Control pololu generic linear Actuator (FB) -vnh5019

I am attempting to control two 4" stroke Pololu Generic Linear Actuators with a Dual VNH5019 Motor Control Shield using an Arduino Uno microprocessor. I have started prototyping the components before I place them into my project. I cannot find any information on the feedback from the actuators except that they are supposed to be a 10 ohm potentiometer. The testing performed has been based on running the VNH5019 Demo software that exercises the actuator motors very well. I then modified the code to add the feedback connecting the white wire to gnd, the yellow wire to Arduino 5V, and the Blue feedback wire to A5. The FB values are read on every cycle of the FOR loops so the total movement can be monitored. The results are consistent for both Actuators and on both the Arduino Uno and the FB were also read on a separate Mega.

The results did not map to the analog (0 to 1024) that would be expected from a 10 ohm potentiometer but were much more narrow. In a 2.5" range, starting at 1.1" and traveling to 3.6" the readings were 464 to 624 which is a count of 64 per inch. This extrapolates over 4 inches to a start value of 400 and end value of 656. There seems to be periodic single low or high readings not in the linear progression. The actuators were tested individually to help rule out a bad potentiometer and the Arduino setup was tested with a 10 ohm potentiometer giving the expected results.

Please help clarify what is the expected output of the Feedback loop and any information to help control the actuators with VNH5019 motor controller.

Thanks,
JSallas

Hello, JSallas.

The output you are describing is reasonable. There is a potentiometer mechanically linked to the actuator, so the potentiometer output voltage corresponds linearly with the actuator position. Only a portion of the pot is used, however (it will not necessarily be at its extremes when the actuator is, so the output is not the full 0 to 5 V).

Can you clarify what you mean by this? Are you saying you occasionally get outlier readings? It sounds like it might just be noise; are these outliers repeatable? It’s good practice to average a few of your analog readings together to filter out noise. Are you doing this?

- Ben

Ben,

Thanks for the input. I am not currently averaging readings but will try it also will perform a reasonableness test and throw out any unreasonable out of range readings.

This is an interesting project to automate a tennis ball machine with an Arduino. It will be able to control ball spin, speed, and time for each shot with servos. The elevation and direction will be controlled by the linear actuators. The communications and remote control will use WIFI. I plan to develop an IPAD app to be able to build shot sequences to create realistic situational practice scenarios.

Thanks for your help! I’m sure there will be more crisis.
JSallas

That sounds very interesting. I’d love to hear more about your project as it progresses, and don’t hesitate to post here if you run into any trouble!

- Ben

I have the same issue with the generic linear actuators.
I noticed the problem is caused by the potentiometer chosen. The potentiometer used is omter 3590S with total mechanical travel of 3600 degrees. However the gearing inside only rotates the pot for 360 degrees in the 4inch version.

Anyone know if the concentric brand version have the same problem?

Where did you find the specification that the potentiometer is 10 ohms? I’ve been considering use of the linear actuator for a vent-opening device, and according to the data sheet, for the 4 inch actuator, the pot output is given as 2.35K ohms/inch.

It would be great if you could connect an ohmmeter between the white (ground) and blue (RS) wires, and report the ohms measured at each end of the actuator’s range.

Thanks, Jim

Hi, Jim and Fisharman.

I just tested a 4" generic linear actuator and it looks like Fisharman is correct. The resistance between the wiper and one side of the pot was approximately 4.5k with the actuator at extreme and 5.5k at the other (it is a 10k pot). The Concentric linear actuators are better in this regard: you get almost the full range of the feedback pot for all of the Concentric lengths.

- Ben

Thanks Ben,

According to your measurements and as I suspected, JSallas’ assumption of a 10 ohm pot is not correct: the variation is about 1000 ohms for the extremes of motion. It is unfortunate that such a narrow range of the pot is used.

Jim

Sorry, when I said he was correct, I just meant his conclusions were consistent with one turn of a ten-turn pot. It definitely isn’t a 10 ohm pot.

I agree that the narrow range is unfortunate, and we will check with the manufacturer to see if this is something they can easily improve.

- Ben