Height values with Trimble R10

That is concerning.
It has been almost 2 years since I used one but static processing with TBC checked in ok then. TBC selected the antenna model.
I don't think I used RTX though, and we don't have OPUS

squowse, post: 435582, member: 7109 wrote: That is concerning.
It has been almost 2 years since I used one but static processing with TBC checked in ok then. TBC selected the antenna model.
I don't think I used RTX though, and we don't have OPUS

I haven't done TBC comparison yet, but I certainly will.

I was just tasked with sitting on some existing control.

I do not know what type of receiver was used for the given numbers.

Sending in the occupied R10 file to OPUS returned an EL HGT of 1584.092m, Record was 1584.089m.

I don't see this error that you are during my experience with the R10's just completed a FEMA survey; a 1st order bench was on site, we tied it from control established from a first order bench/HARN point ten miles distant and hit the on-site bench by .01', in this case applying GEOID09 to the heights, but holding the elevation of the HARN point and not the ellipsoid height on the data sheet.

Although, I will say I simplify the export process for OPUS by using tools supplied by Trimble,,,,,,,,,,,,,, no Rinex conversions.

OPUS values and RTX-PP values usually differ for me in by .05' to .15' vertically. The few times I've checked known bench marks the RTX-PP usually is closer to elevations than OPUS, but there isn't a huge difference.

These kind of "anomalies" are very worrisome, and not as rare as they should be.

Loyal

MightyMoe, post: 435587, member: 700 wrote: I was just tasked with sitting on some existing control.

I do not know what type of receiver was used for the given numbers.

Sending in the occupied R10 file to OPUS returned an EL HGT of 1584.092m, Record was 1584.089m.

I don't see this error that you are during my experience with the R10's just completed a FEMA survey; a 1st order bench was on site, we tied it from control established from a first order bench/HARN point ten miles distant and hit the on-site bench by .01', in this case applying GEOID09 to the heights, but holding the elevation of the HARN point and not the ellipsoid height on the data sheet.

Although, I will say I simplify the export process for OPUS by using tools supplied by Trimble,,,,,,,,,,,,,, no Rinex conversions.

OPUS values and RTX-PP values usually differ for me in by .05' to .15' vertically. The few times I've checked known bench marks the RTX-PP usually is closer to elevations than OPUS, but there isn't a huge difference.

Well, if I do a survey and use all R10's, and tie in some benchmarks, etc, and then adjust holding the benchmarks, then it isn't an issue as the "error" will cancel out. However, when using OPUS, or RTX, or mixing receivers, antennas, etc, or holding CORS, then it would be a problem.

I was thinking at first it was my R10, but then I tried a different R10. Not ready to say with certainty, I am going to make more observations, process in TBC, etc. I was just posting this preliminary data to see if anyone else had similar issues.

What made me suspicious is when we occupied some benchmarks with R8 and then R10 and got results different by about 3 or 4 cm. Of course that happens, but it was repeating itself on different projects. Most projects that I use an R10 on, I only use an R10. It is in those old school "network" type surveys with multiple occupations with different receivers that I began to notice it.

OPUS is not as authoritative as we would all like for it to be. A few years ago I tested a receiver on a pedestal for a couple of weeks and found that the phase center offsets were not being applied properly. It affected horizontal and vertical accuracy. The suggestions I received from NGS were extremely disappointing. Might not be the receiver.

John Hamilton, post: 435589, member: 640 wrote: Well, if I do a survey and use all R10's, and tie in some benchmarks, etc, and then adjust holding the benchmarks, then it isn't an issue as the "error" will cancel out. However, when using OPUS, or RTX, or mixing receivers, antennas, etc, or holding CORS, then it would be a problem.

I was thinking at first it was my R10, but then I tried a different R10. Not ready to say with certainty, I am going to make more observations, process in TBC, etc. I was just posting this preliminary data to see if anyone else had similar issues.

What made me suspicious is when we occupied some benchmarks with R8 and then R10 and got results different by about 3 or 4 cm. Of course that happens, but it was repeating itself on different projects. Most projects that I use an R10 on, I only use an R10. It is in those old school "network" type surveys with multiple occupations with different receivers that I began to notice it.

I know what you mean, it might be possible that the control given me was processed using a R10, it didn't say on the data sheet I was given.
I have been mixing R10, R8, 5800 and L1/L2 antennas, and have not seen anything like you are, but I will continue to look for it.

I processed it in TBC. One thing that didn't go as planned is that TBC now automatically concatenates separate files that are consecutive for the same point. Since I had 24hr files, and the CORS (PAAP) is also 24 hour files, it processed only 5 baselines, one to the 5700, two to the R7 (not sure, I think there may have been a gap), and one to each of the R10's.

Delta Ellip H from the CORS:
5700 with Zephyr Geodetic: 28.741 m (6 days and 5 hours of data)
R7 with Zephyr Geodetic 2: 28.748/28.743 m (9 hours/4 days of data respectively)
R10 #1: 28.768 m (24 hours)
R10 #2: 28.769 m (7 hours)

Looks like about 0.025 m difference.

I believe the whole idea of antenna models is to eliminate these issues. I selected NGS absolute. I might try again using other models.

As I said before, these are all exactly on the same point, the tribrach and adapter were not changed, and the ARP's were flush against the adapter (the R10's had the extra 0.05 m adapter, which was taken into account).

Not to sound patronizing, but you are referencing your HI to the bottom of the R10, not the bottom of the quick release, correct? The quick release is 5cm.

- just saw your last post, so never mind. This is definitely concerning however I 've never seen it. May have to do some testing.... I just ran a static network with 5 R8s and one R10 so I have some data I can look at.

Have you tried all three measure points?
I would give that a shot, let Trimble figure out the final number when you process or send it off.

ARP is the standard reference point, it is NOT setup on a tripod.

The one I referenced above was measured to the lever, of course that would be difficult on your pedestal, sounds like there is something going on during the measurement or conversion.

John Hamilton, post: 435570, member: 640 wrote: I have been suspecting that there is something wrong with the antenna info/offsets for the Trimble R10. I am seeing about 3 to 4 cm error.

So, I recently ran multiple day long sessions on my pedestal. Same setup, same tribrach and adapter, different antennas/receivers.Submitted to OPUS-S.

Here is what I get for ellipsoidal height:
R7+Zephyr Geodetic 2: 342.486 (mean of 5 days, sdev=0.007 m, spread=0.018 m)
5700+Zephyr Geodetic: 342.486 (mean of 7 days, sdev=0.007 m, spread=0.019 m)
R10: 342.528 m (mean of 3 days, sdev=0.004 m, spread=0.007 m)

Yesterday I put a different R10 on the ped, it is too soon to send to OPUS.

I did send all of the data (including the other R10) to the Trimble RTX processor, and got similar results, and also horizontal differences of about 1 cm. Here are plots of the RTX solutions. (the R10's are 5432 and 4240, 5700 is 8940, and the R7 is 3772):

Vertical:

and horizontal:

This is alarming to me. I might be able to understand the OPUS differences depending on the antenna model used. But RTX is a Trimble processor, and one would think that things are correctly modeled. Clearly there is a difference in the height and to a lesser extent the horizontal position of the R10 versus the Zephyr Geodetic antennas (one is GPS only, the other is GNSS). I am going to collect more data , and use some other receivers.

Hi John. I'd like to follow the developments of this issue with the R10 heights as you pursue your investigation. I'm kind of new to using this Tapatalk app, so I'm hoping leaving this reply will give me alerts to your thread here. Thanks - Jesse

Sent from my iPhone using Tapatalk

I am in Mexico for a week, so i won't be doing anything until after I get back. I would like to hear if anyone else has seen anything like this.

It is not just an OPUS issue, I saw it using RTX, which you would think would NOT have this problem, or that someone would have noticed it before. We are not talking mm here, rather several cm.

I wanted to do some testing since John's post raised important questions.

I ran two sessions over the same control point, the same tripod and tribrach, two Trimble units-an R10 and R8.

I only did 5 hours for each unit, the R10 in the morning and the R8 in the afternoon.

These are the resulting ellipsoid heights in meters.

R10-Opus 1142.685, RTX-PP 1142.667
R8- Opus 1142.662, RTX-PP 1142.661

It does appear that when sending to OPUS there is something going on with the R10 measure point, I also processed the point against the nearby CORS station using TBC and came up R10-1142.683 and R8-1142.686.

The points used by OPUS were identical for each OPUS session.

The difference of 2.4cm is about 1 inch.

horizontally there was 5mm's difference.

Interesting that you get the same height (6 mm difference) in RTX for the R10 and the R8. I did not submit an R8 file, just files from an R7 and a 5700. I will collect some r8 data over the next few days and submit. I have to lug a big battery out there. I am planning on running 12V power to the pedestal. I bought the wire, just need to pull it through the conduit. I ran an antenna cable into the building, so I can run anything with an external antenna inside, but the R8 and R10 need power out there.

I may have read something about this in GPS for land surveyors . I don't know what revision it was but the author talked about GPS units not knowing exact antenna phase center, and as new models came out they just get closer and closer to solving the problem. So if you mix older models with new ones would that make a difference? Just lent the book out so can't check, iam probably way off base but this is why I go to this forum, to learn.
Flame suit on!

John Hamilton, post: 438146, member: 640 wrote: Interesting that you get the same height (6 mm difference) in RTX for the R10 and the R8. I did not submit an R8 file, just files from an R7 and a 5700. I will collect some r8 data over the next few days and submit. I have to lug a big battery out there. I am planning on running 12V power to the pedestal. I bought the wire, just need to pull it through the conduit. I ran an antenna cable into the building, so I can run anything with an external antenna inside, but the R8 and R10 need power out there.

I used a cowbell for these sessions, it would probably run for 24 hours, although I haven't tested it that long.

superfly, post: 438410, member: 8421 wrote: So if you mix older models with new ones would that make a difference?

Antenna phase center modeling is pretty well-established science, so I'd be inclined to think that some sort of data entry error in the model file is to blame.