Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
[Context is second PPS input.] brayn...@gmail.com said: > I think it depends if they are phase aligned or offset. If the interrupts > are occurring at the same time, I would anticipate increased jitter. If they > are offset I would not expect any negative impact. > A simple way to demonstrate the effect would be to feed the same pulse to > two different PPS pins. The ISRs will compete when both interrupts fire. If you put a long cable between them, you can change the order of the interrupts by swapping things. That may need a real driver and good termination. -- These are my opinions. I hate spam. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
I don't think so - should just work Did you already have ntpd running when you attached? And do you have the full PPS module stack including ldisc? It works on every platform I got. :-) As to better than Internet time - I cannot get it closer than 2 or 3 ms. I have MANY internet time sources whose offset is not close to that bad On Wednesday, December 10, 2014, Paul wrote: > On Wed, Dec 10, 2014 at 12:17 PM, > > wrote: > > > Quite simply - if you either execute ldattach pps /dev/ttyUSBx, or are > > running GPSD (recommended) it will bind the USB appropriately to a ppsapi > > instance. > > > > I can get a /dev/ppsN but ppstest says time-out. I saw some hints that the > USB DCD > > might depend on the chipset . I don't like gpsd so I try to avoid it but I > might try it. > > Thanks. > > > Do not do it. You will not approve of the results. > > > > I don't intend to use it in production. > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
So just some comments based on personal experience with the raspi. Your experience may vary. I think it depends if they are phase aligned or offset. If the interrupts are occurring at the same time, I would anticipate increased jitter. If they are offset I would not expect any negative impact. A simple way to demonstrate the effect would be to feed the same pulse to two different PPS pins. The ISRs will compete when both interrupts fire. It's trivial to add more PPS inputs to the raspi. I'd be curious if anybody has experience to share with multicore setups with interrupts pinned to different cores. - Brian > On Dec 10, 2014, at 10:55 AM, Laszlo Hanyecz wrote: > > Time-nuts, > > Is it a bad idea to have more than one PPS source on a single machine? Would > this cause additional jitter when trying to compare the timestamps on two > sources? I understand that USB can't deliver a real PPS, but what about > using onboard serial ports or a PCI card? > > Thanks, > Laszlo > > > -Original Message- > From: time-nuts [mailto:time-nuts-boun...@febo.com] On Behalf Of Paul > Sent: Wednesday, December 10, 2014 18:43 > To: Discussion of precise time and frequency measurement > Subject: Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy? > >> On Wed, Dec 10, 2014 at 12:17 PM, wrote: >> >> Quite simply - if you either execute ldattach pps /dev/ttyUSBx, or are >> running GPSD (recommended) it will bind the USB appropriately to a >> ppsapi instance. > > I can get a /dev/ppsN but ppstest says time-out. I saw some hints that the > USB DCD > > might depend on the chipset . I don't like gpsd so I try to avoid it but I > might try it. > > Thanks. > >> Do not do it. You will not approve of the results. > > I don't intend to use it in production. > ___ > time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
> Le 10 déc. 2014 à 19:55, Laszlo Hanyecz a écrit : > > Time-nuts, > > Is it a bad idea to have more than one PPS source on a single machine? Would > this cause additional jitter when trying to compare the timestamps on two > sources? I understand that USB can't deliver a real PPS, but what about > using onboard serial ports or a PCI card? > You would have to test I guess but the potential is there for an issue. It can be circumvented by configuring a 1PPS offset if the receiver allows it , or configuring a reasonable cable delay in the receiver so that the pulses are not simultaneous . The latter possibility exists for the Venus 8 receiver , but I cannot find a command to force a 1PPS offset. > Thanks, > Laszlo > > > -Original Message- > From: time-nuts [mailto:time-nuts-boun...@febo.com] On Behalf Of Paul > Sent: Wednesday, December 10, 2014 18:43 > To: Discussion of precise time and frequency measurement > Subject: Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy? > > On Wed, Dec 10, 2014 at 12:17 PM, wrote: > >> Quite simply - if you either execute ldattach pps /dev/ttyUSBx, or are >> running GPSD (recommended) it will bind the USB appropriately to a >> ppsapi instance. >> > > I can get a /dev/ppsN but ppstest says time-out. I saw some hints that the > USB DCD > > might depend on the chipset . I don't like gpsd so I try to avoid it but I > might try it. > > Thanks. > >> Do not do it. You will not approve of the results. >> > > I don't intend to use it in production. > ___ > time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
Time-nuts, Is it a bad idea to have more than one PPS source on a single machine? Would this cause additional jitter when trying to compare the timestamps on two sources? I understand that USB can't deliver a real PPS, but what about using onboard serial ports or a PCI card? Thanks, Laszlo -Original Message- From: time-nuts [mailto:time-nuts-boun...@febo.com] On Behalf Of Paul Sent: Wednesday, December 10, 2014 18:43 To: Discussion of precise time and frequency measurement Subject: Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy? On Wed, Dec 10, 2014 at 12:17 PM, wrote: > Quite simply - if you either execute ldattach pps /dev/ttyUSBx, or are > running GPSD (recommended) it will bind the USB appropriately to a > ppsapi instance. > I can get a /dev/ppsN but ppstest says time-out. I saw some hints that the USB DCD might depend on the chipset . I don't like gpsd so I try to avoid it but I might try it. Thanks. > Do not do it. You will not approve of the results. > I don't intend to use it in production. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
On Wed, Dec 10, 2014 at 3:37 PM, Hal Murray wrote: > > Linux has two APIs to PPS. > I'm using gpsd 3.9 which uses PPS-API if available. It pushes LD 18 and just like using ldattach there's no output. Both result in: [ 7668.796593] pps pps1: new PPS source usbserial0 [ 7668.796624] pps pps1: source "/dev/ttyUSB0" added and the creation of (in this case) /sys/devices/virtual/pps/pps1. There are no events though. ...Support for USB serial > devices is not so good. > Yes, I read something that suggests that not all chipsets are supported by the DCD patch to the USB serial driver. I think I've lost interest again. I'll just run the PPS into a gpio pin. But if anyone does get USB-DCD working with Linux I'd appreciate any details. Thanks. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
david-tay...@blueyonder.co.uk said: > But having the PPS on the DCD over USB is not as useless as you might first > think, because in tests here using the DCD/PPS over USB produced better > results with NTP than an internet connection alone. It is worth checking - > your results may differ. The Ethernet is also on USB so it will have the USB jitter as well as any jitter from the network. Even if the remote NTP system is perfect (or at least very good relative to the R-PI), I'd expect a local PPS via USB to be slightly better than internet time. tic-...@bodosom.net said: > I'll admit, to my shame, that I have yet to deduce how to use USB provided > DCD for PPS. I've looked, really I have but to no avail. Linux has two APIs to PPS. gpsd uses TIOCMIWAIT, an ioctl that lets a userland program wait for the PPS/DCD change. You can feed that to ntpd via SHM. The ATOM and NMEA drivers in ntpd use the API described in RFC 2783. It's in sys/timepps.h On Fedora, it comes from the ps-tools-devel package. This needs a running ldattach 18 /dev/xxx for each PPS source. The interrupt driver grabs a timestamp so the timing accuracy should avoid most of the jitter associated with getting to userland. I think most real serial ports have support for both. Support for USB serial devices is not so good. I haven't checked recently. I think TIOCMIWAIT support is generally better. -- These are my opinions. I hate spam. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
On Wed, Dec 10, 2014 at 12:17 PM, wrote: > Quite simply - if you either execute ldattach pps /dev/ttyUSBx, or are > running GPSD (recommended) it will bind the USB appropriately to a ppsapi > instance. > I can get a /dev/ppsN but ppstest says time-out. I saw some hints that the USB DCD might depend on the chipset . I don't like gpsd so I try to avoid it but I might try it. Thanks. > Do not do it. You will not approve of the results. > I don't intend to use it in production. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
Don’t do it. 😊 Quite simply - if you either execute ldattach pps /dev/ttyUSBx, or are running GPSD (recommended) it will bind the USB appropriately to a ppsapi instance. Do not do it. You will not approve of the results. From: Paul Sent: Wednesday, December 10, 2014 9:03 AM To: Discussion of precise time and frequency measurement On Wed, Dec 10, 2014 at 11:41 AM, David J Taylor < david-tay...@blueyonder.co.uk> wrote: > But having the PPS on the DCD over USB > I'll admit, to my shame, that I have yet to deduce how to use USB provided DCD for PPS. I've looked, really I have but to no avail. I see that the code has been in the driver circa Linux 2.6 but I'm just not making the "connection". ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
I'll admit, to my shame, that I have yet to deduce how to use USB provided DCD for PPS. I've looked, really I have but to no avail. I see that the code has been in the driver circa Linux 2.6 but I'm just not making the "connection". Can't comment on Linux, but in Windows the COM port driver provides an event when the DCD line changes state, and Dave Hart's code stores that value to timestamp the NMEA data when it arrives. Does Linux not support an interrupt from a virtual COM port? Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
On Wed, Dec 10, 2014 at 11:41 AM, David J Taylor < david-tay...@blueyonder.co.uk> wrote: > But having the PPS on the DCD over USB > I'll admit, to my shame, that I have yet to deduce how to use USB provided DCD for PPS. I've looked, really I have but to no avail. I see that the code has been in the driver circa Linux 2.6 but I'm just not making the "connection". ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
Gents- I don’t know if everyone is aware, but the USB PPS out is basically useless on both my units - at least 200us off if not more. Make sure you’re using the PPS OUT and not trying to measure on the DCD of the USB int, as is easy to be drawn to do. NS = Neil, Thanks for that reminder. I've been measuring both on the co-ax output pin and on the debug header (with the same or very similar results). The difference between my two GPSDO has settled at 170-180 ns. I suppose I will now need a third! But having the PPS on the DCD over USB is not as useless as you might first think, because in tests here using the DCD/PPS over USB produced better results with NTP than an internet connection alone. It is worth checking - your results may differ. Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
Gents- I don’t know if everyone is aware, but the USB PPS out is basically useless on both my units - at least 200us off if not more. Make sure you’re using the PPS OUT and not trying to measure on the DCD of the USB int, as is easy to be drawn to do. NS From: David J Taylor Sent: Tuesday, December 09, 2014 7:35 AM To: Discussion of precise time and frequency measurement From: Dave Martindale In my case, the LTE-Lite had been operating for at least a week before I made my "accumulate mode" measurement, and the Thunderbolt had been operating for at least a month. But both antennas were in poor locations - not bad enough to lose lock any time I was watching, but nowhere close to a clear view of most of the sky. I never saw the 1 PPS disappear while I was watching it. I wonder if your LTE-Lite ever finished its survey and switched into 1D/position hold mode? A GPS operating in 3D mode can indeed fail to get a position fix with 5 satellites being received, if they have bad geometry (e.g. all are in the same plane in space) because the solution will have horrible DOP values. But a timing-mode GPS in position hold mode knows its own (antenna) position, and only needs one visible satellite to continue to provide timing outputs. We don't know how the LTE-Lite's disciplining algorithm is tuned. If frequency stability was considered to be more important that timing, the algorithm may limit the maximum frequency offset that can be used to correct a timing error. Watching the scope output in real time, I can see the time offset between the two 1 PPS pulses change with time, but it always changes rather slowly, so the maximum frequency difference I've seen is quite small. (I no longer have the equipment set up, so I can't provide a quantitative number). - Dave === Dave, Thanks for that background. I'm sure Said must be on holiday (or unwell) otherwise he would have chipped in! Mine did finish the survey - eventually - and I saw this as the positions which were emitted being identical, and also that the survey light was extinguished. But at the moment the Lock OK light is out, and the survey light is out. Four satellites are showing at strength 27 or above, but no position is being emitted. PPS is present, about 170 ns late compared to the Rapco 1904M. The other GPS receivers are showing normal lock and a positional output. I suspect you are correct about the algorithms - the device being optimised for frequency rather than timing. Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
From: Dave Martindale In my case, the LTE-Lite had been operating for at least a week before I made my "accumulate mode" measurement, and the Thunderbolt had been operating for at least a month. But both antennas were in poor locations - not bad enough to lose lock any time I was watching, but nowhere close to a clear view of most of the sky. I never saw the 1 PPS disappear while I was watching it. I wonder if your LTE-Lite ever finished its survey and switched into 1D/position hold mode? A GPS operating in 3D mode can indeed fail to get a position fix with 5 satellites being received, if they have bad geometry (e.g. all are in the same plane in space) because the solution will have horrible DOP values. But a timing-mode GPS in position hold mode knows its own (antenna) position, and only needs one visible satellite to continue to provide timing outputs. We don't know how the LTE-Lite's disciplining algorithm is tuned. If frequency stability was considered to be more important that timing, the algorithm may limit the maximum frequency offset that can be used to correct a timing error. Watching the scope output in real time, I can see the time offset between the two 1 PPS pulses change with time, but it always changes rather slowly, so the maximum frequency difference I've seen is quite small. (I no longer have the equipment set up, so I can't provide a quantitative number). - Dave === Dave, Thanks for that background. I'm sure Said must be on holiday (or unwell) otherwise he would have chipped in! Mine did finish the survey - eventually - and I saw this as the positions which were emitted being identical, and also that the survey light was extinguished. But at the moment the Lock OK light is out, and the survey light is out. Four satellites are showing at strength 27 or above, but no position is being emitted. PPS is present, about 170 ns late compared to the Rapco 1904M. The other GPS receivers are showing normal lock and a positional output. I suspect you are correct about the algorithms - the device being optimised for frequency rather than timing. Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
In my case, the LTE-Lite had been operating for at least a week before I made my "accumulate mode" measurement, and the Thunderbolt had been operating for at least a month. But both antennas were in poor locations - not bad enough to lose lock any time I was watching, but nowhere close to a clear view of most of the sky. I never saw the 1 PPS disappear while I was watching it. I wonder if your LTE-Lite ever finished its survey and switched into 1D/position hold mode? A GPS operating in 3D mode can indeed fail to get a position fix with 5 satellites being received, if they have bad geometry (e.g. all are in the same plane in space) because the solution will have horrible DOP values. But a timing-mode GPS in position hold mode knows its own (antenna) position, and only needs one visible satellite to continue to provide timing outputs. We don't know how the LTE-Lite's disciplining algorithm is tuned. If frequency stability was considered to be more important that timing, the algorithm may limit the maximum frequency offset that can be used to correct a timing error. Watching the scope output in real time, I can see the time offset between the two 1 PPS pulses change with time, but it always changes rather slowly, so the maximum frequency difference I've seen is quite small. (I no longer have the equipment set up, so I can't provide a quantitative number). - Dave On Tue, Dec 9, 2014 at 5:13 AM, David J Taylor < david-tay...@blueyonder.co.uk> wrote: > I work with Said at Jackson Labs. I've been reading the time-nuts > discussion for a few years, but rarely chime in. I saw this discussion and > wanted to make a couple points. > > * The LTE Lite time accuracy specification corresponds with the Skytraq GPS > receiver's specs page which I have attached. The specification is for the > output directly from the GPS receiver available on the LTE Lite Eval > Board's JP1 connector pin 12. This specification assumes optimal antenna > placement and thermal conditions, and position hold mode. It is also an RMS > (1-sigma) measurement not a peak-to-peak measurement. > > * The GPSDO-generated 1PPS on the LTE Lite Eval Board's J1 connector has a > phase offset to the GPS raw 1PPS that is shown in the PJLTS message (2nd > field). The GPSDO functions to drive this phase offset to zero. But at a > given time--especially shortly after power up--the offset may 100 ns or > more. > > Keith > == > > Keith, > > Thanks very much for chiming in, as it has resolved what we are seeing, > particularly your second comment. > > One thing I do notice is that the device appears less sensitive than some > other GPS devices I have. Perhaps "sensitive" isn't the correct word, but > looking at the NMEA output it seems to indicate bursts of "no/invalid > position" a lot more often than I would expect. This is shown by the all > the signal strength bars being grey rather than some of them being blue. > I've also seen times when five or more satellites are above strength 29, > and yet there is no position shown. This also seems to stop the generation > of the PPS output, which would be not so good when driving an NTP server. > > I am wondering whether this is due to overly stringent criteria being set > for a "position found", at least for my location and antenna location, and > if this is the case, whether there is any chance of relaxing those > criteria. I'm guessing not, as the device will not accept any serial input. > > You will have gathered that my main interest is time rather than > frequency, and it seems that other GPS devices give PPS outputs which are > nearer to UTC but they have considerably more jitter. I'm only seeing this > on the 'scope - likely my PCs would bother with a microsecond either way. > > > Cheers, > David > -- > SatSignal Software - Quality software written to your requirements > Web: http://www.satsignal.eu > Email: david-tay...@blueyonder.co.uk > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/ > mailman/listinfo/time-nuts > and follow the instructions there. > ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
I work with Said at Jackson Labs. I've been reading the time-nuts discussion for a few years, but rarely chime in. I saw this discussion and wanted to make a couple points. * The LTE Lite time accuracy specification corresponds with the Skytraq GPS receiver's specs page which I have attached. The specification is for the output directly from the GPS receiver available on the LTE Lite Eval Board's JP1 connector pin 12. This specification assumes optimal antenna placement and thermal conditions, and position hold mode. It is also an RMS (1-sigma) measurement not a peak-to-peak measurement. * The GPSDO-generated 1PPS on the LTE Lite Eval Board's J1 connector has a phase offset to the GPS raw 1PPS that is shown in the PJLTS message (2nd field). The GPSDO functions to drive this phase offset to zero. But at a given time--especially shortly after power up--the offset may 100 ns or more. Keith == Keith, Thanks very much for chiming in, as it has resolved what we are seeing, particularly your second comment. One thing I do notice is that the device appears less sensitive than some other GPS devices I have. Perhaps "sensitive" isn't the correct word, but looking at the NMEA output it seems to indicate bursts of "no/invalid position" a lot more often than I would expect. This is shown by the all the signal strength bars being grey rather than some of them being blue. I've also seen times when five or more satellites are above strength 29, and yet there is no position shown. This also seems to stop the generation of the PPS output, which would be not so good when driving an NTP server. I am wondering whether this is due to overly stringent criteria being set for a "position found", at least for my location and antenna location, and if this is the case, whether there is any chance of relaxing those criteria. I'm guessing not, as the device will not accept any serial input. You will have gathered that my main interest is time rather than frequency, and it seems that other GPS devices give PPS outputs which are nearer to UTC but they have considerably more jitter. I'm only seeing this on the 'scope - likely my PCs would bother with a microsecond either way. Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
Hi If you read the NIST papers where they have looked at the PPS accuracy compared to UTC, the results are not all that good. The assumption that any one GSDO is “correct” compared to UTC is *not* a good one. The consistency of a GPSDO is quite good. That’s a very different thing than it’s accuracy (delta to UTC). In the case that absolute error relative to UTC is a requirement, you need a local UTC reference. The antenna delay setting is then used to “align” all of your GPSDO’s against your reference. On many GPSDO’s the antenna delay adjustment is a 100 ns resolution sort of thing. Again, it’s important to understand that these boxes were all made for cell service. That’s not an application where exact traceability to UTC is needed. Simply having all the sites run the same (consistent) GPSDO is perfectly adequate. If you have two brands of GPSDO, figure out the offset between them, still no need for “real” UTC. The “UTC” specs you see are one sigma bounds on the wander. Offset / centering of that peak are an unknown that is buried deep in the fine print. Bob > On Dec 8, 2014, at 10:12 AM, David J Taylor > wrote: > > From: Dave Martindale > > What is the source of the 1 PPS you are comparing against? > > I compared my LTE-Lite to an old Thunderbolt (original model, single 24 V > input with internal DC to DC converters, Piezo oscillator). At the time, > the Thunderbolt had been running for a few months, while the LTE-Lite had > been running for a week or so. Antennas were sitting on the window ledge > of a west-facing window, so relatively poor sky coverage. I connected the > PPS outputs from the two GPSDOs to two channels of a digital scope and left > it running in "accumulate" mode. A couple of the resulting displays are > attached below (I hope). Yellow trace is the Thunderbolt PPS, also the > trigger source. The LTE-Lite is the cyan trace. Each image shows signals > accumulated over a period of about 8-12 hours. > > As you can see, the relative timing of the two 1 Hz signals wanders by > about +- 100 ns around a midpoint value, but at this midpoint the LTE-Lite > is around 50 ns later that the Thunderbolt. (I call it a "midpoint" > because it's judged by eye as halfway between the two recorded extremes. I > don't have a record of the individual measurements, so I can't calculate > mean or median). The Thunderbolt's antenna cable is perhaps 10 feet > shorter than the LTE-Lite's, so that accounts for ~15 ns (Thunderbolt > antennas compensation is set to zero). > > So, at my house, the LTE-Lite is about 50 ns late (or the TB is 50 ns > early). That's one cycle of the LTE-Lite 20 MHz TCXO - coincidence? > [] > - Dave > === > > Dave, > > My comparison is against a Rapco 1904M, which is another GPSDO. That does > agree on a causal measurement with a number of simple GPS/PPS units I have. A > u-blox LEA-6T shows about 80 ns later than the 1904M, and a u-blox NEO-6M > between 50 ns early and 200 ns late, both after being on for just a few > minutes, and with no special care in antenna placing. > > Do you think that your measurement (~35 ns offset) is consistent with the > LTE-Lite specification: > > "1 PPS Timing Accuracy from GPS receiver" > "<8ns to UTC RMS (1-Sigma) GPS Locked" > > and the specification of the Thunderbolt? > > Cheers, > David > -- > SatSignal Software - Quality software written to your requirements > Web: http://www.satsignal.eu > Email: david-tay...@blueyonder.co.uk > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
I work with Said at Jackson Labs. I've been reading the time-nuts discussion for a few years, but rarely chime in. I saw this discussion and wanted to make a couple points. * The LTE Lite time accuracy specification corresponds with the Skytraq GPS receiver's specs page which I have attached. The specification is for the output directly from the GPS receiver available on the LTE Lite Eval Board's JP1 connector pin 12. This specification assumes optimal antenna placement and thermal conditions, and position hold mode. It is also an RMS (1-sigma) measurement not a peak-to-peak measurement. * The GPSDO-generated 1PPS on the LTE Lite Eval Board's J1 connector has a phase offset to the GPS raw 1PPS that is shown in the PJLTS message (2nd field). The GPSDO functions to drive this phase offset to zero. But at a given time--especially shortly after power up--the offset may 100 ns or more. Keith Keith On Mon, Dec 8, 2014 at 7:12 AM, David J Taylor < david-tay...@blueyonder.co.uk> wrote: > From: Dave Martindale > > What is the source of the 1 PPS you are comparing against? > > I compared my LTE-Lite to an old Thunderbolt (original model, single 24 V > input with internal DC to DC converters, Piezo oscillator). At the time, > the Thunderbolt had been running for a few months, while the LTE-Lite had > been running for a week or so. Antennas were sitting on the window ledge > of a west-facing window, so relatively poor sky coverage. I connected the > PPS outputs from the two GPSDOs to two channels of a digital scope and left > it running in "accumulate" mode. A couple of the resulting displays are > attached below (I hope). Yellow trace is the Thunderbolt PPS, also the > trigger source. The LTE-Lite is the cyan trace. Each image shows signals > accumulated over a period of about 8-12 hours. > > As you can see, the relative timing of the two 1 Hz signals wanders by > about +- 100 ns around a midpoint value, but at this midpoint the LTE-Lite > is around 50 ns later that the Thunderbolt. (I call it a "midpoint" > because it's judged by eye as halfway between the two recorded extremes. I > don't have a record of the individual measurements, so I can't calculate > mean or median). The Thunderbolt's antenna cable is perhaps 10 feet > shorter than the LTE-Lite's, so that accounts for ~15 ns (Thunderbolt > antennas compensation is set to zero). > > So, at my house, the LTE-Lite is about 50 ns late (or the TB is 50 ns > early). That's one cycle of the LTE-Lite 20 MHz TCXO - coincidence? > [] > - Dave > === > > Dave, > > My comparison is against a Rapco 1904M, which is another GPSDO. That does > agree on a causal measurement with a number of simple GPS/PPS units I have. > A u-blox LEA-6T shows about 80 ns later than the 1904M, and a u-blox NEO-6M > between 50 ns early and 200 ns late, both after being on for just a few > minutes, and with no special care in antenna placing. > > Do you think that your measurement (~35 ns offset) is consistent with the > LTE-Lite specification: > > "1 PPS Timing Accuracy from GPS receiver" > "<8ns to UTC RMS (1-Sigma) GPS Locked" > > and the specification of the Thunderbolt? > > Cheers, > > David > -- > SatSignal Software - Quality software written to your requirements > Web: http://www.satsignal.eu > Email: david-tay...@blueyonder.co.uk > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/ > mailman/listinfo/time-nuts > and follow the instructions there. > Venus838LPx-T-Specs.pdf Description: Adobe PDF document ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
From: Dave Martindale What is the source of the 1 PPS you are comparing against? I compared my LTE-Lite to an old Thunderbolt (original model, single 24 V input with internal DC to DC converters, Piezo oscillator). At the time, the Thunderbolt had been running for a few months, while the LTE-Lite had been running for a week or so. Antennas were sitting on the window ledge of a west-facing window, so relatively poor sky coverage. I connected the PPS outputs from the two GPSDOs to two channels of a digital scope and left it running in "accumulate" mode. A couple of the resulting displays are attached below (I hope). Yellow trace is the Thunderbolt PPS, also the trigger source. The LTE-Lite is the cyan trace. Each image shows signals accumulated over a period of about 8-12 hours. As you can see, the relative timing of the two 1 Hz signals wanders by about +- 100 ns around a midpoint value, but at this midpoint the LTE-Lite is around 50 ns later that the Thunderbolt. (I call it a "midpoint" because it's judged by eye as halfway between the two recorded extremes. I don't have a record of the individual measurements, so I can't calculate mean or median). The Thunderbolt's antenna cable is perhaps 10 feet shorter than the LTE-Lite's, so that accounts for ~15 ns (Thunderbolt antennas compensation is set to zero). So, at my house, the LTE-Lite is about 50 ns late (or the TB is 50 ns early). That's one cycle of the LTE-Lite 20 MHz TCXO - coincidence? [] - Dave === Dave, My comparison is against a Rapco 1904M, which is another GPSDO. That does agree on a causal measurement with a number of simple GPS/PPS units I have. A u-blox LEA-6T shows about 80 ns later than the 1904M, and a u-blox NEO-6M between 50 ns early and 200 ns late, both after being on for just a few minutes, and with no special care in antenna placing. Do you think that your measurement (~35 ns offset) is consistent with the LTE-Lite specification: "1 PPS Timing Accuracy from GPS receiver" "<8ns to UTC RMS (1-Sigma) GPS Locked" and the specification of the Thunderbolt? Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
From: Bill Dailey Or use 1ns per foot of antenna cable. That will get you closer. == I was using 5ns per metre to allow for velocity factor. Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
Or use 1ns per foot of antenna cable. That will get you closer. Sent from mobile > On Dec 8, 2014, at 8:16 AM, David J Taylor > wrote: > > Hi > > On *most* GPSDO’s, the simple answer is “there is a cable delay adjustment to > align it”. Without some sort of reliable representation of UTC (at the ns > level) it’s tough to measure. If you happen to live at USNO or NIST, you can > access that sort of timing. For the rest of us - not so easy. The NIST two > way GPS “modem” setup is about the only practical method that I know of. > > The PPS out of any GPSDO will be much lower jitter than the pps from a normal > GPS module. > > Bob > == > > Bob, > > Thanks for your comments. The antenna location and cable lengths are very > similar (either 0, 5m or 10m) so I was expecting somewhat less than 50 ns > difference. 200+ ns is rather more than I expected for the LTE-Lite, so I > did wonder whether anyone else had measured it. > > Although I've not checked it rigorously, most of the GPS/PPS units here of > various brands are within under 100 ns of each other, hence my expectation. > > Cheers, > David > -- > SatSignal Software - Quality software written to your requirements > Web: http://www.satsignal.eu > Email: david-tay...@blueyonder.co.uk > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
Hi On *most* GPSDO’s, the simple answer is “there is a cable delay adjustment to align it”. Without some sort of reliable representation of UTC (at the ns level) it’s tough to measure. If you happen to live at USNO or NIST, you can access that sort of timing. For the rest of us - not so easy. The NIST two way GPS “modem” setup is about the only practical method that I know of. The PPS out of any GPSDO will be much lower jitter than the pps from a normal GPS module. Bob == Bob, Thanks for your comments. The antenna location and cable lengths are very similar (either 0, 5m or 10m) so I was expecting somewhat less than 50 ns difference. 200+ ns is rather more than I expected for the LTE-Lite, so I did wonder whether anyone else had measured it. Although I've not checked it rigorously, most of the GPS/PPS units here of various brands are within under 100 ns of each other, hence my expectation. Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
What is the source of the 1 PPS you are comparing against? I compared my LTE-Lite to an old Thunderbolt (original model, single 24 V input with internal DC to DC converters, Piezo oscillator). At the time, the Thunderbolt had been running for a few months, while the LTE-Lite had been running for a week or so. Antennas were sitting on the window ledge of a west-facing window, so relatively poor sky coverage. I connected the PPS outputs from the two GPSDOs to two channels of a digital scope and left it running in "accumulate" mode. A couple of the resulting displays are attached below (I hope). Yellow trace is the Thunderbolt PPS, also the trigger source. The LTE-Lite is the cyan trace. Each image shows signals accumulated over a period of about 8-12 hours. As you can see, the relative timing of the two 1 Hz signals wanders by about +- 100 ns around a midpoint value, but at this midpoint the LTE-Lite is around 50 ns later that the Thunderbolt. (I call it a "midpoint" because it's judged by eye as halfway between the two recorded extremes. I don't have a record of the individual measurements, so I can't calculate mean or median). The Thunderbolt's antenna cable is perhaps 10 feet shorter than the LTE-Lite's, so that accounts for ~15 ns (Thunderbolt antennas compensation is set to zero). So, at my house, the LTE-Lite is about 50 ns late (or the TB is 50 ns early). That's one cycle of the LTE-Lite 20 MHz TCXO - coincidence? I also have an old Garmin GPS-25 board. This is a navigation GPS, without timing features, but it does have a 1 PPS output. I've included one capture of GPS-25 vs. Thunderbolt. The jitter is much worse; most (but not all) traces are within +- 400 ns of the Thunderbolt (note the different horizontal sweep). And there is also an overall bias: the Garmin receiver appears to be about 100 ns late on "average" compared to the TB. Unfortunately, I don't have any other way to measure which GPSDO has the more accurate PPS, and which one is responsible for most of the jitter. (A man with two GPSDOs never knows what time it is, precisely). I do have a big old 5 MHz OCXO pulled from a Transit receiver which is probably quite stable, but it is 0.2 Hz off nominal frequency and is not adjustable. Viewed on a scope alongside either GPSDO output, the 5 MHz phase shifts by one cycle every 5 seconds, too fast to make any comparison by eye of the stability of either GPSDO. - Dave On Mon, Dec 8, 2014 at 7:17 AM, David J Taylor < david-tay...@blueyonder.co.uk> wrote: > On the 10MHz LTE-Lite, how far out from true UTC would the PPS be expected > to be? > > It seems to be about 200+ ns late on my unit, although it is much more > stable than a typical GPS/PPS produces. > > Thanks, > David > -- > SatSignal Software - Quality software written to your requirements > Web: http://www.satsignal.eu > Email: david-tay...@blueyonder.co.uk > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/ > mailman/listinfo/time-nuts > and follow the instructions there. > ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] 10MHz LTE-Lite - PPS accuracy?
Hi On *most* GPSDO’s, the simple answer is “there is a cable delay adjustment to align it”. Without some sort of reliable representation of UTC (at the ns level) it’s tough to measure. If you happen to live at USNO or NIST, you can access that sort of timing. For the rest of us - not so easy. The NIST two way GPS “modem” setup is about the only practical method that I know of. The PPS out of any GPSDO will be much lower jitter than the pps from a normal GPS module. Bob > On Dec 8, 2014, at 7:17 AM, David J Taylor > wrote: > > On the 10MHz LTE-Lite, how far out from true UTC would the PPS be expected to > be? > > It seems to be about 200+ ns late on my unit, although it is much more stable > than a typical GPS/PPS produces. > > Thanks, > David > -- > SatSignal Software - Quality software written to your requirements > Web: http://www.satsignal.eu > Email: david-tay...@blueyonder.co.uk > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.