Valtra 8550 hydraulic autosteer project

Though I’d share my story here as well now that it’s in the working stage finally. The victim is a Valtra 8550 from 2000 with just over 16 000 hours on the clock, so bit of a new life for the old lady :blush:


Started off with the Kaupoimod PCB assembly some months back:

IMG_4502 IMG_4539

For the autosteer valve, I purchased a Danfoss PVG16 unit with PVHC propo block, so it’s just two coils that are driven with PWM. End plate is configured for closed center and includes the LS shuttle valves and so the orbitrol LS line goes through the external LS. Also bought an adjustable pressure switch to the LS line from the orbitrol. Fabricated a bracket under the stairs so it sits next to the priority valve.

Piping is so that steering L/R, P, and T lines are all tied directly to orbitrol. This way the steering is protected with the orbitrol shock valves even in autosteer mode. LS line from orbitrol to steering valve LS external input, LS line from steering valve to priority valve. Currently, no mechanical cut-off anywhere, will put a ball valve to the steering valve P port.

Valve unit:
IMG_0141 IMG_0143

Bracket under the stairs. Raw material is recycled from a Valmet 903 bottom protection plate that my dad bought new in 1981 so bit of heritage in the system :wink:
IMG_4554 IMG_4557

Piping in the tractor:
5C593CC3-3213-4AC0-9151-4D83AD423BD6 IMG_0144

Finally, put in the steering angle sensor with a bracket on the front axle:
IMG_0139 IMG_0140

And fitted the box and the tablet in the cab. The green switch above the PCB box is a main power switch for the whole system, and the red button visible under the tablet is the autosteer switch.
IMG_0145 IMG_0146

So far so good, have only been running some small tests in the back yard for now trying to figure out the steering parameters. Driving out from the shop had my first moments of desperation as there was no pressure in the steering, after two hours of banging head to the wall I figured the relief valve in the PVG was in the factory setting so only 20 bars in the system, easy fix with a 4 mm allen wrench in the end :smile:


How will you install a pressure switch on the LS line so that using a Danfos valve will not disconnect the Autoster? Check valve?

There’s a built-in shuttle valve in the PVG LS line, so no need for a separate shuttle valve. When LS pressure rises in the propo valve, the shuttle prevents it from going in the wrong direction. Seems to work fast enough so that the initial pressure spike doesn’t reach the switch.

Finally up and running! Seems to work well now, tried even putting a check valve to the LS line (tried 1/4 and 3/8) but that messes up the original steering, feels sticky as the LS pressure is throttled somewhat by the check valve. Ended up increasing a bit the pressure setting in the switch and works fine.

Had to do a bit more on the software side, though. Using 10 bit PWM helped a lot in smoothing out the response as the spool travels from min to max in quite a narrow range. Still it was either lagging behind or way too aggressive, along with heating up the oil and sending pressure spikes when steering changed direction, no matter what the settings.

All problems were finally solved by setting a 0.4 degree deadband in the .ino, just sets the pwm to zero when in the deadband. Now the spool has time to return to neutral before steering the other way and all the pressure pulsations are gone. Same with the oil heating up, as the spool was unable to vent the LS line to tank when the ’residual’ PWM kept it off the zero mark.

Also the 45 degree tilt on the housing sort of works, subtracted sin(45 deg)*16384 from the roll value. Not exact, but good enough, played around with the antenna height to get it roughly right. Need to do the math and check if I can get it closer in the code.


Great to see an “old” Valtra getting along with modern technology ! Here (belgium) one fendt dealer once said to me (trying to sell his machines):

Valtra is not bad but archaic !!! :grin:

so we still have that quote in mind with my neighboor when we talk about those “50s” serie (I have a 8150). I can see you had a paintjob too and some trouble with those hood panels… thoses small isues that every owner has too… I would like to ask: how did you fix your Ram mount for the toughbook in the cab ? I will post some pictures of my setup later.
I also have a T163 Versu (not the same hydraulic pump (LS) even if the schematic is the same for the brake/steer part …but the hydraforce valve seems to leave some strange LS signal pressure that i don’t really like (pump pressure noise and little strange feeling on the steering wheel). So when you say

the spool was unable to vent the LS line to tank when the ’residual’ PWM kept it off the zero mark
I don’t really understand that, could you explain a bit more ?


It’s original paint! The side panel my dad ran over, managed to source a new one, just needs a coat of paint. We bought the thing new back in 2000 so it’s sort of an emotional bond having spent a large part of my youth in that cabin :grin: Bought the new T4 series when it came out in 2015, some progress had happened in 15 years…

As for the hydraulics, this is my theory: The Danfoss valve I have vents the LS line to the tank when in neutral position. When in autosteer without the deadband, the valve is constantly off-center. So it starts giving hydraulic pressure to steering when it’s at around 20% PWM duty cycle. The PWM in the standard setup only goes to zero when the error is 0.0, otherwise it’s giving something like 10-20% duty. This doesn’t move the spool enough to make the wheels turn, but the spool is not in neutral, so the LS line cannot vent to tank as it should when there’s no steering command. As a result, even though there’s no steering movement, the valve is commanding for oil flow. Now, with the deadband, the PWM goes directly to zero and the spring in the valve pushes the spool immediately to center and the LS pressure drops instantly.

As for your issues with the T163, are you using which type of hydraforce valve? If it’s not venting the LS line to tank, you will have constant pressure in the LS line, which it shouldn’t do. Do you have a check valve in the LS line from the orbitrol to the valve? I had exactly the same symptoms in the orbitrol steering when I added the check valve to the LS line from the orbitrol. It seems to restrict the flow too much and make sthe steering “jerky”, my theory (once again :blush: ) for that is that the check valve doesn’t move smooth enough and cause the LS pressure to fluctuate.

I see that the way of controlling the LS with a valve slider is not a good solution in an application where extremely small valve openings are needed. The LS control with L R check valves is accurate and does not cause such problems. This is probably why difficulties in regulating these valves can arise if there are time offsets between L / R and LS opening. Atos valves that I have already installed on many machines always work smoothly and precisely and the adjustment is very easy!
Next week my friend will wear my kit on Valtra so I will report how it works.

Discard the check valve from the orbitrol line or use a throttle check valve.

Then you can easily set the flow value to discharge the LS line and at the same time reduce the losses from the backflow of control pressure through the orbitrol.

Yeah I took it out after trying out. The ball shuttle valve in the PVG valve body is working very fast so no problems now!

The big thing was adding the deadband so that the valve works as it’s intended regarding the LS line vent to tank. I guess this is very much dependent on the valve design, spring force vs coil force.

@baraki next install I’ll try the Atos setup you have. Seems like a good combo. For this one, I’ll still add the safety on/off valves to A/B lines instead of a 6/2 way. Easier to install (at least in this case) and then you have the protection for steering from the orbitrol shock valves when in autosteer mode.

Bit of an update to the system, changed the valve driver to PVEA-CI and driving that from Kaupoimodv4 pcb with a mcp2515. Bno085 IMU with the tare fuction used so that the PCB can sit in an angle.

Needed to crank the steering pressure down to make the valve behave. Now it’s a compromise so that manual steering gets enough pressure and autosteer isn’t too aggressive. Should replace the PVB valve body with a version that has a separate control for the LS pressure or put a on/off valve to the pressure line. Now both steering pressures are limited by the PVG relief valve.

Next step is upgrading everything to CAN, already have the WAS and pressure sensor (STW M-01) or then try out wasless…

image image image


Finally got to real work and did a bit of fertilizer spreading with the setup, seems to be working fairly well, maybe some adjustment to the steering parameters for Stanley/PP but the steering is following the target value spot on. The pressure compensated valve works very well regardless of soil moisture, duals etc.

Have a new PVB valve body on order that has an individual LS pressure adjustment, so should get rid of the final steering pressure problems.

Hi Nut,

I’m with an hydraforce SP10-57C, i have the symptoms of delays and overshoot (which comes together I believe). I do have a throttle check valve and it does work ok when i do a test on static on soft ground or front axle up!
anybody heard of this hydraforce valve: (SP10-57CF-0-N-00). The F stands for fast (or min. lap) ?

What is the benefit of the danfoss or Atos (baraki’s way of doind it ?) The current necessary for it ? I do like the hydraforce for the small required space necessary.

You mean you have a check valve in the LS line? What kind of delay, it’s not following the target angle? Probably too low min PWM

As for Danfoss, it’s all about the closed loop control, there’s a displacement sensor on the spool so you always get the same spool movement with the same command.

yes, I have a throttle check valve between the orbitrol and the Hydraforce prop valve on the LS line.
Well, if i “close” the throttle check valve i do have a reactive steering but i can hear the hydraulic under load (which means warming of the oil) if i do loose it up, well, the steering is less reactive due to the amount of the LS signal needed to change the priority check valve sending pressure to the steering/brake flow circuit.

As for Danfoss, it’s all about the closed loop control, there’s a displacement sensor on the spool so you always get the same spool movement with the same command.

So you can have a better “standart” start /max PWM ?

Very nice :+1:
It seems to be a lot of parts in this Danfoss valve
Do you have the parts number for your assembly?
Try to figure it out on today but i’m not sure if I get the right parts

The throttle check method works very well, but not with every priority valve. Oil flows from the LS line priority valve towards the orbitrol, in some valves the nozzle is very small and the leakage is very small and with these the control works very well. If the nozzle is large and the leakage towards the orbitrol is large, steering this way is problematic. I am a bit surprised that you have a problem with Valtra, I made a few pieces without any problem. I deal with excessive leakage mainly at JD. Or is the priority valve worn out? Do the test, unscrew ls from the privet valve, direct the oil into a bucket and see how large the leak is.

I’m installing PVFC valve to replace the PVG internal shuttle valve. The only problem left is the pressure shocks to the LS line when spool goes to neutral. I think is due to the leakage flow being completely blocked when the PVG is working.

This valve one in the converted config on page 4. That’s how it’s done on OSPE / EHi valves so should work, we’ll see. Port 1 goes to orbitrol instead of tank and there should ve a constant leakage flow

1 Like

I think adding a little leakage to the Ls on the Ls line from PVG would solve your problem.

Might be, but I would need to put an external shuttle valve then in any case, as I’m now using the internal shuttle and it’s impossible to put a bleed into the LS in the PVG valve body. So I’d need a shuttle and then a bleed orifice with a tank connection.

With the PVFC you have a constant flow to the orbitrol (which it should have) in all conditions and I picked on up for around 60 eur at a spare parts retailer so not too bad priced either.

Deleted as you already have the pdf I just saw