Insane Hydraulics

Theme Image

A Different Way to Steer with a Danfoss PVG32 Valve

Today I want to talk about a slightly different way to control proportional electrohydraulic DCV actuators with ratiometric input. I will describe a small "urgent" project in which I used this special control, and although it used the Danfoss PVG32 series valve with PVEH actuators, I am sure that the principle that I am about to describe can be used in any other proportional DCV with ratiometric control.

What is a ratiometric input anyway?

A proportional DCV with a ratiometric input considers it's value (voltage) in relation to the supply voltage (as opposed to a simple analog input that measures an absolute value), which makes it immune to supply voltage variations when the source of the control signal is some kind of a voltage divider (i.e. a potentiometer).

This is very convenient for mobile equipment (where supply voltage varies all the time) and allows for the creation of simple and reliable potentiometer-based bidirectional proportional controllers (joysticks). For the PVEH the input value of 50% of the supply voltage means "spool centered", and 25% and 75% of the input shift the spool fully to A and B respectively.

The Story

Allow me to explain how a need for this control appeared in the first place.

It all started when a successful tasty greenery (salad, lettuce, watercress, etc.. ) producing company decided to build a small harvester of their own design.

If you've never seen a watercress harvester and are imagining a CLAAS Lexion in your head - dial down your scale. Wa-a-ay down! Greeneries like the watercress are grown in narrow water-filled beds, and the harvesters have a tiny footprint and are often more tall than wide (at least the ones that carry an operator on top of them). They almost look like a tall chair placed on top of a pair of short tracks.

Several companies helped to put the project together, and in about a year, the first shiny prototype was being tested in the field. The hydraulic system was based on a Danfoss series 45 open-loop pump with a classic LS controller coupled to a Danfoss PVG32 with PVEH and PVEO actuators, which were controlled with two Prof 1 series joysticks.

The tests, however, revealed an interesting problem - operators could not control the tracks with needed precision. And the origin of the problem didn't lie in the control system itself, but rather in the way the combination of the machine-operator behaved. This phenomenon could be called parasitic feedback, but in the case of that particular machine I called it:

One, Two... Dead!

Let me explain what I mean by this.

The harvester was very tall in relation to the track length, and the operators' chair was on the very top, so any uneven terrain or any acceleration and deceleration would cause violent swings of the operator's chair, and since the operator was holding the joystick handle (I would, actually, replace the word "holding" with the word "clenching for balance") - any precise control was impossible.

Here's a typical learning curve of a new operator. Putting the machine in motion required pushing the joystick forward, so when the tracks would begin moving the tall frame would swing back, causing the operator to sway and push on the joystick all the way out, which, naturally, would result in a bigger acceleration, making things even more comical. After realizing that the machine is running away, the newbie would let go of the control, which would immediately halt the tracks, with the harvester in the "loaded" (dangerously tilted backward) position, after which the "devil's apprentice" (as the drivers called it) would straighten itself in a no less violent fashion, catapulting the operator precisely into the front blade, which was essentially a fast-moving and extremely sharp band saw - hence - one, two ... dead.

After a couple of close calls, an operator would eventually manage to drive the machine around the yard, but still, any harvesting was out of the question, because the machine would have to move at a steady slow speed, and would only need to make small course corrections along the way.

A different control solution was desperately needed, and I was asked to find a "quick and dirty" fix. So, I went to the farm, saw the machine in operation, even tried it - and indeed saw that the required control precision was completely impossible with the current joysticks.

After giving it some thought - I pitched an idea. What if we created a system that would allow the operator to choose a desired speed on a dial control while allowing him to gently steer the harvester with another dial ( since the harvester is on tracks - the steering would be performed by track speed control). I also proposed that it should be controlled by an on/off pedal switch, so that when the operator lifts his food the machine stops. Since the dials would not be touched constantly, but only occasionally for small speed/course corrections - all the control problems should go away.

The client liked the idea and gave me a "go".


Using a PLC would be nice, of course (who doesn't like a PLC!) but was out of the question - way too expensive! What I needed was a simple and most analog solution in the world. I have controlled PVG32s with simple industrial potentiometers before with great success - however, this time it was slightly different - this time I needed to control two modules at the same time, and I would have to find a way not only to control the magnitude of the signal for both of the valve sections (speed control), but I would also need to control the division of the signal between the two outputs (steering control).

My idea was - since the input of the PVEH control module has some internal impedance, which is high, but not infinite (in our case about 12 kOhm), I could use it to my advantage and use another potentiometer to divide the signal between my two 12 kOhm signal consumers - giving me exactly the steering effect that I wanted. Here's a schematic that shows the principle and the values that I used:

This schematic illustrates just the principle of the signal division. In my bord I also used a couple of relays to switch the control between the existing joysticks and the dials - so the operator would use the joystick to drive the machine around, and would only use the dials for harvesting.

Then, since I was extremely short on time - I placed all of the components on a prototype board (once again - a DIY technology in an industrial project!), fond a nice black enclosure, and installed it in the electric board of the machine. Thank god they used a standard industrial cabinet, which made the installation easy. Isn't it a beauty?

Here you can see the dials:

The control board exceeded all my expectations! The speed, direction, and steering control was absolutely perfect and very steady. The harvester itself did have other problems, and a ton of them (like any prototype would) but the principle that I was testing - i.e. the way to control two ratiometric inputs with two simple potentiometers connected in series - one for the signal magnitude. and another for signal division - worked perfectly!

And the solution was pretty cheap two. I used Eaton M22 potentiometers, which are IP66 and heavy-duty, and I got them for about 30 bucks each. Try getting a prof 1 joystick for that money!

So - you definitely can control signal division between two ratiometric inputs with a simple potentiometer. Bulletproof and cheap!