Hello,

I have just begun working with robotics and engineering and am a complete novice, so please forgive any ignorance or obvious points I should know as I am still learning.

I am working on making sure something a certain distance (25 ft), stops, and will come back to the starting point at the push of a button. Currently the machine is using vision sensors to attempt to stop the product at the correct distance, but the reaction time varies depending on the size and weight of the product. I figured the best way to improve this would be to attach a rotary encoder to one of the rolling wheels in order to better calibrate the distance by turns of the encoder wheel. Is there a way using the rotary encoder to control the motor to go only to the set distance , stop, and eventually return the exact same distance?

Any advice or tips are fully welcome, even if you are pointing out how dumb I am for missing something so obvious. Thank you!

 

This may help for a start...
https://circuitdigest.com/microcontroller-projects/arduino-based-encoder-motor-using-pid-controller

 

Optical-Sensors are most likely the best way to go.
Maybe your's are just set-up, or designed poorly.
.
.
.

 

If point-to-point motion control is not accurate enough, you may want the accurate motion control and positioning, as per CNC precision, then this can be done with either a open loop stepper motor , or true closed loop with a quadrature encoder and PID controlled servo motor.
There is a PC based HMI CNC control, Mach3, that could be used if necessary.
This is a simple step-dir based system, where the loop is closed in the drive.

 

What level of precision do you need?

 

During my last years before retirement we were doing quite a bit of precision measurement using equipment manufactured by Micro-Episolon. Really tight accurate precision laser systems. Deep pocket high end stuff but great stuff. On earlier distance and displacement measuring we actually used some yoyo or string pots which also were accurate and worked well. Draw wire displacement pots. Not sure if draw wire has a 75 foot runout. They have or did have a great selection engineers to help you.

For the most part we just ran an analog output proportional to distance into a code so we always knew where something was in travel and could have different things done at any given points. Newer systems now u8se I2C which is great. You decide what you want or need, talk to an app engineer and start figuring a bill of materials. Again, really good, accuarte repeatable stuff doesn't come cheap but it's all about what you want.

Ron

 

. On earlier distance and displacement measuring we actually used some yoyo or string pots which also were accurate and worked well. Draw wire displacement pots.
Ron

I worked on a System that used this principle, but had quadrature encoders as measuring devices, the device works on a principle similar to a steel measuring tape, with the encoder driven by the centre rotating pin.
It was used in a hydraulic lifting system for working on railway box cars.

 

If there is no slip in the system at any point then an incremental encoder measuring either wheel rotations or motor revolutions can simply drive an up/down counter, with a "home" detector at the start point to assure repeating. That way the system can know as the end is approached and slow for an exact distance move. The counter can be either an actual electronic counter or done in software, and the encoder resolution will depend on the accuracy required.
I designed a system like that using a linear motor with an integral linear encoder and it was repeatable over 4 inches to within 0.001 inch, which was the encoder resolution..

 

I designed a system like that using a linear motor with an integral linear encoder and it was repeatable over 4 inches to within 0.001 inch, which was the encoder resolution..

In my experience of quite a few decades working with industrial servo's, , with the constraints laid down by the OP, the system would require a PID loop for any kind of accuracy IMO.

 

In my experience of quite a few decades working with industrial servo's, , with the constraints laid down by the OP, the system would require a PID loop for any kind of accuracy IMO.

Certainly a PID controller would be a good way to control, but simply reducing the speed (gain??) as the endpoint is approached would be adequate. Gain proportional to error would work.

 

Certainly a PID controller would be a good way to control, but simply reducing the speed (gain??) as the endpoint is approached would be adequate. Gain proportional to error would work.

The OP mentioned the size and weight of the load varies, so the accuracy may also vary. Depends on what is required for final position accuracy.

 

Now I see that there is no mention about cycle time, and so a speed control with the speed being position controlled will bring about fairly accurate positioning every time.