Linear encoder (absolute)

Needed: an absolute linear encoder for the range 4 inches (about 10 cm) that works with the Jrk 21v3 feedback input.

The resolution should be about a quarter of a mm [correction, 1/40], which is about what the Jrk can handle in the given range (10 cm / 4096).

Surely someone has used a linear encoder withJrk 21v3 or v12. Can you recommend one?

Pololu carries the Sharp GP2D120 optical distance sensor (Pololu # 1136). I couldn’t find the resolution on Sharp’s data sheet. One potential problem for my application is that this device takes time to read the distance. The Sharp data sheet says 39 ms or about 4 hundredths of a second.

The object whose distance will be sensed moves up to 5 cm/sec, so in .04 sec it might move 2 mm. The speed will be monitored using the formula “change in distance divided by change in time.” In the application I have in mind it won’t matter if there’s a slight delay between the actual speed and the computed speed, however it’s essential that one varies smoothly with the other. Too much jitter in the distance measurement will throw off the speed measurement.

Every once in a great while – apparently a very great while – someone might like to use a linear encoder with a Pololu Jrk 21v3 motor controller. If I understand the controllers specs, the feedback terminal on these controllers requires a voltage (0 to 5 V) or some sort of pulse input that I don’t understand, they are not designed for sine-cosine quadrature signals.

A company called Celesco makes linear potentiometers that look like they would work. They claim the resolution is “essentially infinite” – I suspect they mean “virtually infinite”. The following information is from Celesco’s website.

… Item … Range, … Ohms, … Linearity
… MLP-12 … 0.5 in, … 1.25 K, … ±1%
… MLP-25 … 1 in, … 2.5 K, … ±1%
… MLP-50 … 2 in, … 5 K, … ±1%
… MLP-75 … 3 in, … 7.5 K, … ±0.5%
… MLP-100 … 4 in, … 10K, … ±0.5%
… MLP-125 … 5 in, … 10K, … ±0.5%
… MLP-150 … 6 in, … 10K, … ±0.5%

They’re rated at 42 V, so 5 V can be put across them.


The linear actuators we carry work with the jrk motor controllers. I don’t have anything else to suggest, but anything that fits within the voltage and current requirements of the jrks and has an analog feedback voltage between 0 and 5 V should work. If it uses a pot, I recommend you look for one with a 1k - 10k pot, though it doesn’t have to be exactly in this range.

Also, note that 10 cm / 4096 would be a resolution of 1/40 mm, not a quarter millimeter. Our linear actuators use potentiometers for the feedback, so the resolution you can get with the jrk is a function of the total feedback voltage range (some our generic actuators, for example, might only range from 2 to 4 V when the feedback pot is powered at 5 V).

As for the Sharp sensor, you can see a graph of its output voltage as a function of distance. Note that the voltage does not change linearly with distance, so you should expect the resolution you can achieve to vary with distance (e.g. the voltage difference from 5 cm to 10 cm is over 2 V, while the voltage difference from 20 cm to 40 cm is less than half a volt).

- Ben


Thanks for your reply. I inserted your resolution correction (1/40 mm instead of 1/4 mm) in brackets in my first post.

I’ll be using a voice coil linear actuator with the Jrk motor controller. Unlike Pololu’s geared linear actuators with feedback the voice coil doesn’t come with a distance sensor built in.

The Celesco linear pots mentioned earlier are expensive. In single quantities from the manufacturer:
CLP-25 … $279
CLP-100 … $294
MLP-25 … $358
MLP-100 … $422
and I imagine other “industrial” brands would be high priced too. Probably the price would be a lot less from a distributor, but still pricey.

I can’t really comment on what would work best for your application, but I suspect you do not need a pot that is so expensive (or at least that you will be limited by other factors long before you take full advantage of the features that make the pots so expensive). My advice would be to try an inexpensive linear pot to start with and see if that is good enough.

- Ben