Judging from the pressure of the repressed bile...

The one thing that is abundantly clear is the collective frustration that people feel about providing help while critical information is withheld for "proprietary" reasons.

It becomes a silly game that is not fun or productive.

I am not calling out the TS here, just making a general statement.

 

We can surmise the following:

• the object is either cylindrical or spherical.
• the object is likely cylindrical

How? They could as easily be square or rectangular. If I was like I was back in my working days making things in a factory, seeing that much variance in the parts being made, instead of detecting the parts that were different, the normal thing to do would be to fix the reason they were varying.

 

The one thing that is abundantly clear is the collective frustration that people feel about providing help while critical information is withheld for "proprietary" reasons.

It becomes a silly game that is not fun or productive.

Or it could also be that the TS is trying to get us to do his "homework assignment", by posting it as a work related problem.

 

How? They could as easily be square or rectangular. If I was like I was back in my working days making things in a factory, seeing that much variance in the parts being made, instead of detecting the parts that were different, the normal thing to do would be to fix the reason they were varying.

Because you don’t measure the diameter of a square or rectangular object. The TS explicitly stated he wanted to measure the diameter.

 

I think a 7 bit Gray code encoder would be one solution. As the TS need one degree resolution over 90 degrees he need 91 positions. 7 bits would give up to 128 positions. Many years ago I made an 8 bit Gray code encoder for an antenna position indicator. I drew it out on paper (I think twice full size.) and filled in the positions to block the light with Indian ink. I suspect now you could find a program online to generate the pattern. It was then converted to a transparency which was mounted between a light source and a group of 8 photo detectors.

Les.

 

If the former, then a rotary encoder would be high on my list of approaches.

I would be looking at that idea also, you can get low priced incremental encoders such as CUI AMT versions up to 1k pulses/rev and others.
One channel only would be required.
Would need referencing at power up.
Max.

 

Because you don’t measure the diameter of a square or rectangular object. The TS explicitly stated he wanted to measure the diameter.

Should have gone back to the original post before stating that I guess. But the second part of my post still applies, why such a big fluctuation in sizes?

And if these are in a 'track' how close can the measurement even get when using a contact probe?
We made wire insulation that was measured coming out of an extruder, using a laser measuring device, no contact needed. And the accuracy was way closer than is needed here.

 

If you want to measure 0-90° rotational angle you can use a reflective optical sensor.