Position-Force loop 0-120000N

Having a project for a testbench that will control a cylinder in a position-force control loop.

My customer want to copy the setup from a similar testbench and make a new with the same capacity. Their current testbench is operated manually where they can (with a light hand and fine adjustment) get the machine to pull down to 500N.
My question is, would it be possible to automate this, with an RMC ofcourse, and still atleast be able to pull down to 500N (they would like go less if possible) as well as pulling up to it’s max 120000N capacity.

I’m no expert in hydraulics, but I think that when someone needs to pull 500N they doesn’t choose a 120000N cylinder.

There isn’t any data available for any of the hydraulic components yet as this is still in a very early state.


You are correct that it can be difficult to control to a value that is a very small percentage of the maximum capacity. The two main factors are the feedback resolution and the gain of the actuator. When I say the gain of the actuator, I mean how much the actuator responds to a small change in the command signal.

With high-resolution feedback and a high-response valve, I would expect that it is possible to control to 500 N, even on a 120000 N cylinder. The main question is how much tolerance is the system allowed to have when it is at 500 N. Some items that will improve the controllability are:

  1. Use a dual-range load cell to measure the force. This type of load cell is two load cells in one, with one load cell for the entire range, and the other for about 1/5th of the range. This provides higher resolution for the lower range. The lower range load cell also has overload protection so it can handle the high load, but only measure in the lower range. The RMC motion controller has a feature called ‘custom feedback’ that will allow to the RMC to use both inputs from the load cells, and seamlessly switch between them as the load increases and decreases. Here is an example of a dual-range load cell: https://www.interfaceforce.com/products/load-cells/2101-dual-range-standard-compression-load-cell/

  2. Use a small, high-performance valve. With a smaller valve, finer control is possible. For precise force control, the valve should be very high-response with a zero-lapped or under-lapped spool.

  3. Use two valves if necessary. If one small valve is not enough for the dynamic range of the system, it is possible to use a large valve and a small valve, and switch between them when necessary. I am guessing that for your system, one small valve is sufficient, but I don’t have enough information to know for sure.

0 N should be achievable. Why should this be difficult?
The problem is feedback resolution but think about this.
When the actuator is stopped the sum of forces acting on the piston is ZERO.
However, the sum of forces on both sides of the piston just need to be within the static friction to be stopped.
When stopped, the pressure x area on both sides is about the same except for seal friction.
If the seal friction is 500N then it will be hard to get the pulling force to exactly what you want below 500N unless the cylinder has little or no friction.
Also, if a load cell is used, instead of pressure sensors, the accuracy will increase because the friction does not affect the measurement.
However, now we get into the accuracy of the load cell 550/1200000 is 0.4% full scale.
The load cell better have much better resolution than that