homemade autosteering egnos

stef

Member
Location
belgium
Jd 30 series don't use that valve
Mr PS
The controller / software you've written, is that translating PVED-CL commands?
I don't use any built-in valve from JD, instead I use an "external" valve which I installed myself.
This external valve is piloted by the controller, and this controller exchanges messages (I also built these protocol layer) with the main software. running on windows.
 

sleepy

Member
Location
Devon, UK
I don't use any built-in valve from JD, instead I use an "external" valve which I installed myself.
This external valve is piloted by the controller, and this controller exchanges messages (I also built these protocol layer) with the main software. running on windows.

Have you considered open sourcing your design, to allow other people to help develop it?

GPS systems are still sold into agricultural markets for far more than they are worth, simply because the manufacturers can get away with it. The Chinese are hopefully about to start pushing the price down.
 
I built/implementded something similar to PVED-CL commands, I have set of commands to manage the SteeringControler and another set to manage the SprayerSectionsController.
I don't use any built-in valve from JD, instead I use an "external" valve which I installed myself.
This external valve is piloted by the controller, and this controller exchanges messages (I also built these protocol layer) with the main software. running on windows.
Wow! That is really seriously impressive! Can you show us a picture of the external valve you have used?

What sort of controller have you used for section control?

So you've written all the control software, the protocol layer and command set for 2 controllers all on your own - presumably in your spare time? This gets better and better!
 

stef

Member
Location
belgium
Wow! That is really seriously impressive! Can you show us a picture of the external valve you have used?

What sort of controller have you used for section control?

So you've written all the control software, the protocol layer and command set for 2 controllers all on your own - presumably in your spare time? This gets better and better!

here is a picture of the valve.

I also built the sprayer 's controller (SSR switch); the sprayer's valves (Berthoud sprayer) are commanded by 12V (no-canbus on this model);

Yes, it took me a lot of hours to achieve this...
 

Attachments

  • 20151125_071506.jpg
    20151125_071506.jpg
    143.6 KB · Views: 545
This seems way ahead of the main GPS manufacturer's quite crude retro fit electric steering wheels. Why don't they use a valve like you have? It seems to be a much tidier solution.
 

Robt

Member
Location
Suffolk
A valve can be fitted but it's then hard to change between vehicles and it's normally more expensive.
We are fitting one to a Bateman next week. Do about one a month
 

stef

Member
Location
belgium
Both systems have pro & con.

I choosed the electrovalves for these reasons:
  • no noise, no moving parts in cab
  • no additional wires/connectors from steering wheel to controller
  • no additional space obstruction in cab
  • no manual "physical" handling to engage/disengage the system (electricSteeringWheel needs physical contact with steeringwheel)
  • higher precision control (sliding issue between electricSteeringWheel and steeringwheel)
  • don't want to move to much device when switching from vehicle to another, in my case I only move the windows tablet (which is used also for other purpose than autosteering).Electrovalve is installed once than forget it.
  • electrovalve can be found as EOM parts, not yet the case for electricSteeringWheel
  • global price depends on the number of vehicles you need to equip, threshold is around 2-3 vehicles?
 
Location
North
Would be nice to know more about the different valve systems. I assume OEM steering valves have some sensor that tells the driver is using the steering wheel, that sensor output should be easy to use on a self made steering controller too? Is there any significant price difference between an OEM valve and this self built one?

Another question: does this system make use of the steering angle sensor? I understand most steering wheel based systems do not, which again is a significant disadvantage specifically at low speeds or when the tractor does not always follow the front wheels (ploughing on a furrow or front press etc.).

I thought RTK would not need any other changes in addition to a new GNSS receiver with RTK capability? Would the system need to know about the higher accuracy?

Some of your SW features seem much more advanced than those on many OEM screens, very impressive.
 

stef

Member
Location
belgium
Would be nice to know more about the different valve systems. I assume OEM steering valves have some sensor that tells the driver is using the steering wheel, that sensor output should be easy to use on a self made steering controller too? Is there any significant price difference between an OEM valve and this self built one?

Another question: does this system make use of the steering angle sensor? I understand most steering wheel based systems do not, which again is a significant disadvantage specifically at low speeds or when the tractor does not always follow the front wheels (ploughing on a furrow or front press etc.).

I thought RTK would not need any other changes in addition to a new GNSS receiver with RTK capability? Would the system need to know about the higher accuracy?

Some of your SW features seem much more advanced than those on many OEM screens, very impressive.


I use proportional hydraulic valves from hydraforce, no feedback provided (a picture has already been posted some days ago, see above)

Yes, I use a steering angle sensor, indeed this device helps the steering model to calculate/reach the right path.

In "theory", a RTK solution only needs a Receiver with rtk capability, but I'm not sure that the actual steering model will fit at 2cm (vehicle dynamic movement's model is quite important while trying to reach this precision)
 
Last edited:

stef

Member
Location
belgium
most of the time around 5-10 cm which is +/- enough to seed cereals.
Main problem is the drift which accumulates over time; this is not really a problem as long as you follow always the next nearest track; drift effects can be noticed when you "jump" to a more distant track in order to avoid headland maneuvers. The picture was taken in a field of 10ha, seeded in around 7hours and I "only" experiment one drift of 15cm, I started to seed at one side of the field then I continue from the other side, the join boundary is on the picture "drift".
drift is not predictable, sometimes only a few cm, sometimes 1meter... this is one of the egnos system limitation/constraint

  • picture of the equipment used to seed.
  • picture taken between to track.
  • picture which shows the drift after some hours.
 

Attachments

  • between2tracks.JPG
    between2tracks.JPG
    610.4 KB · Views: 370
  • drift.JPG
    drift.JPG
    520.1 KB · Views: 357
  • 20141017_122812.jpg
    20141017_122812.jpg
    383.4 KB · Views: 346
Location
North
I guess you should have added that the drift is from the EGNOS receiver, not at all from your self-made equipment or SW. Accuracy I cannot comment, sounds like path-to-path accuracy is pretty much all from the EGNOS receiver too, the steering system having no significant contribution. That is, a commercial system based on EGNOS would not work any better (neither any worse).
 

holwellcourtfarm

Member
Livestock Farmer
VERY interesting thread. Huge congratulations on what you have achieved. Roughly how much do you think it has cost you to get to this stage? Commercial kit to do this to the same accuracy seems to be around £7k - £9k.

Does your software allow you to download/save a record of what you have done?
 

stef

Member
Location
belgium
I guess you should have added that the drift is from the EGNOS receiver, not at all from your self-made equipment or SW. Accuracy I cannot comment, sounds like path-to-path accuracy is pretty much all from the EGNOS receiver too, the steering system having no significant contribution. That is, a commercial system based on EGNOS would not work any better (neither any worse).

you're totally right, accuracy & drift is due to egnos receiver/correction.
Then, you have to fine tune the parameters of the sw in order to stick as much as possible to the theoretical track path calculated.
 

stef

Member
Location
belgium
VERY interesting thread. Huge congratulations on what you have achieved. Roughly how much do you think it has cost you to get to this stage? Commercial kit to do this to the same accuracy seems to be around £7k - £9k.

Does your software allow you to download/save a record of what you have done?

as you can imagine, major investment is time spent to develop, try, test, fix all the stuffs...

actually, you can save each session (a work) and continue it later. Unfortunately, the files are binaries and not compatible with any systems (this could be a future functionality (export/import data)). I did spend time on this topic because there was so much work to do... however binaries files are very useful for me, in order to test/replay (simulate) a set of data at my office desk.
 

stef

Member
Location
belgium
Are there gyros in your system to account for slopes?
I had a version running with an imu, but results are not yet sufficient, too much emu drift over time; I think I should invest on a high grade emu instead of the one I use. however I already noticed that tilt/roll has a large impact on position accuracy (antenna is on the cab roof)
 

SFI - What % were you taking out of production?

  • 0 %

    Votes: 105 40.5%
  • Up to 25%

    Votes: 94 36.3%
  • 25-50%

    Votes: 39 15.1%
  • 50-75%

    Votes: 5 1.9%
  • 75-100%

    Votes: 3 1.2%
  • 100% I’ve had enough of farming!

    Votes: 13 5.0%

May Event: The most profitable farm diversification strategy 2024 - Mobile Data Centres

  • 1,775
  • 32
With just a internet connection and a plug socket you too can join over 70 farms currently earning up to £1.27 ppkw ~ 201% ROI

Register Here: https://www.eventbrite.com/e/the-mo...2024-mobile-data-centres-tickets-871045770347

Tuesday, May 21 · 10am - 2pm GMT+1

Location: Village Hotel Bury, Rochdale Road, Bury, BL9 7BQ

The Farming Forum has teamed up with the award winning hardware manufacturer Easy Compute to bring you an educational talk about how AI and blockchain technology is helping farmers to diversify their land.

Over the past 7 years, Easy Compute have been working with farmers, agricultural businesses, and renewable energy farms all across the UK to help turn leftover space into mini data centres. With...
Top