@rftop I did some tests using different RTK methods, I’m going to start with two single baseline tests. I also should note that I tried giving the corrections to the instrument through bluetooth on SW maps and it reminded why I stopped trying that. It would randomly lose bluetooth connection every couple minutes if I wasn’t actively watching my phone which makes it unusable for me. These tests are all still through the internal connection.
Here is a single baseline correction using a CORS site which is ~38 km from my instrument in the same network as the one I showed last week:
It could be chance but this does look a little better than the plot I showed above. It still does have a pretty dramatic drop of ~25 cm, could this just be related to the long baseline?
Here is a different single baseline correction using a CORS site which is ~24 km away:
Again, could be chance but this looks a bit tighter than the above plots. I’m a little confused about the drops and jumps that are observed at 18:45, 19:23, and 19:33. Does anyone have an explanation for those?
What would be helpful to have some insight on is could all the differences in these plots and the one from last week be explained by multi-path problems and distance from the CORS site?
I can tell you I am using a 3 cors station network solution and my facet heights are usually crap like .1 or more between two points, yet my postcard with l5 antenna will be close to 0.03 h and less than .1 vertical. The other day I saw 0.79 vertical between two points set by my facet. I never use it for vertical accuracy for survey work I do.
Yesterday however when I was locating property corners for someone I used the facet and my phone and walked right to the corners on my calc. So horizontally it’s been a great tool.
And this is after getting the arp correct in the software. Benchmarks are +/- 0.15 or more whereas my postcard can hit it less than 0.05 most times.
Definitely useful to have some confirmation of my experience. Are your units meters? I.e. the postcard gets ±5 cm most of the time? That would be exactly what I’m looking for. Are you using the antenna suggested in the hookup guide for the postcard (helical or surveying)?
I would be very interested to see your comparison of the four devices. Surveying is a small part of my job, but I need fairly high accuracy vertically for my surveys. I can’t justify paying >$5k for an instrument but I can assemble something from components if I know there are components that will work.
I am a professional land surveyor in my state and I use a total station for most accurate things but I use the postcard for finding pins, establishing a baseline in state plane coordinates and most things. It has kept up with most name brand equipment in my tests.
@sparky really just needs to build an L5 antenna, postcard or mosiac, and Lora into a unit and make it water proof and rugged.
I think the mosaic facet is also on that level but again waterproof I like to survey in the rain so my tablet touchpad is a royal pain to work with
All my experience before experimenting with the Facet was total station as well. I recognize this is a much more accurate method but I’m surveying in tide gauges that NOAA uses and their method is for GNSS so I wanted to try to replicate it more affordably.
That’s a great setup you’ve got there, I may try to replicate it for further testing to see if I can reach 0.1 feet vertical as well. It looks pretty simple: is it just the postcard with a portability shield, the survey antenna, and a lora module? I’ll buy the components for testing and then try to waterproof it later.
Yes and you do not need Lora unless you dont have a good real time network. Here in Oregon we have a great one so I can usually not need a base station. But I’ve been experimenting to pull out all the tricks.
I do 3 minute observations and average them for stuff I need. Or rapid static/static would be another way.
I have not tried static with the facet to see if verticals would be on par. I do know the antenna offsets on the sales page are not correct.
I noticed they are a couple cm different from the NGS calibration file. I’m assuming the NGS file is correct. But this wouldn’t impact the consistency of my measurements.
I’ve tried all kinds of different versions of this - rapid static, static, with and without corrections, OPUS processing, NRCAN processing - and sometimes I’ve been within an acceptable error range but mostly I have not been and it’s very difficult to predict when the data I’m gathering is good before I go back to my office and process it. Like you have mentioned, the horizontal seems to be great on the Facet but unfortunately the horizontal is secondary to the vertical in my application.
What software are you using?
Luckily, the postcard is an inexpensive option that I can still justify testing out, hopefully I get as good results as you do. It would still be great if I could figure out a way to use the Facet for this though, I suppose I could observe all day on one spot…
So in the woods the facet can outperform the postcard? At this point I just need something to work under an open sky, the majority of the locations I need to survey are in the open because they are coastal but I do have some rivers that are under tree canopy too. I’ll get into that when once I have an established method. I have an auto level and can certainly do that if needed.
It’s entirely possible I’m making a mistake here because I’m new to this. These are plots of the data from the sd card on the Facet. Regardless of whether I’m using corrections or not, the output is always the same format but there’s definitely a difference in the way the data looks. For one, the software I’m using plots RTK fix points in green, RTK float points in yellow, and DGPS points in blue. I posted a plot above which is very noisy and shows mostly blue points which is an observation without corrections. All the rest of the plots with green points are using corrections. That being said, these are not processed at all, and they are the “raw” observations that come from the instrument after the RTK correction. I’m displaying them just to show how much drift the instrument is observing over my measurement period. Let me know if I’m missing a step or should be considering something else.
These are all at least a half mile from the water, they’re just tests in my yard which, sadly, is not right on the ocean.
You might not want to bother with this, but here’s a simple way to verify the precision (repeatability) of the System, that everyone can perform:
Install the Facet on a position/point that can be repeated.
Store RTK Fixed solutions only in SW Maps. Do this many times, spread-out over time. The idea is you want different Satellite geometry, different atmospheric conditions, etc.
Export the Points to a spreadsheet with UTM turned ON.
UTM allows you to quickly subtract the smallest value from the entire data set (X, Y, and Z individually).
You can easily see the range of values, because all you are left with are a couple of centimeters over the entire data set (hopefully).
Yeah it’s a good method. I may give it a shot but I’m fairly sure what the outcome will be. I essentially have built up a database of measurements on various fixed locations that I’ve tested the instrument on over the last few months, using different occupation lengths, different correction sources, and different processing methods. I know that’s not exactly the same as what you outlined but it has proven to me that the methods I’m using haven’t been consistent enough. While I was doing one of the tests late last week, I manually recorded the elevation from SW maps periodically and these are the numbers I got:
Time - elevation (m) - fix type
3:13:25 - 24.630 - fix
3:16:20 - 24.720 - fix
3:21:35 - 24.745 - fix
3:26:40 - 24.586 - fix
3:34:05 - 24.644 - fix
3:39:00 - 24.655 - fix
Even over the course of 26 minutes, with the instrument in RTK fix, the elevation was varying by 16 cm, which you can see in this plot (it’s in UTC so 3:00 above = 19:00):
I just looked at the Facet documentation and there is the line:
“The NMEA sentences will have increased accuracy of 14mm horizontal and 10mm vertical accuracy”
I thought maybe I was misremembering that the vertical accuracy was also supposed to be essentially 1 cm. So I guess this is a question for @sparky but I’m curious what the conditions might be that would give you that level of vertical accuracy. I’m certainly not within 10 km of the CORS station here and I have a slightly obstructed southern sky so I recognize I don’t have ideal conditions. I have used this instrument in much clearer settings however and ~10 km from a CORS site without obviously getting the accuracy I need.
But I probably need to do these tests again under more ideal circumstances to know for sure.
