Re: [time-nuts] tyco electronics A1025
On 2014-12-14 10:29, Francesco Messineo wrote: On Sun, Dec 14, 2014 at 6:12 PM, Bob Camp kb...@n1k.org wrote: that's not meant as a time nut stratum 1. It's just a free gps module I would like to recycle as a needed stratum 1 server for a small network. Of course if I can find informations on it. I know there're better options, but in this case anything would do. … and NTP is not a GPSDO or a Cs replacement. My guess is that there is no PPS out of the device. It would be very unusual if there was. Finding the NEMA output pin should be possible with an oscilloscope. At that point, a simple serial connection to the server is about all you need. Bring up the NEMA driver and it is running. It is unlikely that any further optimization would be possible, even with the (maybe) 290 page data sheet on the part. I would not let the lack of a data sheet stop you in this case. Hook up the output to a PC with a terminal program and see what you get. The main problem would be if you need to find the serial input pin to change what it puts out, hopefully you do not. The A1029, which is a newer model, has indeed a PPS output and I've been able to find a datasheet for it but the pinout isn't anything like the A1025. I planned to reverse engineer the pinout, but I'd like at least not to be forced to try to guess the power pins. Maybe someone still has the data for this older module. One article mentions the A1029 as a drop in replacement for the A1025, as an early auto receiver with gyro and dead reckoning nav holdover, but that may refer to the complete module, and you may have just the GPS. The GPS could have provided PPS for DR nav, and some TE model specs offer TCXOs, which may also have been required for DR timing holdover, but may not have been part of the GPS. Those GPS seem to have been standard STMicroelectronics parts with firmware customization for functions and additions, and offered proprietary $PSTM NMEA sentences. If you can read off the STM part STA 2... (perhaps under a patch antenna) you may be able to search for more details. -- Take care. Thanks, Brian Inglis ___ 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] GPS position averaging
Hello, all I have a couple of GPSDOs (KS, z3805a) hooked up to a fixed outdoor antenna through an HP 58516 splitter. I notice that when I powercycle the GPSDOs (which I generally try to avoid), the position they come up with differs from time to time in at least the last three digits (and height in particular varies with up to 30m). I would like to get a better idea of what my actualt antenna location is, so that I can manually set the position in the GPSDOs. For that purpose I set up a Ublox M8N with logging in U-center, and I've collected about 250K readings over several days. For each reading I also log number of SV's used, as well as HDOP, VDOP. PDOP, thinking I would filter the list and end up with a subset of really good fixes.The idea is then to average the best readings and use that as my position. My question is twofold; 1) is this for some reason a bad idea? And 2) How do I average the numbers? I can not put my finger on it but it feels wrong to average lat, long and height independently. Any hints are greatly appreciated. Thank you! Ole P ___ 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] GPS position averaging
olep...@gmail.com said: My question is twofold; 1) is this for some reason a bad idea? And 2) How do I average the numbers? I can not put my finger on it but it feels wrong to average lat, long and height independently. It seems like an interesting idea to me. It would be interesting to run a test: setup a pair of identical GPSDOs running off the same antenna but using different locations and see if you can see any difference in the PPS or 10 MHz output. You can do a filtering pass to dump anything with fewer than 4 satellites. The survey code has a mode where it says bad geometry. I don't know the fine print on how that works. I can get my elevation off a topo map. I'm not sure how to translate that to GPS coordinates. Something like that might be another opportunity for filtering. The refclock part of ntpd filters a batch of samples by discarding outliers. Time is only one dimension, so the recipe is pretty simple: sort, compute the average, discard either the top or bottom, whichever is farther from the average. I'm not sure how to do something similar in 2 or 3 dimensions. It might be interesting to make histograms of HDOP and friends. (before and after any filtering you can come up with?) That may cover the bad-geometry case. -- 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] GPS position averaging
I am new to this list and to this topic, but it seems to me that if one wants to come up with an average of a set of spatial measurements, one would use distance as the parameter to be averaged. The distance would presumably be that from a fixed spatial reference point (0,0,0). One would then take the square root of the sum of the squares of latitude, longitude and height. My problem with this is that I can't see what one would use for the reference point, but maybe that is not important (or maybe it is!). DaveD On 12/15/2014 2:48 AM, Hal Murray wrote: olep...@gmail.com said: My question is twofold; 1) is this for some reason a bad idea? And 2) How do I average the numbers? I can not put my finger on it but it feels wrong to average lat, long and height independently. It seems like an interesting idea to me. It would be interesting to run a test: setup a pair of identical GPSDOs running off the same antenna but using different locations and see if you can see any difference in the PPS or 10 MHz output. You can do a filtering pass to dump anything with fewer than 4 satellites. The survey code has a mode where it says bad geometry. I don't know the fine print on how that works. I can get my elevation off a topo map. I'm not sure how to translate that to GPS coordinates. Something like that might be another opportunity for filtering. The refclock part of ntpd filters a batch of samples by discarding outliers. Time is only one dimension, so the recipe is pretty simple: sort, compute the average, discard either the top or bottom, whichever is farther from the average. I'm not sure how to do something similar in 2 or 3 dimensions. It might be interesting to make histograms of HDOP and friends. (before and after any filtering you can come up with?) That may cover the bad-geometry case. ___ 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] GPS position averaging
I would recommend determining your terrain's elevation using a topo map and estimating or measuring antennas height above ground. And then excluding survey results with wacko altitudes before averaging. While we often set the elevation angle mask high for timing purposes, for survey purposes especially altitude it will likely be best to do survey with no or very low (10 degree default?) elevation mask. The Z3801A manual tells how to set elevation mask for at least some of your units. Tim N3QE On Mon, Dec 15, 2014 at 4:03 AM, Ole Petter Ronningen olep...@gmail.com wrote: Hello, all I have a couple of GPSDOs (KS, z3805a) hooked up to a fixed outdoor antenna through an HP 58516 splitter. I notice that when I powercycle the GPSDOs (which I generally try to avoid), the position they come up with differs from time to time in at least the last three digits (and height in particular varies with up to 30m). I would like to get a better idea of what my actualt antenna location is, so that I can manually set the position in the GPSDOs. For that purpose I set up a Ublox M8N with logging in U-center, and I've collected about 250K readings over several days. For each reading I also log number of SV's used, as well as HDOP, VDOP. PDOP, thinking I would filter the list and end up with a subset of really good fixes.The idea is then to average the best readings and use that as my position. My question is twofold; 1) is this for some reason a bad idea? And 2) How do I average the numbers? I can not put my finger on it but it feels wrong to average lat, long and height independently. Any hints are greatly appreciated. Thank you! Ole P ___ 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] The Trapezoidal Clocking
Hi If you treat it as a sine, you certainly get further. There are some neat sine wave input chips. The next layer to the problem is that the trapezoid may be slower than (or faster than ) a sine wave depending on just what’s going on with the system. That will make it a bit difficult to know what sort of a clock you have at each tap. Bob On Dec 14, 2014, at 10:40 PM, Magnus Danielson mag...@rubidium.dyndns.org wrote: Hi, Agree. It's not ideal. In it's simplicity it achieves first degree delay compensation, but it is not the best of signals you get. You need to treat the signal as being essentially a sine, and overcome the slew-rate. It may be a useful technique besides it's limits. Cheers, Magnus On 12/15/2014 03:45 AM, Bob Camp wrote: Hi …. but … Low edge speeds = poor signal to noise = high jitter. The result is a clock that aligns, but has high(er) jitter compared to a conventional square wave clock. Most of the “lower phase noise / lower jitter” progress in CMOS logic has gone hand in hand with faster edge rates. Bob On Dec 14, 2014, at 8:27 PM, Magnus Danielson mag...@rubidium.dyndns.org wrote: Bob, On 12/14/2014 02:10 AM, Robert Darby wrote: This is an paper that may be of some interest to those interested in clock distribution. The author, Jinyuan Wu has done considerable work on FPGA TIC's with Fermi Lab. http://www-ppd.fnal.gov/EEDOffice-w/Projects/ckm/comadc/TrapezCLK1b.pdf How incredibly cunning! Cheers, Magnus ___ 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. ___ 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] Did a member of time-nuts buy this?
On Dec 15, 2014, at 2:27 AM, Mike Monett timen...@binsamp.e4ward.com wrote: Hi On Dec 10, 2014, at 11:12 PM, Mike Monett timen...@binsamp.e4ward.com wrote: [...] Can you tell me some of the ones that do? I have yet to see one for under $2K that does it correctly. I don't have the cash to buy ones at those sort of prices. Some have reported that the old Motorola UT's will do it. The samples I have tried have not done very well. I may have not had them running right who knows. OK, that pretty much eliminates the NIST data. or means that more research is needed on single sat boards. There are a *lot* of them out there. What is a single sat board and can you give some examples? A GPS module that will correctly and accurately report the time solutions for each and every satellite that it’s using as part of the total timing solution. An alternative might be to take RINEX data out of a module and post process that yourself. The intent is to be able to say “this sat told me I’m 15.32 ns fast”. You then look at somebody else’s data and they say “same sat told me I’m 17.82 ns slow”. 5) Feed that into your control loop equation. There's another term I need to research! It's not a simple control process, but it's not all that terrible either. It just takes a bit of work to optimize. Figure a few months to a few years for the optimization depending on what you have for issues along the way. OK, so I figure out how to do this. How do I tell if this is making the gpsdo more accurate? In other words, how do I get the ADEV without having an H-Maser? You get a Cs (or other atomic standard) or you compare several different GPSDO's against each other. The preference would be for groups of three so you can rule out ones that are not doing what they should. The ideal would be 3 TBolts, 3 same model Rb's, 3 same model OCXO's and more than three groups overall. Each has their own ADEV curve. Comparing devices with vastly different ADEV is not the best way to go. Yes, that makes good sense. [...] There are many new dacs, op amps, and bipolar microwave devices that offer much lower noise than current designs use. I think the performance can be improved in some areas with new components and design techniques, and I have equipment and time to explore. There are only two points that system noise really comes into the GPSDO design: 1) The TDC must have enough resolution 2) The DAC on the EFC must have noise below the OCXO I was more thinking of the Rubidium physics package. Most of the Rubidiums on eBay are pretty old. The lamp must have constant intensity. The null detector has to identify a very shallow drop. The microwave signal needs to have excellent phase noise. These all need low noise, low flicker components, and significant advances have been made in recent years. The surrounding circuitry could probably benefit from a redesign to take better advantage of low noise components. For example, SRD's are much noisier than NLTL's. Probably little can be done to improve the Rubidium cell, but it should be the limiting factor and not the electronics. The small Rb’s are designed for a target cost. There is very little interest in fancy Rb’s. That low cost target does indeed impact the physics package. The TDC is limited by the basic resolution of the GPS system. It's easy to build one that has far more resolution than needed / useful. Some people have little faith:) Which is indeed a good thing at first. The GPS output is (at best) a 100 ps sort of thing. Each time I say that others will pop up and suggest that it’s closer to 1 ns. Either way, the demands on the TDC are not all that major. The long term (as opposed to 1 second) error on the GPS signal are likely closer to a couple of ns than to the number 0.1 or 1 ns level. The DAC issue is normally solved with a $4 part. Unless you have an OCXO with a very wide EFC range, it's noise is unlikely to be an issue. DAC resolution can be addressed to any desired level with two DAC's (fine and coarse). (Once you get going, you only use the fine DAC). The real fun and games revolves around the software used to implement the filtering / control loop between the GPS and the OCXO (or Rb, or TCXO, or MEMS, or VCXO, or Cs or âŠ) Focus on the software. Bob I have a few Rubidiums and OCXOs I'd like to get running for a month or so to stabilize. During this time, I'd like to monitor the performance to discover any bad units and see which are the best ones. For example, TVB shows a 100:1 variation in ADEV in FE-405B Rubidiums. The section is titled Variation in FE-405B in http://leapsecond.com/pages/fe405/ I’m not a 405 expert, but I believe the 405 is OCXO based. Clearly, it would be futile to try to use a bad unit in setting up a gpsdo. I need some means of measuring the performance of these units while they are running. A
Re: [time-nuts] tyco electronics A1025
On Mon, Dec 15, 2014 at 9:09 AM, Brian Inglis brian.ing...@systematicsw.ab.ca wrote: On 2014-12-14 10:29, Francesco Messineo wrote: The A1029, which is a newer model, has indeed a PPS output and I've been able to find a datasheet for it but the pinout isn't anything like the A1025. I planned to reverse engineer the pinout, but I'd like at least not to be forced to try to guess the power pins. Maybe someone still has the data for this older module. One article mentions the A1029 as a drop in replacement for the A1025, as an early auto receiver with gyro and dead reckoning nav holdover, but that may refer to the complete module, and you may have just the GPS. The GPS could have provided PPS for DR nav, and some TE model specs offer TCXOs, which may also have been required for DR timing holdover, but may not have been part of the GPS. Those GPS seem to have been standard STMicroelectronics parts with firmware customization for functions and additions, and offered proprietary $PSTM NMEA sentences. If you can read off the STM part STA 2... (perhaps under a patch antenna) you may be able to search for more details. I've found a couple of articles saying the A1025 indeed has PPS output as I suspected. However, none of them reports any hint about the pinout of this module. The module itself is soldered on a small daughter board, so I can't look on the other side for possible part numbers other than the tyco electronics one. The daughter board has some transistors, passives, a 74HC14 and a small component, probably a power supply regulator that I can't identify. It's a microchip part marked CS05351CK (CS0535 1CK in two rows). Reverse engineering is not progressing much also because of this unknown microchip part. Regards Frank ___ 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] GPS position averaging
The Ublox M8N can send out raw measurements with messages TRK-MEAS and TRK-SFRBX for observation and navigation data: you can postprocess them to improve the accuracy. http://www.rtklib.com/rtklib_support.htm On Mon, Dec 15, 2014 at 12:43 PM, Tim Shoppa tsho...@gmail.com wrote: I would recommend determining your terrain's elevation using a topo map and estimating or measuring antennas height above ground. And then excluding survey results with wacko altitudes before averaging. While we often set the elevation angle mask high for timing purposes, for survey purposes especially altitude it will likely be best to do survey with no or very low (10 degree default?) elevation mask. The Z3801A manual tells how to set elevation mask for at least some of your units. Tim N3QE On Mon, Dec 15, 2014 at 4:03 AM, Ole Petter Ronningen olep...@gmail.com wrote: Hello, all I have a couple of GPSDOs (KS, z3805a) hooked up to a fixed outdoor antenna through an HP 58516 splitter. I notice that when I powercycle the GPSDOs (which I generally try to avoid), the position they come up with differs from time to time in at least the last three digits (and height in particular varies with up to 30m). I would like to get a better idea of what my actualt antenna location is, so that I can manually set the position in the GPSDOs. For that purpose I set up a Ublox M8N with logging in U-center, and I've collected about 250K readings over several days. For each reading I also log number of SV's used, as well as HDOP, VDOP. PDOP, thinking I would filter the list and end up with a subset of really good fixes.The idea is then to average the best readings and use that as my position. My question is twofold; 1) is this for some reason a bad idea? And 2) How do I average the numbers? I can not put my finger on it but it feels wrong to average lat, long and height independently. Any hints are greatly appreciated. Thank you! Ole P ___ 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] GPS position averaging
On Mon, Dec 15, 2014 at 10:52 AM, Azelio Boriani azelio.bori...@gmail.com wrote: The Ublox M8N can send out raw measurements with messages TRK-MEAS and TRK-SFRBX for observation and navigation data: you can postprocess them to improve the accuracy. And searching the archives for rtklib should find some relevant posts with pointers to post-processing. Naturally the output of rtklib doesn't quite agree with the Canadian (CSRS-PPP) results although that could just be operator error. ___ 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] Manual repository
Users of my web site may have had difficulties getting some of the files in the last couple of months. Particularly, those files that were recently uploaded. Typical symptom would be a file that was uploaded, was removed from the Recent Upload folder (indicating that I moved the file to it's normal folder) but the file would not show up in search, while being in the normal folder. This problem has now been fixed. Let me know if you come across any new issue. Thank you Didier KO4BB www.ko4bb.com/manuals ___ 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] GPS position averaging
Take a look at the code in Lady Heather that does a 48 hour precision survey. It calculates a weighted median position of fixes over 1 hour periods then calculates a final position from those medians. The algorithm was developed by having people around the world with Thunderbolts and quality antennas and surveyed antenna positions log data for 48 hours. I processed the data and did some voodoo to come up with the weights, etc that minimized the error in the calculated positions from the surveyed positions. The 48 hour survey interval minimizes the effects of multi-path and satellite orbits, etc. ___ 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] tyco electronics A1025
I reply to myself, On Mon, Dec 15, 2014 at 2:41 PM, Francesco Messineo francesco.messi...@gmail.com wrote: I've found a couple of articles saying the A1025 indeed has PPS output as I suspected. However, none of them reports any hint about the pinout of this module. The module itself is soldered on a small daughter board, so I can't look on the other side for possible part numbers other than the tyco electronics one. The daughter board has some transistors, passives, a 74HC14 and a small component, probably a power supply regulator that I can't identify. It's a microchip part marked CS05351CK (CS0535 1CK in two rows). I identified this component, it's an MCP1700 LDO 3.3V voltage regulator in SOT-89 package. So now I can power the board, connect an antenna and scope signals. It should be easy now. Best regards Frank ___ 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] GPS position averaging
Hi Simple answer - there is no simple answer. You will get significantly better results by intelligently processing the data than by a giant “average everything in sight” approach. Also consider that various GPS modules seem to have offsets in their nav solution. The NIST papers on various GPSDO’s show this pretty clearly. Simply having a “correct” solution may not help as much as you might think. Of course if you want a truly correct solution, your local surveyor is indeed your best friend. He probably has a GPS gizmo that with about a half hour or sitting there, tell you where you are to under an inch. Who knows how much that will cost. It probably depends on how interested you can get him in coming over and looking at your toys. Bob ___ 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] GPS position averaging
On Mon, 15 Dec 2014 17:55:20 -0500, you wrote: Hi Simple answer - there is no simple answer. You will get significantly better results by intelligently processing the data than by a giant average everything in sight approach. Also consider that various GPS modules seem to have offsets in their nav solution. The NIST papers on various GPSDOs show this pretty clearly. Simply having a correct solution may not help as much as you might think. Of course if you want a truly correct solution, your local surveyor is indeed your best friend. He probably has a GPS gizmo that with about a half hour or sitting there, tell you where you are to under an inch. Who knows how much that will cost. It probably depends on how interested you can get him in coming over and looking at your toys. But is it the closest to the 'true' position that you really want, or the best estimate of where the particular GPS you are testing thinks it is? Angus ___ 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] GPS position averaging
Hi On Dec 15, 2014, at 6:39 PM, Angus not.ag...@btinternet.com wrote: On Mon, 15 Dec 2014 17:55:20 -0500, you wrote: Hi Simple answer - there is no simple answer. You will get significantly better results by intelligently processing the data than by a giant average everything in sight approach. Also consider that various GPS modules seem to have offsets in their nav solution. The NIST papers on various GPSDOs show this pretty clearly. Simply having a correct solution may not help as much as you might think. Of course if you want a truly correct solution, your local surveyor is indeed your best friend. He probably has a GPS gizmo that with about a half hour or sitting there, tell you where you are to under an inch. Who knows how much that will cost. It probably depends on how interested you can get him in coming over and looking at your toys. But is it the closest to the 'true' position that you really want, or the best estimate of where the particular GPS you are testing thinks it is? I would *guess* that the position reported by the GPS would be the better alternative. The distances (errors) involved are often fairly small. Proving which one is correct may not be easy. Obviously, the ideal case is the one where there is no error. If the error exists it says that the firmware has some sort of bug. Of course *my* code never ever has bugs … :) Bob Angus ___ 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] GPS position averaging
This may be one of those cases where the human brain's ability to see patterns would help to see the distribution and eliminate outliers. Plot a manageable set of X and Y points, and another set of X and Z points. To make the three dimensions plot as one in a statistical distribution (bell) curve, use the magnitude of the vector. What do the surveyors instruments do that can get them down to a few centimeters in half an hour? You might also want to study what's been said about the causes of variance. The atmosphere is not uniform. Consider the relative effect of position on time, where one foot or 30 cm is equivalent to 10E-9 second. Also consider the 'averaging' period of the PLL filter in the disciplined oscillator. I'm no expert, especially not in math. These hints are thoughts that come to mind. Bill Hawkins -Original Message- From: Ole Petter Ronningen Sent: Monday, December 15, 2014 3:04 AM I would like to get a better idea of what my actualt antenna location is, so that I can manually set the position in the GPSDOs. For that purpose I set up a Ublox M8N with logging in U-center, and I've collected about 250K readings over several days. For each reading I also log number of SV's used, as well as HDOP, VDOP. PDOP, thinking I would filter the list and end up with a subset of really good fixes.The idea is then to average the best readings and use that as my position. My question is twofold; 1) is this for some reason a bad idea? And 2) How do I average the numbers? I can not put my finger on it but it feels wrong to average lat, long and height independently. Any hints are greatly appreciated. ___ 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] GPS position averaging
Hi On Dec 15, 2014, at 7:19 PM, Bill Hawkins b...@iaxs.net wrote: This may be one of those cases where the human brain's ability to see patterns would help to see the distribution and eliminate outliers. Plot a manageable set of X and Y points, and another set of X and Z points. yup To make the three dimensions plot as one in a statistical distribution (bell) curve, use the magnitude of the vector. What do the surveyors instruments do that can get them down to a few centimeters in half an hour? The ones around here use a survey grade GPS and log the data to a file. They do a bunch of post processing to get the solution to a desired level of accuracy. Last time I had it done, they had more fun looking at the GPSDO’s than running their fancy GPS. You might also want to study what's been said about the causes of variance. The atmosphere is not uniform. Ionospheric issues are one of the things that gets corrected for to make a more accurate result. Consider the relative effect of position on time, where one foot or 30 cm is equivalent to 10E-9 second. …. and some of the errors between solutions are 10 feet. Also consider the 'averaging' period of the PLL filter in the disciplined oscillator. The problem shows up when the GPS constellation is in a configuration that accentuates the error. That’s often a 12 / 24 / 48 hour sort of thing. Bob I'm no expert, especially not in math. These hints are thoughts that come to mind. Bill Hawkins -Original Message- From: Ole Petter Ronningen Sent: Monday, December 15, 2014 3:04 AM I would like to get a better idea of what my actualt antenna location is, so that I can manually set the position in the GPSDOs. For that purpose I set up a Ublox M8N with logging in U-center, and I've collected about 250K readings over several days. For each reading I also log number of SV's used, as well as HDOP, VDOP. PDOP, thinking I would filter the list and end up with a subset of really good fixes.The idea is then to average the best readings and use that as my position. My question is twofold; 1) is this for some reason a bad idea? And 2) How do I average the numbers? I can not put my finger on it but it feels wrong to average lat, long and height independently. Any hints are greatly appreciated. ___ 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] GPS position averaging
Hi Actually there may be a Time Nuts relevant reason for wanting to know the true / correct / survey location. If your desire is to know UTC to ns or something like that, an indicated position error *probably* relates directly to a time error. Being able to correct the position error may allow you to better estimate the UTC time error. Bob On Dec 15, 2014, at 6:39 PM, Angus not.ag...@btinternet.com wrote: On Mon, 15 Dec 2014 17:55:20 -0500, you wrote: Hi Simple answer - there is no simple answer. You will get significantly better results by intelligently processing the data than by a giant average everything in sight approach. Also consider that various GPS modules seem to have offsets in their nav solution. The NIST papers on various GPSDOs show this pretty clearly. Simply having a correct solution may not help as much as you might think. Of course if you want a truly correct solution, your local surveyor is indeed your best friend. He probably has a GPS gizmo that with about a half hour or sitting there, tell you where you are to under an inch. Who knows how much that will cost. It probably depends on how interested you can get him in coming over and looking at your toys. But is it the closest to the 'true' position that you really want, or the best estimate of where the particular GPS you are testing thinks it is? Angus ___ 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] GPS position averaging
kb...@n1k.org said: The problem shows up when the GPS constellation is in a configuration that accentuates the error. Thatâs often a 12 / 24 / 48 hour sort of thing. It would be interesting to compare the results from daytime vs nighttime, or today vs yesterday. -- 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] GPS position averaging
Hi On Dec 15, 2014, at 8:07 PM, Hal Murray hmur...@megapathdsl.net wrote: kb...@n1k.org said: The problem shows up when the GPS constellation is in a configuration that accentuates the error. That’s often a 12 / 24 / 48 hour sort of thing. It would be interesting to compare the results from daytime vs nighttime, or today vs yesterday. There can be seasonal issues (trees sprout leaves …) as well. Bob -- 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. ___ 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] GPS position averaging
On Mon, Dec 15, 2014 at 6:39 PM, Angus not.ag...@btinternet.com wrote: But is it the closest to the 'true' position that you really want, or the best estimate of where the particular GPS you are testing thinks it is? I don't understand what you're saying here. Can you phrase it differently? ___ 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] Did a member of time-nuts buy this?
Hi Here’s another way to sort this all out: For ADEV, you probably want to get down to ~ 1x10^-13 at 1 second. If the system improves by 1/tau, it will be adequate for anything else you want to do. It’s a system that would need to resolve 100 fs, so most of the “counter” devices that die out around 10 ps would be ruled out. For long term, you might decide on 1x10^-15 at 100,000 seconds. If it also goes as 1/tau you would have 1x10^-13 at 1,000 seconds. Not good enough for ADEV, but quite adequate for things like aging. Any of the 10 to 100 ps range devices would do fine in this case. Sort into two piles and then sub-sort. Bob On Dec 15, 2014, at 2:27 AM, Mike Monett timen...@binsamp.e4ward.com wrote: Hi On Dec 10, 2014, at 11:12 PM, Mike Monett timen...@binsamp.e4ward.com wrote: [...] Can you tell me some of the ones that do? I have yet to see one for under $2K that does it correctly. I don't have the cash to buy ones at those sort of prices. Some have reported that the old Motorola UT's will do it. The samples I have tried have not done very well. I may have not had them running right who knows. OK, that pretty much eliminates the NIST data. or means that more research is needed on single sat boards. There are a *lot* of them out there. What is a single sat board and can you give some examples? 5) Feed that into your control loop equation. There's another term I need to research! It's not a simple control process, but it's not all that terrible either. It just takes a bit of work to optimize. Figure a few months to a few years for the optimization depending on what you have for issues along the way. OK, so I figure out how to do this. How do I tell if this is making the gpsdo more accurate? In other words, how do I get the ADEV without having an H-Maser? You get a Cs (or other atomic standard) or you compare several different GPSDO's against each other. The preference would be for groups of three so you can rule out ones that are not doing what they should. The ideal would be 3 TBolts, 3 same model Rb's, 3 same model OCXO's and more than three groups overall. Each has their own ADEV curve. Comparing devices with vastly different ADEV is not the best way to go. Yes, that makes good sense. [...] There are many new dacs, op amps, and bipolar microwave devices that offer much lower noise than current designs use. I think the performance can be improved in some areas with new components and design techniques, and I have equipment and time to explore. There are only two points that system noise really comes into the GPSDO design: 1) The TDC must have enough resolution 2) The DAC on the EFC must have noise below the OCXO I was more thinking of the Rubidium physics package. Most of the Rubidiums on eBay are pretty old. The lamp must have constant intensity. The null detector has to identify a very shallow drop. The microwave signal needs to have excellent phase noise. These all need low noise, low flicker components, and significant advances have been made in recent years. The surrounding circuitry could probably benefit from a redesign to take better advantage of low noise components. For example, SRD's are much noisier than NLTL's. Probably little can be done to improve the Rubidium cell, but it should be the limiting factor and not the electronics. The TDC is limited by the basic resolution of the GPS system. It's easy to build one that has far more resolution than needed / useful. Some people have little faith:) The DAC issue is normally solved with a $4 part. Unless you have an OCXO with a very wide EFC range, it's noise is unlikely to be an issue. DAC resolution can be addressed to any desired level with two DAC's (fine and coarse). (Once you get going, you only use the fine DAC). The real fun and games revolves around the software used to implement the filtering / control loop between the GPS and the OCXO (or Rb, or TCXO, or MEMS, or VCXO, or Cs or âŠ) Focus on the software. Bob I have a few Rubidiums and OCXOs I'd like to get running for a month or so to stabilize. During this time, I'd like to monitor the performance to discover any bad units and see which are the best ones. For example, TVB shows a 100:1 variation in ADEV in FE-405B Rubidiums. The section is titled Variation in FE-405B in http://leapsecond.com/pages/fe405/ Clearly, it would be futile to try to use a bad unit in setting up a gpsdo. I need some means of measuring the performance of these units while they are running. A dozen HP5370's would be out of the question. I did some research to find the different methods available and decide which has the lowest per-channel cost and best performance. Here are some of the references I found. I discarded most of the poor ones and tried to keep only the ones that talk about measurements in the picosecond or femtosecond range.
Re: [time-nuts] GPS position averaging
Hi On Dec 15, 2014, at 8:29 PM, Paul tic-...@bodosom.net wrote: On Mon, Dec 15, 2014 at 6:39 PM, Angus not.ag...@btinternet.com wrote: But is it the closest to the 'true' position that you really want, or the best estimate of where the particular GPS you are testing thinks it is? I don't understand what you're saying here. Can you phrase it differently? I proposed two possible solutions to the problem: 1) Do the long term average with your specific GPS gizmo. 2) Get a survey grade GPS in and determine the “real” location. The NIST papers do indicate that #1 does not equal #2 for the gear they tested. So the question - If they are not going to be equal, which one do you pick? Bob ___ 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] Choke Rings and Points North
With all the discussion about surveys position accuracy, I have a question about my choke ring antenna. There is an arrow marked N on the underside of the rings. How accurately does the alignment need to be to North? True north or magnetic north (my thinking says True North)? Does the directional accuracy affect the precision survey? I'm assuming that it would have no effect on the accuracy of the 10 MHz frequency output. Or am I completely off base? Dave M ___ 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] GPS position averaging
On 12/15/14, 5:35 PM, Bob Camp wrote: Hi On Dec 15, 2014, at 8:29 PM, Paul tic-...@bodosom.net wrote: On Mon, Dec 15, 2014 at 6:39 PM, Angus not.ag...@btinternet.com wrote: But is it the closest to the 'true' position that you really want, or the best estimate of where the particular GPS you are testing thinks it is? I don't understand what you're saying here. Can you phrase it differently? I proposed two possible solutions to the problem: 1) Do the long term average with your specific GPS gizmo. 2) Get a survey grade GPS in and determine the “real” location. The NIST papers do indicate that #1 does not equal #2 for the gear they tested. So the question - If they are not going to be equal, which one do you pick? if you've got observables and they're in RINEX format, you can do offline processing through JPL's GIPSY thing.. If you're looking at data that's a week old, they've had plenty of time to process all the observed satellite behavior, calculate ionospheric corrections, changes in earth's rotation due to earthquakes and tsunamis, and all that sort of stuff. http://apps.gdgps.net/ Oddly, the picture of the guy on the cellphone in the left side of the banner looks like Yoaz Bar-Sever, who runs the gdgps stuff. Why he's on a phone, I'm not sure.. apps is more of a file/web sort of thing. ___ 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] Choke Rings and Points North
Hi If: 1) You are doing a survey of a number of points in an area 2) You want to hit the “1/5 mm accuracy” that your system is rated to :) 3) You really do care Then: You point the arrow north to the best of your ability to make all the antenna errors show up in the same direction. It becomes an offset to the entire grid of points rather than an error randomly added to each point. You do a correlation process at your stationary survey (DGPS) locations to remove the error. If you are not doing that sort of survey work … don’t worry about it. (Apologies to those who do this for a living, you do indeed care, it’s just the time side that does not care). I’m sure that “best of your ability” is indeed quite good. Bob On Dec 15, 2014, at 8:46 PM, Dave M dgmin...@mediacombb.net wrote: With all the discussion about surveys position accuracy, I have a question about my choke ring antenna. There is an arrow marked N on the underside of the rings. How accurately does the alignment need to be to North? True north or magnetic north (my thinking says True North)? Does the directional accuracy affect the precision survey? I'm assuming that it would have no effect on the accuracy of the 10 MHz frequency output. Or am I completely off base? Dave M ___ 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] Choke Rings and Points North
On 12/15/14, 5:46 PM, Dave M wrote: With all the discussion about surveys position accuracy, I have a question about my choke ring antenna. There is an arrow marked N on the underside of the rings. How accurately does the alignment need to be to North? True north or magnetic north (my thinking says True North)? Does the directional accuracy affect the precision survey? I'm assuming that it would have no effect on the accuracy of the 10 MHz frequency output. Or am I completely off base? If you're using a standard antenna, they've characterized them for the change in phase center with respect to the direction the signals are coming from. It's assumed you'll install it level, so elevation is taken care of. The remaining uncertainty is the azimuth, hence the north arrow. Now we can find out how much of a nut you really are. On choke ring antennas, I think the maximum shift in phase center with look direction is on the order of single digit millimeters, or a few ps. And how accurately do you know what direction is north. That could be a whole project in itself, ranging from moss on trees, to shadows of sticks and rocks, to observations of Polaris through a theodolite, and so forth. ___ 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] GPS position averaging
On Mon, Dec 15, 2014 at 8:35 PM, Bob Camp kb...@n1k.org wrote: So the question - If they are not going to be equal, which one do you pick? The better one. If the math is sound the presumably the better position results in better time. If there are correlated defects then I suppose the GPS provided solution may be better. None of the various surveys for my primary antenna (which is not optimally located) agree and all are wrong compared to carrier phase solutions. The delta between the various units (Trimble, uBlox and whatever is in the Fury) is greater than the delta between my indoor and outdoor antenna positions in xy. Z not so much. ___ 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] GPS position averaging
On Mon, Dec 15, 2014 at 9:01 PM, Jim Lux jim...@earthlink.net wrote: if you've got observables and they're in RINEX format, you can do offline processing through JPL's GIPSY thing.. According to the upload form APPS still requires dual frequency. The underlying Gipsy system may not have that constraint but I opted used CSRS-PPP to check the RTKLIB output rather than worry about it. ___ 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] GPS position averaging
On 12/15/14, 6:46 PM, Paul wrote: On Mon, Dec 15, 2014 at 9:01 PM, Jim Lux jim...@earthlink.net wrote: if you've got observables and they're in RINEX format, you can do offline processing through JPL's GIPSY thing.. According to the upload form APPS still requires dual frequency. The underlying Gipsy system may not have that constraint but I opted used CSRS-PPP to check the RTKLIB output rather than worry about it. ___ Interesting.. I'll ask them later this week when I'm back on lab. I was under the impression that they can do single frequency GIPSY processing. ___ 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] Choke Rings and Points North
With all the discussion about surveys position accuracy, I have a question about my choke ring antenna. There is an arrow marked N on the underside of the rings. How accurately does the alignment need to be to North? True north or magnetic north (my thinking says True North)? Does the directional accuracy affect the precision survey? I'm assuming that it would have no effect on the accuracy of the 10 MHz frequency output. Or am I completely off base? Dave M Dave, I agree with the others. If you're using dual-frequency GPS with post-processing it might make a few mm or ps difference. This only matters for absolute measurements, or multiple measurements where you plan to compute differentials. It also depends on the antenna. You can find the phase vs. angle profiles at http://facility.unavco.org For one example of an antenna calibration see: www.unavco.org/projects/project-support/development-testing/publications/trimchoke/trimchoke.pdf The GPS pros take their mm very seriously. If all you're using is a fixed location and commodity GPSDO then it makes no difference at all, not even close. But to prove us wrong, put the antenna on a 17 hour turn-table, collect data for 6 months, and then see if you see any 17h peaks in the FFT! /tvb ___ 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] Choke Rings and Points North
Jim Lux wrote: On 12/15/14, 5:46 PM, Dave M wrote: With all the discussion about surveys position accuracy, I have a question about my choke ring antenna. There is an arrow marked N on the underside of the rings. How accurately does the alignment need to be to North? True north or magnetic north (my thinking says True North)? Does the directional accuracy affect the precision survey? I'm assuming that it would have no effect on the accuracy of the 10 MHz frequency output. Or am I completely off base? If you're using a standard antenna, they've characterized them for the change in phase center with respect to the direction the signals are coming from. It's assumed you'll install it level, so elevation is taken care of. The remaining uncertainty is the azimuth, hence the north arrow. Now we can find out how much of a nut you really are. On choke ring antennas, I think the maximum shift in phase center with look direction is on the order of single digit millimeters, or a few ps. And how accurately do you know what direction is north. That could be a whole project in itself, ranging from moss on trees, to shadows of sticks and rocks, to observations of Polaris through a theodolite, and so forth. Thanks for the explanations. I'm not terribly concerned about time, other than knowing when it's time to eat and sleep... I'm more of a frequency nut than a time nut. I have a USGS map and recent survey of my property, so I know where North is, to a pretty good certainty. Thanks!! Dave M ___ 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] Lady Heather and comm port setup...
Hi, I think the answer is probably no, but is there a sanctioned way of coercing the fine Lady into unusual comm port settings? For instance, can I get her to do: 9600, 8, Odd, 1 without resorting to the whip? Thanks! -Chuck Harris ___ 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] Lady Heather and comm port setup...
Hi, I think the answer is probably no, but is there a sanctioned way of coercing the fine Lady into unusual comm port settings? For instance, can I get her to do: 9600, 8, Odd, 1 without resorting to handcuffs, and the whip? Thanks! -Chuck Harris ___ 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.
