0
\$\begingroup\$

What's an appropriate sensor for cheaply and accurately measuring an angle of rotation?

I'm trying to measure the angle of a joint in a robotic limb using a small potentiometer mounted at the axis of rotation, and although it's small and inexpensive, it's accuracy and susceptibility to noise is horrible.

I need something equally small and cheap, but much more accurate.

The only other technology that comes close are hall effect sensors. I've seen some SoC that can detect a rotating magnet above them and output the angle as a signal. However, I can't find any of them packaged in a form that's as small as a potentiometer.

Do any of these come in small package sizes? What's their accuracy compared to a potentiometer measured with an ADC? How much more immune would a hall effect sensor be?

\$\endgroup\$
  • \$\begingroup\$ An optical encoder is probably going to be your best bet for a balance of cost and accuracy. My very first experience with rotary encoders was with what was then just called a "selsyn." Aka synchro. It was for my WW II (1944 commission) naval radar unit. But for any real accuracy, you will pay a lot in cost and space and... rarity. I suspect few things are as boutique, now, as a selsyn. I've used a lot of optical encoders and they are worth having. However, I have no experience using them with robotic arms so I cannot speak to that specific use. \$\endgroup\$ – jonk Aug 21 '18 at 5:15
  • \$\begingroup\$ Anybody want a selsyn? Reasonable cost, but as @jonk says, they're huge, probably far too big for your robot elbow. \$\endgroup\$ – Neil_UK Aug 21 '18 at 6:01
  • 1
    \$\begingroup\$ rotary annular magnetic sensors offer crazy resolution at reasonable price. Renishaw and Austria Microsystems are names to look for. \$\endgroup\$ – Henry Crun Aug 21 '18 at 8:11
  • 1
    \$\begingroup\$ "Cheaply and accurately" don't usually go together. You will only get good answers if you specify the required resolution, accuracy, and cost with numerical values. -1 \$\endgroup\$ – Elliot Alderson Aug 21 '18 at 15:30
  • 1
    \$\begingroup\$ @ElliotAlderson, By cheap, I mean having a cost < $30/unit. By accurate, I mean a digital signal with at minimum 1024 CPR, preferably absolute position. \$\endgroup\$ – Cerin Aug 21 '18 at 16:41
2
\$\begingroup\$

Have you thought about using inductive angular sensors. Specifically, inductive angular resolver?

They are used also in robotics as in automotive industry and are well known for its high resolution and can work well in noisy environment.

Principal how they work you can find on following link: https://en.wikipedia.org/wiki/Resolver_(electrical)

\$\endgroup\$
1
\$\begingroup\$

There are very accurate and fairly low cost absolute angle encoders based in the old Inductosyn patents. I have not found a source for the controller chip used or the specific designs, but the internals are in the digital protractors available now and are definitely worth researching or duplicating.

I took one of these apart and found it is indeed based on the Inductosyn rotary encoder principle. https://www.ebay.com/itm/iGaging-11-Electronic-Digital-Protractor-Goniometer-Angle-Finder-Miter-Gauge/130837893293?epid=28016984636&hash=item1e768bd0ad:g:gmkAAOSwnDZT-naA

\$\endgroup\$
  • \$\begingroup\$ it may be possible to hack one of these protractors ... then just drill some holes in the metal ruler part and mount it on the robotic arm \$\endgroup\$ – jsotola Aug 21 '18 at 5:48
  • \$\begingroup\$ That looks interesting, but I really don't want to hack a protractor to get a sensor. Is the sensor component by itself sold anywhere? \$\endgroup\$ – Cerin Aug 21 '18 at 13:58
  • \$\begingroup\$ Could you edit your post and replace the eBay link with more durable references, please? \$\endgroup\$ – Nick Alexeev Aug 22 '18 at 14:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.