-
So, I was Pegtesting my GPS-RTK yesterday and I found….
WOW, I bet that got your attention! Half of you are thinking “you can’t pegtest gps” and the other half, or probably a lot less that half, are thinking “what’s a pegtest”. OK, you can call it something else if you want but it’s basically still a pegtest. I have been noticing for quite some time that I can shoot a bunch of control points one day and set another point farther down the road for tomorrow. then when I set up on the new point the next day and tie a few shots from the day before, my vertical is off. I never could quite put my finger on what was happening. I decided it was time to pegtest my GPS.
This is a semi-permanent control point that I use regularly on a ranch. It’s a 5’+ rebar in the middle with two wooden post and two cattle panels to protect the base from cattle.
I welded the head of a 5/8″ bolt to a 3/8″ x 6″ lag screw for a GPS mount, one for each wooden post. They were adjusted to be precisely the same height shooting them with a level and shot to be 6.280 feet above the top of the rebar. Setting my Trimble SPS851 base with a Trimble Zephyr Geodetic Model 2 antenna (Pt # 57971) on the rebar, I shot the post A-B-A-B to establish what could be repeatable. My rover is the Trimble SPS881, generally considered the same as an R8.
Rebar 423.647, Calculated elevation of nut using the level shot of 6.280 is 429.927
Elevations using GPS RTK and 3+ minute observation.
Post A1 430.020
Post B1 430.016
Post A2 430.022
Post B2 430.022That gives me 4 shot with elevations ranging .006 feet but off by 0.089 to 0.095 from the elevation I shot with the level.
The only thing this is not taking into account is the depth of the punch mark in the rebar.
To eliminate one of the first things a skeptic will say, I decided to shoot from one post to the other with both height’s at zero to the bottom of antenna mount.
Then I set the base on Post A, starting it with the coordinates of
A2 and shot B11 430.099
B2, shot A11 430.094
A2, Shot B12 430.106
B2, Shot A12 430.101
A2, Shot B13 430.104
B2, Shot A13 430.101
I shot the first two before a pdop spike at 5pm. I had the mission planning on my phone and knew when to expect it. An hour later I came back and shot the last 4.That is 6 shot averaging 430.101 with the farthest off be .007 feet but being 0.079 feet higher that the 430.022 that the base was set on. By alternating the setup, it would rule out any physical error in the way I conducted the test.
I think this proves that the vertical offset in one of my receiver antennas is off 0.08 feet, or possibly a combination of the error in both receivers is off that much.What I have not figured out is how to determine which receiver is off. Any ideas?
The Zephyr vertical offset is 0.2804 on the label but measuring the thickness of it to be 0.28 feet, that would put the antenna phase center on the very top surface of the antenna. Is that even possible?
Comments on the way I conducted the test?
Thanks,
James -
33 Replies
-
Are you using Trimble Access? That SPS 881 is GLONASS capable, right? It’s NOT the same as an R8 – it’s the same as an R8 Model 2 (R8GNSS) – there’s about a 10mm difference in the phase center offsets between the two.
According to the drawings, etc., on the NGS antenna calibration site the Zephyr Geodetic II has a phase center offset from the ARP of 0.279′.
-
I would think that one of the factors you are not addressing, is the GPS orbital errors. They are systemic. Rtk has its limits. It has inherent random errors. It would be a better test, to process 24 hrs, post processed.
I think. -
This is using Survey Controller 12.50 and the Antenna Type selection for the rover in the survey style is “R8 GNSS/SPS88x”;Measured to “Bottom of antenna mount” ; Part number 60158-00
I’m just measuring it with a tape measure with a flat bar on the kitchen table but 0.279′ puts the antenna phase center on the top surface of the antenna.
James
-
Nate The Surveyor, post: 353828, member: 291 wrote: I would think that one of the factors you are not addressing, is the GPS orbital errors. They are systemic. Rtk has its limits. It has inherent random errors. It would be a better test, to process 24 hrs, post processed.
I think.Relative to each other, the GPS orbital errors would be the same wouldn’t they?
Where post processing might be useful in this case is to identify which unit has a bad number. Post process both with zero height. -
JaRo, post: 353822, member: 292 wrote: that would put the antenna phase center on the very top surface of the antenna. Is that even possible?
Yes, it’s even possible for the phase center to be above the antenna.
-
Nate The Surveyor, post: 353828, member: 291 wrote: It would be a better test, to process 24 hrs, post processed.
I think.One disadvantage of working for a construction company is convincing them that post processing software is a good thing.
James -
But I can see where you are coming from with them very close to each other they should index out the orbital errors. interesting. Maybe I’ll try it thanks.
-
JaRo, post: 353822, member: 292 wrote: The Zephyr vertical offset is 0.2804 on the label but measuring the thickness of it to be 0.28 feet, that would put the antenna phase center on the very top surface of the antenna. Is that even possible?
The antenna phase center is the electrical phase center, not a physical object. So as Jim said, it’s possible for it to be above the actual antenna.
-
Lee.
Your comment about Access does bring up another idea. I can repeat the test using the data collector with Access that I got from you just to see if anything changes. We have an R10 that is UHF floating around the company somewhere. Adding it to the mix may also answer some questions. Or add more questions.James
-
I would definitely try that, just to make sure Survey Pro isn’t doing anything weird.
-
The wandering thoughts of a cattle producer got me confused as I thought the word you used in the title of this thread was “pregtest”. Preg testing is standard practice in the cattle industry to find out as early as possible whether or not the cow is pregnant.
View the standard process at: [MEDIA=youtube]0bSk0KM3Ya0[/MEDIA]
-
I don’t know if it is possible with the gear in question, BUT if you can put user defined antenna definitions into the gear, I would start by using the NGS absolute calibrations loaded to each unit and see if that works better.
The receivers are measuring the deltas from the electrical phase center of each antenna, if that isn’t determined correctly in relation to the ARP, the results will suffer. Having the model of that electrical phase center/ARP relationship from an independent testing such as the NGS may be helpful in resolving the issues.
Using different models of antennas on the base and rover with a less than perfect antenna definition will cause havoc, if both where the same model it wouldn’t matter what definition you used or for that matter any definition, the differences would then simply be in manufacturing. Personally, I believe antennas at CORS are the biggest contributor to unexpected results, especially in the vertical. This is why RTN vertical has more issues vertically than a local base (even when very close to a CORS). Despite NGS and manufacturer’s modeling all the various antennas, the antenna differences can still be an issue at times IMO. I can’t tell you how many times I have found the wrong antenna listed on a CORS datasheet, you know where the photo doesn’t match the station log or the RINEX header doesn’t match either, etc. Antennas get changed, often new photos and paperwork don’t. It is a large job for an agency to track all the antennas, especially if there are multiple agencies or people involved, stuff does slip through the cracks despite best efforts of staff.
SHG
-
JaRo, post: 353822, member: 292 wrote: WOW, I bet that got your attention! Half of you are thinking “you can’t pegtest gps” and the other half, or probably a lot less that half, are thinking “what’s a pegtest”. OK, you can call it something else if you want but it’s basically still a pegtest. I have been noticing for quite some time that I can shoot a bunch of control points one day and set another point farther down the road for tomorrow. then when I set up on the new point the next day and tie a few shots from the day before, my vertical is off. I never could quite put my finger on what was happening. I decided it was time to pegtest my GPS.
This is a semi-permanent control point that I use regularly on a ranch. It’s a 5’+ rebar in the middle with two wooden post and two cattle panels to protect the base from cattle.
I welded the head of a 5/8″ bolt to a 3/8″ x 6″ lag screw for a GPS mount, one for each wooden post. They were adjusted to be precisely the same height shooting them with a level and shot to be 6.280 feet above the top of the rebar. Setting my Trimble SPS851 base with a Trimble Zephyr Geodetic Model 2 antenna (Pt # 57971) on the rebar, I shot the post A-B-A-B to establish what could be repeatable. My rover is the Trimble SPS881, generally considered the same as an R8.
Rebar 423.647, Calculated elevation of nut using the level shot of 6.280 is 429.927
Elevations using GPS RTK and 3+ minute observation.
Post A1 430.020
Post B1 430.016
Post A2 430.022
Post B2 430.022That gives me 4 shot with elevations ranging .006 feet but off by 0.089 to 0.095 from the elevation I shot with the level.
The only thing this is not taking into account is the depth of the punch mark in the rebar.
To eliminate one of the first things a skeptic will say, I decided to shoot from one post to the other with both height’s at zero to the bottom of antenna mount.
Then I set the base on Post A, starting it with the coordinates of
A2 and shot B11 430.099
B2, shot A11 430.094
A2, Shot B12 430.106
B2, Shot A12 430.101
A2, Shot B13 430.104
B2, Shot A13 430.101
I shot the first two before a pdop spike at 5pm. I had the mission planning on my phone and knew when to expect it. An hour later I came back and shot the last 4.That is 6 shot averaging 430.101 with the farthest off be .007 feet but being 0.079 feet higher that the 430.022 that the base was set on. By alternating the setup, it would rule out any physical error in the way I conducted the test.
I think this proves that the vertical offset in one of my receiver antennas is off 0.08 feet, or possibly a combination of the error in both receivers is off that much.What I have not figured out is how to determine which receiver is off. Any ideas?
The Zephyr vertical offset is 0.2804 on the label but measuring the thickness of it to be 0.28 feet, that would put the antenna phase center on the very top surface of the antenna. Is that even possible?
Comments on the way I conducted the test?
Thanks,
JamesNot sure what’s going on here, so many things can cause this.
What I can say is that we have always checked our set-up (peg them if you will), and we don’t see what you are seeing.
Lately the vertical is getting very stable, not like the old days of RTK.
I would do more testing till I figure it out, taking many shots is interesting but something else is going on here, your GPS should check your levels much better than that. We usually see them less than .03′ doing what you are doing, not good enough for curb and gutter, but for many tasks plenty good.You are thinking about it and doing checks, some just assume it works and head out.
I really don’t see static as something that will catch this issue.
-
When I first got this setup, my first experience with GPS, The salesman set it up with the antenna 55971 in the survey style. It was several years later before I caught it and that was from sending in files to opus. At that time I emailed NGS because on the opus page, the name was correct on one antenna and the part number was correct on the other. This is the response I got back from NGS.
You should select TRM 57971.00 None Converted from Absolute igs05_1480.atx. Trimble used to use lead in the solder for the 55971, but stopped. Trimble says the model you have is identical in every way except for no lead in the solder, but did not submit samples of the revised model to us for testing. The calibration for your model is from an IGS listing, converted and modified to be compatible with NGS relative calibrations. I will agree with you, the description in OPUS for your model is lacking, but the part numbers are correct.
Charles Geoghegan
Not sure it makes any difference. I may try the 55971 antenna just out of curiosity.
-
Well you sure got me to thinking now I can set both of these up on the same level and I can set the rover up underneath the base that would give me a double check about horizontal and vertical I’m going to mess with it seems like a good idea thanks for the commentary.
-
my antenna has relative calibration and its electrical phase center is 36.4 mm above top of physical antenna radiator.
-
Yuriy Lutsyshyn, post: 353913, member: 2507 wrote: my antenna has relative calibration and its electrical phase center is 36.4 mm above top of physical antenna radiator.
WOW, I would not have thought that was possible.
James -
NGS relative calibration of my antenna:
MRAN1+NONE 11-312 s/n: 0003
2.5 0.3 74.3
0.0 -0.4 -0.3 0.3 1.1 1.8 2.4 2.7 2.6 2.2
1.3 0.0 -1.7 -3.7 -5.9 -8.1 -10.3 0.0 0.0
0.0 0.0 000.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0actually it is 37 + 1.59 – 74.3 = 35.71 above radiator plus variations caused by sat elevation
You might have applied wrong calibration . Also relative and absolute calibrations are different.
-
I’ve noticed that the NGS and IGS modelied phase center offsets from the ARP of my old Zephyr Geodetic antenna are different from what Trimble specifies and uses in their Survey Controller and office software. Also that NGS OPUS doesn’t appear to use the RINEX header info, but NASA APPS does.
-
astrodanco, post: 353930, member: 7558 wrote: NGS and IGS modelied phase center offsets from the ARP of my old Zephyr Geodetic antenna are different from what Trimble specifies
That’s true for all of the Trimble antennas I’ve worked with. As long as you don’t mix offset sources, I figure the differences cancel out.
Log in to reply.