Thinking of trying to set up OpengradeX to assist digging drain trenches with my 14 tonne excavator. I’ve hopefully sorted the code to resolve bucket tip position from receiver position, heading, boom, stick and bucket angles. Planning to pass modified gps data to a standard OpengradeX installation.
Just wondering if anyone else has tried this?
Very interested in your project!
I haven’t but I tried working out the angles in Powershell - was thinking 3 IMUs (one on each boom section) and a GPS on the cab could make it work?
I’m torn between IMU’s and angle sensors but IMU’s would have to be very good and stable as small angles make big differences on the end of a 3m dipper connected to a 5m boom.
IMU’s would be more durable. Either way, getting it actually working is my first step. Accuracy can be worked on.
As an update. I’ve (hopefully!) got the code to accept a GGA sentence then modify it to reflect the bucket blade edge. Just need to decide the best (low cost and robust) way to accurately measure angles on the machine. My options up to now are Landrover sensor as used in AOG for steer angle, an IMU on each limb or maybe a linear potentiometer. I guess a combination would maybe be best? The bucket joint has quite a large range of movement and is most vulnerable so maybe an IMU candidate?
Will your antenna be on the counter weight? There’s a system out there that mounts the antenna on the top of the boom/stick pivot. This eliminates some calculations.
The calculations are relatively easy but I see how this would make it simpler. Just a bit vulnerable maybe? An IMU would be needed on the stick I guess. The whole arrangement would then only depend on stick orientation (relative to north), stick angle and bucket angle. Eliminating errors from both stick pivot and boom pivot sensing.
For my usage, I’d be reasonably concerned about the vulnerability issues. BUT maybe it would be a good step forwards IF I was struggling with accuracy on a more conventional cab/weight mounted receiver setup?
Coding and implementation wise it would be easier for sure. Mounting the antenna directly on the end of the stick ram would also keep it’s orientation roughly correct too. With just a small but predictable variance in position when the stick is moved.
[edit…]
Need to think about this more. Actually quite like the concept. It removes track frame pitch measurement too as the stick pitch is measured independently. Roll is still required although in the real world wouldn’t be very significant as trench digging on any level of roll is not ideal. It would basically require two reasonably well set up IMU’s. For £150 each you can get good, accurate and robust ones, designed for this environment. Probably cheaper in the long run than messing about with linkages and angle sensors…
Got a few spare BN0085’s ( for testing purposes at least) so could actually cobble this together and see what happens. Altering the code is basically just removing stuff and using relative to ground for each angle. Seems to still work with test figures in theory at least.
The product is called Ditch Assist X. They use a pendulum mount for the antenna which keeps it directly lined up with the pivot point at all times. And then they have a module which I assume contains an IMU on both the stick and the dog bone.
I agree with the vulnerability part, but I think the antenna could be caged/protected enough that it would work in most situations.
I have had an antenna mounted up there with no protection whatsoever and only ever worried about it when working in the trees.
I’m going to disregard the pivot point and just use the ram mounting point I think. The F9P antenna is very small anyway and reasonably robust so it will probably be fine. The electronics can hide under the ram so will be protected.
Hopefully I’m not missing anything fundamental!! 
The code seems to compile and produce output that looks correct with test data. I’ve got adapters to send i2c over an ethernet cable for the bucket IMU so can use a BNO085 there for testing.
I would go directly with 2 GNSS receivers.
one on the stick and one on the frame.
At least to always have the correct orientation. It is turning a lot, I fear IMU alone could get lost sometimes.
And the one one the stick is a most IMO. With all the linkages from the frame I guess you would lose some vertical precision quickly
I had wondered about this, and do have two F9P’s lying around at the moment. I’ll start with the stick one and see how it all behaves. It will be a really interesting project anyway.
I do far more excavator work now than tractor stuff, so don’t get as much time on AOG as I did. Grade assistance on my 14 tonner will save me a LOT of time too. We drain with solid, twin wall pipe almost exclusively so it is actually quite forgiving when it comes to trench smoothness. Far more than land flex.
This project is worthy of some expense to me if it works. My laser equipment is good, but time consuming in operation and requires a reset on every grade change. I’ve got at least £3500 invested in that so a bit thrown at this project is definitely worthwhile.
These look a good compromise between price and accuracy. If they are stable enough. Interesting to compare one to the sort of stuff we use already.
Hi, I installed a setup on a 13 ton excavator last fall as a guidance screen and it works very well as a reference when digging trenches. I just installed a GPS antenna on the stick and I made myself a mechanical angle indicator to always position the stick at the right angle to validate my trench height, easy to use as a reference when the slope is minimal.
