1/2" Shaft Encoders

How do most people use encoders on half inch drive shafts? Is it easiest to convert down to a smaller shaft or are there reliable encoders with a 1/2" hole (and if so, any specific suggestions)?


We are using 2 of the KOP rotary encoders, one on a 5/8" shaft and one on a 3/8" shaft. What we did was drill a hole in the end of the larger shaft, take a shaft that is the right size for the encoder, insert that into the hole and epoxy it in. Then all you have to do is build a bracket for the encoder and you’re good to go. We have used this on our Arm for about 2 weeks with pretty extensive testing and haven’t had any problems yet.

We usually do the opposite of that, turning down the end of our large shafts to fit the encoder. With our ancient lathe, we feel better about turning a concentric protrusion on the end than we do about trying to drill a concentric hole. Plus, we can turn it to the right diameter for exactly the fit we want.

Here’s what we do: http://www.youtube.com/watch?v=GwDFD-TjDFU

Drill & tap the end of the shaft 1/4-20, then locktite a 1/4" threaded rod/stud/(bolt with the head cut off)…

We are having problems with this method now. On Saturday, on our 2nd arm joint there was a “dead spot” of about 30 degrees where our encoder wasn’t registering a change in distance. The disc was rotating and it would work once past this dead spot but within it there would be no change. And then after about 2 hours of other work our 1st arm joint had the same problem. I suspect that it is due to a mechanical alignment of the shaft that we inserted, as described above. Our 2nd arm joint had a noticeable angle offset (maybe 3 degrees) although the 1st arm joint did not have anything noticeable. We were unable to troubleshoot this any farther due to running out of time so if someone else has any ideas that would be interesting to verify.

Was the hole put in the end of the shaft using a lathe? Any other method will never be straight enough to work. If it is a joint on an arm in the classic sense, then it may be better suited for a potentiometer anyway. The optical encoders are better for continuous rotation applications like drive wheels.

Potentiometers are usually easy to get at places like Radio Shack. Plus, it has the added advantage of knowing the absolute position of the arm at all times instead of the relative position.

They are easy to mount too, make a little bracket for them and then our team sometimes uses what we call the “Cyber Blue method” where we stretch a very short bit of surgical tubing over each end to act as a flexible coupler.

Wire it up to an analog input and you are done!

Is this a “rule of thumb” or have you actually seen that without using some sort of precision rotation method to change the diameter that even a small angle offset will result in the behavior I described?

We have used potentiometers before. I am normally more of a fan of the encoders as they can deal with being overdriven and are not noisy, but these are fairly limited justifications as you can easily get a multi-turn pot and the aren’t the analog inputs aren’t that bad. We used encoders more just because they were readily available. We ended up just using timing and within about 3 hours had our autonomous program running completely successfully so we will be going with an open-loop control system, as referenced here: http://www.chiefdelphi.com/forums/showthread.php?threadid=91793

A quick and low-precision-required trick is to use some pieces of vinyl tubing. We did this our first few years with potentiometers and encoders because it doesn’t require that the shafts be aligned within thousandths of an inch.

A large piece was heated up and slipped onto the large shaft, and one or two pieces of smaller tubing were used to build up the shaft on the encoder. Zip ties and fabric pins were used to secure the pieces. Obviously this method doesn’t work for through-hole encoders, although I suppose you could mount them with a shaft and bearing and then connect them with tubing if you were desperate.



We still use the rubber tube method, and it works great. We have a small round cylinder that is the same diameter as the encoder shaft machined into the end of the shaft we want to attach the encoder to. We then use a small piece of rubber tube to attach the shaft to the encoder. This allows the encoder to not be aligned perfectly, and for everything to still work nicely.

I suppose if you had a very nice vise with a vertical V-groove and a more-than-decent mill you might be able to pull it off if you can get it centered down the shaft (probably with some fancy edge-finding). The lathe makes it easier to just center-drill, drill and ream all from the tailstock. Or you can choose to turn your own pin onto the end like I mentioned above. Any misalignment at all with the pin could result in dead zones or even scratch the disk on the lenses. Of course if open-loop works for you, go with it :slight_smile: