We are interested in installing potentiometers onto the ends of shafts to measure rotation. I am curious what the best way to attach a pot to the end of the shaft is from your experience.
Post was about a specific encoder, but the concept can work with almost anything.
1 Like
If the shaft stub is already turned down to approx 1/4" (or whatever the diameter of your pot stem is) like many AndyMark gearbox back shafts, rubber tubing or shaft couplers are the go-to answers. If you use a rigid shaft coupler, be sure to design your pot mounting bracket so that it constrains rotation of the pot body but doesn’t over-constrain the alignment.
1 Like
I’m going to real quick recommend using a real encoder - they’re more expensive, but something like the CTRE Mag Encoder or the MA3 encoder are going to be much more accurate and reliable.
1 Like
We have had some success using a short section of 1/4" ID rubber fuel line as a shaft coupler. It’s stiff enough to keep the potentiometer rotating but compliant enough to keep most of the side loads off of the pot bushing.
1 Like
I have to disagree with this. If you’re measuring an arm that goes through 1/2 rotation, you aren’t getting any benefit from an encoder.
In fact, you probably reduce accuracy. With the CTRE Mag Encoder, you get 1024 CPR. That means over the course of your 1/2 rotation, you have 512 discrete points. Since a potentiometer is an analog device, you can get a lot more points! An accuracy in your reading of 0.001 is normal. Over a 5V range (assuming a properly sized pot), that’s 5000 different points.
As for reliability… With an encoder, you have to know where you start. You have to be able to track your end points and reset your readings to ensure things don’t get wacky and cause you to break your arm when it isn’t where you thought it was. With a properly set up potentiometer, you know exactly where the arm is as soon as you turn on the robot, and the position is always known, just by reading the current value of the pot. No need to track counts or reset your readings.
3 Likes
You’ll actually have 2056 discrete points if you seed the relative position with the absolute position when the robot starts.
The ctre mag encoder has both absolute and relative modes.
1 Like
What potentiometer would you recommend to be used with a TalonSRX?
1 Like
Sorry for not a specific part, but there are so many. Both Mouser and Digikey have potentiometers. 5K is a good value. You want to be sure the get the linear type. (They make logarithmic one for the audio industry.) With DigiKey you can use filters to drill down to what you want. Continuous rotation is a good feature so you don’t have to worry about breaking it if your arm over rotates . In stock is also good. You have to watch not to put the cross over from high to low in the used range of rotation. CTRE also makes a break out board for analog sensor. IE potentiometer. They make wiring it simple. Vex also has some potentionmeters on their regular web site. I never actually used one though.
The CTRE mag encoder is great. But in a lot of situations a pot is easier to mount.
A little late now but on first choice, John Deere had some really nice rotary analog encoders. They go on green stuff that almost as much abuse as a FRC robot.
That’s a great idea. Is the shaft slit all the way through or just that portion?
So one could use an absolute encoder to accomplish the same goal as the position is static and not relative to the state the robot starts in? That is good to know, we have used the ctre mag encoder in quadrature mode in the past for our drive base. As for the debate between absolute encoder vs potentiometer is it mostly preference? Which would you say is easier to implement mounting wise? Is an absolute encoder less likely to fail than a potentiometer?
If your shaft is connected via belt or chain to something else on your robot, you might want to run the potentiometer off that belt or chain instead of just off the shaft. For example, McMaster has a #25 sprocket that will connect via set screws to 1/4 inch shaft like what is on a potentiometer. You can also get small timing pullies that will do the same thing. it’s a helpful approach if the end of your shaft is not in a convenient location.
We are running chain so this is an option we will likely discuss and weigh out.
I think we have backed ourselves into a corner of sorts unfortunately. We are using Neos to run our arm because we wanted something lighter to work with. Measuring the motor rotation at the driving sprocket rather than at the shaft doesn’t allow us to account for slippage. I don’t know if there is any way to interface an analog device with them at this point in their development.
Here’s how we solve that problem: https://photos.app.goo.gl/HaVbJExfT57ZYBVB8
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.