5 seconds? It seems reasonable to filter out the 5 seconds, but I can't see how to filter out 1 second without also filtering out the 2 seconds. I think I'd expect a shoulder rather than a bump. It would fall off above 2 seconds and blend in below 2 seconds. My question exactly. I just talked to someone who uses SRS PRS10 Rb's, and apparently SRS admitted to him that the 10s ADEV is fairly bad on the PRS10 units (without any details). It's exactly peaked at 2s, it's got to be something digital like the DAC control of the OCXO etc. Any SRS designers among this group? BTW: while I cannot prove it anymore, I am pretty confident the PRS10 did the same thing while free-running (without 1PPS input). bye, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
From: [EMAIL PROTECTED] Subject: Re: [time-nuts] Some results of PRS10 and Trimble Resolution Date: Wed, 28 Jun 2006 18:39:44 EDT Message-ID: [EMAIL PROTECTED] Hi Magnus, Said, would you have a schematic of an interpolator with say 5ns resolution to share? Go and look up the HP5335A operating and service manual on Agilents site (I have it in original hardprint BTW). There are a few pages in it which quite nicely explain how it works in theory and all. You want to check the pages 8-29 to 8-31, 8-73 to 8-76 and the schematic on page 8-111. The HP5335A has a 1 ns of resolution according to the specs, so I think it matches your question. I also have an HP5335A on the lab-bench. It take some effort to have it under GPIB-control thought, so I can't do any massive tests at this time, otherwise I'm sure I could do some if needed. The HP5335A dates back to 1980 and was certainly not their highest profile counter, but a good and then modern all-around counter for normal laboratory use. The input triggering is however a bit strange. You can crank it up into 12 digit measures and the gate time pot is among the things which still makes it somewhat of a favorite over some modern counters (HP53131 for instance). Hope it helps and that you find the information you needed. Cheers, Magnus ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Thanks Magnus, BTW: how do you get the 5335 to do 12 digits? I can do it on my 5334, but haven't found a way to get the 5335 to do it. thanks, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
My question exactly. I just talked to someone who uses SRS PRS10 Rb's, and apparently SRS admitted to him that the 10s ADEV is fairly bad on the PRS10 units (without any details). I very much agree with his report, although it's not like SRS is trying to hide anything. Look again at the PRS10 specs and nice composite ADEV plot at: http://www.thinksrs.com/products/PRS10.htm and see 2e-11 @1s; 1e-11 @10s; 2e-12 @100s So I'm curious what sort of short-term performance you or others are _expecting_ out of a PRS10, or any low-cost, compact, telecom-market Rb standard for that matter? True, Rb have low drift and Cs have no drift but if you want really good short-term stability use an OCXO, or at least a lab-grade rubidium (like an hp 5065A, about ~5x better than PRS10 at 1s). For a GPSDO, Rb has very good price and holdover performance relative to cesium or quartz. And you can use a much longer GPS time constant with Rb or Cs. However, if long-term holdover is not a key spec for you; if you're after exceptional short-term performance, then a high-end OCXO is probably a better choice than a low-end Rb. What makes the PRS10 a really nice unit is the combination of good performance and a huge set of RS232-accessible irresistible functionality. /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
[time-nuts] Some results of PRS10 and Trimble Resolution
You wrote I still don't really understand why the PRS10 Rubidium has such a bad 2s Allan Deviation with a 7-hour 1PPS filter time, also considering this kind of nice low Phase noise. My thoughts, This is a guess but I would look at the loop filter bandwidth that locks the OCXO to the Rubidium physics package. Outside the loop filter bandwidth the stability will be determined by the OCXO inside the loop filter bandwidth stability would be a function of the physics package, and at the cross over point you get noise contributions from both the OCXO and physics package, and the phase detector. If the loop is DSP based then you can get non-linear quanitzation effects from the DSP as well all the phased detector noise, not usually a factor for average electronics but 10E-11 is not average. Regards, T.R. WR4T - Sneak preview the all-new Yahoo.com. It's not radically different. Just radically better. ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
I will add my previous post only applies if the 2s ADEV is poor while you running open loop with respect to a GPS 1pps signal. Otherwise I think Paul has the more likely answer. Cheers, T.R. WR4T - Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls. Great rates starting at 1¢/min. ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hello Poul-Henning, you wrote: Newer versions of the PRS10 firmware allows you to set the PPS input offset, and with a small program you can transmit the negative sawtooth correction from the GPS to the PRS and that solved the problem for me. How much lower were you able to get the 2s ADEV by applying sawtooth correction to your PRS10 unit? Thanks, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Tom
In the series of sawtooth displays are these all the same M12 and if
so how did it happen that you have this result in a short time?
Thanks.
Fred
N9RKM
On Tue Jun 27 15:45 , 'Tom Van Baak' [EMAIL PROTECTED] sent:
.
For example, a 25 MHz oscillator in a GPS engine allows
the processor to pick, each second, which one of 25 million
edges it wants to be the official 1 PPS edge for that second.
Now 25 MHz frequency is 40 ns period so that gives a 1 PPS
granularity of +/- 20 ns in this example. See:
[1]http://www.leapsecond.com/pages/m12/sawtooth.htm
If that makes sense, then what's sawtooth correction?
...
In short, the sawtooth correction describes, to the best
guess of the receiver, how early or late the hardware
1 PPS was.
/tvb
___
time-nuts mailing list
[EMAIL PROTECTED]
[3]https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
References
1.
file://localhost/tmp/parse.pl?redirect=http%3A%2F%2Fwww.leapsecond.com%2Fpages%2Fm12%2Fsawtooth.htm
2. javascript:top.opencompose('time-nuts@febo.com','','','')
3.
file://localhost/tmp/parse.pl?redirect=https%3A%2F%2Fwww.febo.com%2Fcgi-bin%2Fmailman%2Flistinfo%2Ftime-nuts
___
time-nuts mailing list
time-nuts@febo.com
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
http://www.leapsecond.com/pages/m12/sawtooth.htm Tom In the series of sawtooth displays are these all the same M12 and if so how did it happen that you have this result in a short time? Thanks. Fred Yes, it's the same M12; other M12's are similar. VP's have the same thing but the peak-to-peak is larger. This is all normal and the sawtooth does not indicate anything is wrong. In fact, you can make a good case that the noise of a 1 PPS sawtooth is a good thing in many cases; as we found with Tom Clark's original TAC, it permits simple time interval averaging to give a more precise result in spite of the 1 PPS quantization. Not sure what you mean by a short time. The web page dates back to Nov 2004, I think. But the sawtooth in this and other receivers show up in a matter of seconds so it's not like you have to take a lot of data over a long time to see it (note that most of the plots on that page cover only 2 minutes of elapsed time). /tvb ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution
Hello Tom, We approached the sawtooth correction on the software side: we sample the 1PPS with 6.66ns resolution in our new Fury GPSDO and apply software adjusted sawtooth correction (post-capture), and this yields an easily visible, very significant reduction in the 1PPS capture noise with the Motorola M12+ receivers. You can really see the difference second-to-second when the correction is turned on/off. This yields an average sampling quantization noise of +-3.33ns. The physical sawtooth on the 1PPS signal is actually helpfull in this setup as you mention since it dithers the LSB quantization noise, and thus actually improves the quantization resolution over time (with simple averaging low pass filtering of the captured data). This trick is called dithering in audio etc. Using delay lines may be tricky and expensive, they usually are temperature sensitive, and only yield good results if the capture of the 1PPS is done in fast enough (10ns capture resolution). But if your capture is fast enough anyways, there is no need to use a delay line since the correction can be done in software. Delay lines also have another disadvantage: the 1PPS correction from the GPS can be positive or negative in time, so in theory you would need a negative time delay (Einstein would be happy :). So for the delay line to work correctly, you have to set the 0ns delay tap equivalent to the most negative pulse, all other pulses will incur more than 0ns delay. You are now effectively delaying the average 1PPS by 1/2 the spread of the pulses. Thus the output of the delay line is always late (by about 30ns for commercial receivers) on average for standard GPS receivers. This error will have to be subtracted later. One trick when using Motorola GPS's is to set the cable delay to an additional 30ns to make the GPS receiver itself compensate for this delay by issuing the 1PPS output 30ns too fast. An interesting effect of sampling at +-3.33ns is that the GPS errors such as multipath, atmospheric, and GPS crystal temperature related issues become clearly visible with this kind of resolution... Bye, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
[time-nuts] Some results of PRS10 and Trimble Resolution T
Hello, Hereby some stability results of a PRS10 steered by a Trimble Resolution T GPS. The reference was UTC(VSL) and the filter was set op PT5. Mod.Allan: Tau = 1 s UTC(VSL)-PRS10 6 1.20017E-11 12 4.83028E-12 24 2.26147E-12 48 1.3338E-12 96 9.78699E-13 192 8.39936E-13 384 7.8562E-13 768 7.1999E-13 1536 6.13279E-13 3072 4.92398E-13 6144 3.54263E-13 12288 2.04454E-13 24576 1.46744E-13 49152 6.79256E-14 57600 4.99072E-14 115200 2.83436E-14 230400 1.89469E-14 460800 5.642E-15 It seems a good combination for frequency. If there is intrest I am able to show also the stability of the GPS alone. Best regards, Erik Kroon ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
Hello Erik, I recently measured my PRS10 as well. You can see a big hump (about 1.6E-011) at the 2s measurement intervall. While this is within spec, does anyone know why the unit is so noisy at that particular intervall? Is there a filter setting that would improve this? Could this be power supply injected noise? I have set PT to 7, driving the unit with an Motorola M12+. I measured the ADEV with the new TSC5120A. Other oscillators I have measure at around 1E-012 at 2s, so it's not a setup problem. Thanks, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
Re: [time-nuts] Some results of PRS10 and Trimble Resolution T
Hallo Said, the M12+'s pps jitter has a strong frequency component at a repetion rate of 2 s. This component has a peak to peak value of 2-4 ns and is almost unseen in the raw pps values. It is, however, clearly to be seen if you manage to compute the sawtooth corrected pps values. What you see in the PRS10 output is simply a feedthrough of this frequency component, even with natural pll time constants of some hours or so. While i do not own a PRS10 myself, i use exactly the same software algorithms as the PRS10 including the pre-filter in my GPSDO (FRK-L or LPRO or FTS-1200 or HP10811 against M12+) and i come to very close results regardless which local oscillator i use. The bump gets smaller with higher pll time constants but you will no get rid of it completely. For a quick check, remove the M12+ pps and let the PRS10 run on its own, you should see no bump in this case. Regards Ulrich, DF6JB -Ursprüngliche Nachricht- Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Im Auftrag von [EMAIL PROTECTED] Gesendet: Montag, 26. Juni 2006 19:44 An: time-nuts@febo.com Betreff: Re: [time-nuts] Some results of PRS10 and Trimble Resolution T Hello Erik, I recently measured my PRS10 as well. You can see a big hump (about 1.6E-011) at the 2s measurement intervall. While this is within spec, does anyone know why the unit is so noisy at that particular intervall? Is there a filter setting that would improve this? Could this be power supply injected noise? I have set PT to 7, driving the unit with an Motorola M12+. I measured the ADEV with the new TSC5120A. Other oscillators I have measure at around 1E-012 at 2s, so it's not a setup problem. Thanks, Said ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi- bin/mailman/listinfo/time-nuts ___ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
