Best way of getting an analog readout on linear/rotational movement

[ChristofferB]

Hi all!

I have a little side project in converting a diffraction grating UV-vis spectrometer.

The grating, as seen in picture 1 was rotatable with a motor, and I want to control it via a knob (or possible beter motor in the future) and have a proportional voltage out (to make quick, simple vis spectra with a X-Y recorder).

The mechanism for turning the grating is seen in pic 2. the axle rotates a screw that translates forwards and pushes on a brass cam, that in turn rotates the grating.

If I wanted a voltage out proportional to the rotation, what would be the best approach?

mounting a potentiometer directly on the grating axle? having the screw push on a slide potentiometer?

This is a quick project, so simpler is better.

thanks for the look!

 

[KT88]

That totally depends on your requirements...

[ChristofferB]

Of course it does, as said it should just be 'pretty good' - I'd prefer it to be potentiometer based, I'm just unsure how to mechanically implement it with the least backlash/ slop.

[KT88]

A coupler, similar to the one you used to couple the motor to the rod would be fine.
If you don't have the highest accuracy demand heat shrink tubing (maybe multiple) could also work, if the axes are precisely aligned.
There are potentiometers on the market for that purpose. https://www.megatron.de/

[ChristofferB]

I see, so there is. If I can find them used, might be a good solution!

This may be very 1950's, and a little on the Rube Goldberg end of the scale, but what if some dial string (very rigid) was tied to the cam, wrapped a few times around the axle of a 10 turn linear potentiometer, and then secured with a tension spring? There would need to be some friction on the potentiometer axle of course, maybe heat shrink tubing.



A potentiometer shaft
B thread
C tension spring

[Gribo]

Depending on the accuracy, I would use a shaft encoder.

[ChristofferB]

that's an option, but it would require some amount of circuitry to obtain a voltage out, like this:

[TimFox]

How many turns of the shaft do you need to encode?

[ChristofferB]

About 8 turns.

The string thing actually works remarkably well, until a better solution can be implemented!

[TimFox]

If less than 10 turns, why not a 10-turn pot?

[viperidae]

You could glue a magnet on the pivot of the brass bit and use a hall effect rotation sensor

https://au.element14.com/bi-technologies-tt-electronics/6127v1a60l-5/sensor-hall-0-2v-20v-to-10v-60deg/dp/2319671 is an example

[ace1903]

Can't see well but I think that you have DC motor with embedded optical encoder. I have motor that looks similar and it has optical encoder already mounted.
If you have same model, you can modify encoder cover to add another coupling behind encoder and drive motor shaft by hand by rotating shaft attached to that coupling.
I used STM32F103 to count pulses and to display position and small L298 h bridge to drive motor to desired position.

[S. Petrukhin]

Once upon a time, I had the task of controlling the rotation of the outdoor antenna from a computer. I put a mouse in the antenna drive, pressing the wheel to the motor axis, and connected the buttons to the limit switches.

[Ground_Loop]

I would try a ten-turn Hall effect potentiometer.

[David Hess]

Potentiometers intended for use in angular and linear position sensing are available and work very well.
 

[KT88]

@ ChristofferB: I like your implementation because of it's over all simplicity.
I would use most likelyuse a larger pulley on the pot and stay with a single-turn. This causes less drag and would give better lon-term repeatability (less or no slip - the string could be attached to the pulley). Together with a larger diameter pot you won't see little if any disadvantge in accuracy over a multi-turn at probably lower lower cost...

Cheers

Andreas