Activity Feed › Discussion Forums › GNSS & Geodesy › Removing Geoid for Machine control. (TRIMBLE)
Removing Geoid for Machine control. (TRIMBLE)
Posted by scrass on April 10, 2023 at 5:29 pmHello all,
I have a TBC related question.
As you may or may not know, the Trimble machine control files must contain a .cal or .cfg created from a .dc calibration file. Our particular company subcontracts our control to be setup by a licensed surveyor. They then email me the TBC file with the control network and all applicable information (design etc). In a perfect world I can take the .dc file they generated setting up the site and convert it to a .cfg or .cal for our machines and walk away. However, the problem is this ?? the machines will not read anything with a Geoid. The suggested workflow from Sitech is summarized as follows ??load up a .csv with grid coordinates into a new project, set up base over an unknown position (here point), tie the points from the .csv and calibrate? this method works well for smaller sites; however, the scope of our work is miles long with sometimes hundreds of control points. The data we receive is good data from a licensed surveyor that is also leveled. We are then introducing error by calibrating on top of it. I would like to figure out how to blow out the Geoid without removing the solution or changing any of the data we receive from the surveyor.
One way of doing this is to create a new TBC project, drop in the design, linework, csv (in grid), and then an xml or csv file in (lat long el) or keying in the lat long el for each point and calibrating the project manually in TBC ?? this in essence still uses the Geoid in building the control network however the .dc file it generates is simply a calibration lacking a geoid (What we need for the machines to read it).
Another way of doing this is to go out and use Sitech??s suggested workflow.
Please reach out to me with any and all suggestions, comments, questions so that I can create a successful, accurate workflow.
Sincerely, Sam
rover83 replied 1 year ago 10 Members · 28 Replies- 28 Replies
Two things strike me as curious, if you are working on projects miles long in grading operations, why would the data not be scaled to ground as working on the grid could have a sizeable impact in elevations and distances. A site calibration across miles of road is not going to provide the correct solutions if it is on grid and without a geoid, you will not get the proper corrections. In the case that you present, over a large distance the errors will increase in both horizontal and vertical working on the grid.
Taking the data that your Licensed Surveyor supplies you and modifying it in any way, even by simple conversions, takes the liability off of his back and puts it on yours. Your best solution is to have your LS speak to the technical people of whatever software the blade control works off of so they can arrive at a solution that removes you from trying to make things work. If it can be worked out, all you should have to do is simply upload the data provided without modification or conversion.
I’m not the best expert on this so somebody please correct me if I’m wrong but what I’ve said is what has been taught to me over the years.
I’m not any more familiar with machine control other than knowing it exists and is used. It does not seem correct for GPS machine control systems to be unable to use a geoid. I can’t fathom a machine control manufacturer creating software incapable of using a geoid.
If you’re getting the DC file from the surveyor, and they have already run a calibration that you are merely transforming from one file type to another, why not just have the surveyor calibrate without a geoid, and send you that file?
“…people will come to love their oppression, to adore the technologies that undo their capacities to think.” -Neil PostmanI’ve been told by machine control operators that they have to calibrate just like Scrass says. However, that was some time ago, 2014 as I remember.
If that’s still the case and you have a projected/geoid TBC file from the surveyor, there is no reason to go to the field and calibrate. The better option would be to calibrate in the office and then set on control and check them. Miles long calibrations are not a great idea but if you have many control points simply sit down and do the calibration in TBC. It’s trivially simple to do, create a .csv file of lat, long, heights, a .csv file of N, E, Elevations and go to town.
I’ve been told by machine control operators that they have to calibrate just like Scrass says. However, that was some time ago, 2014 as I remember.
If that’s still the case and you have a projected/geoid TBC file from the surveyor, there is no reason to go to the field and calibrate. The better option would be to calibrate in the office and then set on control and check them. Miles long calibrations are not a great idea but if you have many control points simply sit down and do the calibration in TBC. It’s trivially simple to do, create a .csv file of lat, long, heights, a .csv file of N, E, Elevations and go to town.
The newer Earthworks software supports Geoids, the older Trimble scs900 does not. I agree with all the posts above about not messing with anything, however if I receive a .dc file from a surveyor with site perimeters that include a geoid it will not work with most of our equipment. (Hence the reason for this post) I come from surveying so I understand that using a geoid to build a control network is the best practice – and I intend to use the exact data set the surveyor gives me, however I need to calibrate to it as a mean of keeping the data the same while removing the geoid. I understand the concept of site calibrations and have done them many times in the field with access but am having a hard time figuring out best practices for doing it in the office with two csv (one gnss and one grid) this is where I need help.
Two things strike me as curious, if you are working on projects miles long in grading operations, why would the data not be scaled to ground as working on the grid could have a sizeable impact in elevations and distances. A site calibration across miles of road is not going to provide the correct solutions if it is on grid and without a geoid, you will not get the proper corrections. In the case that you present, over a large distance the errors will increase in both horizontal and vertical working on the grid.
Taking the data that your Licensed Surveyor supplies you and modifying it in any way, even by simple conversions, takes the liability off of his back and puts it on yours. Your best solution is to have your LS speak to the technical people of whatever software the blade control works off of so they can arrive at a solution that removes you from trying to make things work. If it can be worked out, all you should have to do is simply upload the data provided without modification or conversion.
I’m not the best expert on this so somebody please correct me if I’m wrong but what I’ve said is what has been taught to me over the years.
- We don’t work on the grid, we work in a localized (calibrated) site. The surveyor builds the control network and then we take the grid data that he created with a geoid and calibrate to it as a means of matching his same measurements but eliminating the geoid. This is because our system won’t read the geoid. We essentially hold the geoid corrected elevations without the software having to process that they come from a geoid instead it thinks they come from a calibration. Just the way Trimble decided to do it… dumb I know….
If you’re getting the DC file from the surveyor, and they have already run a calibration that you are merely transforming from one file type to another, why not just have the surveyor calibrate without a geoid, and send you that file?
often times that isn’t an option. We can get a .dc or matching pairs in lat long, and I have to build the calibration from that (what I’m mainly asking about) or do it how Sitech wants – by manually going and calibrating with my gnss receiver.
I’m not any more familiar with machine control other than knowing it exists and is used. It does not seem correct for GPS machine control systems to be unable to use a geoid. I can’t fathom a machine control manufacturer creating software incapable of using a geoid.
that’s Trimble for you…. The newer stuff will take a geoid but most of our stuff is pretty old and is not getting replaced anytime soon… hence the asking for help building calibrations in TBC from two data sets .csv in grid and .csv in lat long
If the surveyor did not calibrate but simply did a scale for producing the grid to ground if thats the case in theory simply have them save that scale parameters “save as site” then you should be able to set your base over a known and use that config file in theory. Like a site calibration. I have never tried this with machine control before but in theory it would work mathematically. Now if they are on true grid for miles and the design was on grid then to me this is where using true grid is best if used correctly. As even scaling to ground for miles actually can cause more distortion thats not accurately accounted for because the earth is not flat and atleast the projected coordinate system is and if designed on that system would be better as it already has the commutated distortion taken into consideration. However i have not found many surveyors or engineers that actually do this. It seems they sometimes go the opposite way and pretend ground is more accurate. I don’t know how the construction side is set up and all because most historically did site calculations. But a long linear project might require you to perform multiple calibrations along the route and use equations stations from one calibration to another in order to best fit the existing control. This can be calculated based on the design roughly and existing conditions to figure out just how far you can go between calibrations.
Have you considered hiring the surveyor who set the control do the calibration in TBC for you? Then you can field verify it. If you don’t want to do that there is always YouTube. There are quite a few Trimble training videos there.
Contact me offline [email protected] we have a workflow for you. Sitech and Trimble have been pulling this crap for years, “saying it can’t be done”. The way I see it either they just don’t understand GNSS or they want to sell you all new equipment. Most likely the later, since they get really upset when I teach my clients how to work with Geoids.
Have you considered hiring the surveyor who set the control do the calibration in TBC for you? Then you can field verify it. If you don’t want to do that there is always YouTube. There are quite a few Trimble training videos there.
Yes this is what I plan to do when for this next project – however I would like to figure it out myself as that isn’t always an option.
That’s good to hear.
I can’t understand why geoids were kept out of machine control equipment. In the early days of GPS that probably made sense. Calibrations were about the only way to simulate somewhat decent elevations when Geoid 90-96-99 were all that was available. Calibrations never made sense for horizontal control. We stopped using them for elevation control early on since they never worked in our messy geoid environment.
Then Geoid 03 showed up and changed everything. Geoid 18 is an elegant model, can’t say anything bad about it.
I have never done a calibration in TBC, so I did one this morning, HOLY Cr@P!!!! What a cluster that thing created. We’ve been checking into these control points less than 1cm routinely, sometimes a bit more vertically. The calibration changed some of the elevations more than 3cm. Unacceptable!!!
That’s good to hear.
I can’t understand why geoids were kept out of machine control equipment. In the early days of GPS that probably made sense. Calibrations were about the only way to simulate somewhat decent elevations when Geoid 90-96-99 were all that was available. Calibrations never made sense for horizontal control. We stopped using them for elevation control early on since they never worked in our messy geoid environment.
Then Geoid 03 showed up and changed everything. Geoid 18 is an elegant model, can’t say anything bad about it.
I have never done a calibration in TBC, so I did one this morning, HOLY Cr@P!!!! What a cluster that thing created. We’ve been checking into these control points less than 1cm routinely, sometimes a bit more vertically. The calibration changed some of the elevations more than 3cm. Unacceptable!!!
My understanding is the machine is trying to figure out so much about the machine (blade height, blade rotation, etc.) that they built it that way for the purpose of saving processing power.
@MightyMoe I have the same problems building a calibration in TBC, I must be doing something wrong. If you export the grid points (built with a Geoid) to a csv and then build a calibration using lat long ht (the points that the aforementioned grid points derived from) and a wgs 84 projection, you would essentially be doing what I need… I think lol…
I don’t have any problems building it, the problems come with the calibration not being able to match an inclined plane to the Geoid. I purposefully chose a job near the toe of a huge mountain to see if calibration can handle that environment. The answer is a clear no!!
The process is exporting the given control points in two files LLH, NEE. Then change the point names in one of the files, I did the NEE. I changed point 5 to 5X, point 6 to 6x ect.
Then match LLH point 5 to NEE point 5x during the calibration after importing the .csv files into a “clean” TBC file. Be sure to have all the units correctly set in the new file to match the export file. So the geographic calibration points will be 5, 6, 7, 8 ect. the grid points will be 5X, 6X, 7X ect. The resulting error horizontally will be perfect since you’re using projected NEE values. The vertical will be the one that works in a quiet Geoid environment and won’t work very well in a messy Geoid environment.
I don’t have any problems building it, the problems come with the calibration not being able to match an inclined plane to the Geoid. I purposefully chose a job near the toe of a huge mountain to see if calibration can handle that environment. The answer is a clear no!!
The process is exporting the given control points in two files LLH, NEE. Then change the point names in one of the files, I did the NEE. I changed point 5 to 5X, point 6 to 6x ect.
Then match LLH point 5 to NEE point 5x during the calibration after importing the .csv files into a “clean” TBC file. Be sure to have all the units correctly set in the new file to match the export file. So the geographic calibration points will be 5, 6, 7, 8 ect. the grid points will be 5X, 6X, 7X ect. The resulting error horizontally will be perfect since you’re using projected NEE values. The vertical will be the one that works in a quiet Geoid environment and won’t work very well in a messy Geoid environment.
@mightymoe I understand the matching pairs part, but lets say, hypothetically, I receive a csv file of NEE grid points from a surveyor, and want to build the calibration in TBC – Should I ask for them to give me the local or global LLH?
When I drop the grid csv it wants a project location, then when i drop the LLH it does the same. then I compute the project as you stated and my residuals are way off. I figure this is due to a missed or wrong step. It is my understanding that these coordinates should match exactly as the grid points are derived from the GNSS points plus the various datums and geoids etc. and my goal is to match them together (without using a geoid) only using a an inclined plane calibration. I have to be missing a step.
Thanks.
I did mine using a DOT project scaled to ground using their Project Adjustment Factor. I matched the Global LLH to the Surface Grid Coordinates. Because it’s based on a projection (State Plane) there was no horizontal error (residuals) except for some .0001′ type notations. So you will end up with three different coordinate sets a Global, a Local (they won’t be identical like the projected file) and a Grid.
But the verticals are where it fell apart. They didn’t work at all, please understand that these points lie near a huge mountain range, as you head west you quickly head up the face, it’s not Kansas. If you’re in flat area it should work for restricted areas, I would never push a calibration very far. If you’re having a problem then either you’re entering incorrect information or the original control is bad.
but lets say, hypothetically, I receive a csv file of NEE grid points from a surveyor, and want to build the calibration in TBC – Should I ask for them to give me the local or global LLH?
Someone needs to go out and measure those same points with a GNSS – using the same base station or corrections source, coordinate system and settings that the machinery will use.
The calibration developed from that ties where the machines are with the numbers from the plan.
In my experience the design engineers will provide the benchmark positions, the contractor will have a separate surveyor (inhouse, or consultant) provide the calibration.
The residuals in the calibration give a measure of the accuracy of the engineers benchmark numbers – but a careful surveyor may run an independent check – possibly resurveying surveying those points with conventional instruments and levels. I would also be wanting to see a check survey done on the benchmarks using the calibration.
You could use the original surveyor, if they are good it should not matter, but I prefer to have an independent check or three.
A project many miles long with hundreds of control points and the contractor doesn’t think this is the right time to upgrade machine control equipment. How much money did they leave on the table? Whatever the surveyor provides you with make sure you check it twice. When things go wrong nobody wins even if they’re not the one holding the bag.
Should I ask for them to give me the local or global LLH?
Global, most likely, unless there’s some goofy stuff going on in the project.
A project many miles long with hundreds of control points and the contractor doesn’t think this is the right time to upgrade machine control equipment. How much money did they leave on the table?
Amen. Can’t say how many times I’ve been on large-scale projects that were open-ended T&M, and we ended up renting better gear often enough that we could have purchased it outright.
“We can’t afford a $50K laser scanner!”
“We dropped 20-30K in rental fees over the past year, and this is a 3-year contract.”
“…”
“…people will come to love their oppression, to adore the technologies that undo their capacities to think.” -Neil Postman
Log in to reply.