3
\$\begingroup\$

What is the best way to detect the position / displacement of a linear motor ( solenoid type ) for feedback ? I was thinking of using a hall effect sensor or an Infrared sensor .

\$\endgroup\$
3
  • \$\begingroup\$ try googling linear position transducer \$\endgroup\$ Jul 24, 2013 at 9:19
  • \$\begingroup\$ Are you trying to determine if the motor has moved at all (e.g. a binary thing), or exactly how far the actuator has moved? \$\endgroup\$ Jul 24, 2013 at 9:20
  • \$\begingroup\$ trying to find out exactly how far it has moved \$\endgroup\$
    – subz
    Jul 24, 2013 at 9:23

2 Answers 2

3
\$\begingroup\$

"Best" covers a lot of ground: accuracy, resolution and cost. High resolution was discussed before in terms of LVDTs. If you are willing to trade off resolution for accuracy, then a simple way of doing this is to have a few opto-interrupter sensors spaced at accurately known locations, and attach a flag to the linear motor that blocks the sensors as it moves.

At slightly higher cost, and roughly equal accuracy, but higher resolution is to use an off the shelf rotary encoder with a mechanism that converts linear to rotary motion. The simplest way is by wrapping a string around the shaft of the encoder and fixing the two ends of the string to the ends of the object that moves. Another method is to press a gear onto the encoder shaft and attach a "gear rack" of some type to the linear motion shaft. The gear rack can be an actual rack that mates with the rotary gear or a section of timing belt that is glued flat to act as a rack. This arrangement is generally known as rack and pinion rack and pinion

\$\endgroup\$
2
  • 1
    \$\begingroup\$ +1. Neat animation. I kept waiting for it to switch directions. Ahh well, another 10 minutes wasted! ;-) \$\endgroup\$ Jul 24, 2013 at 20:15
  • \$\begingroup\$ Wish I could take credit for it, but it's from the rack & pinion wikipedia entry. \$\endgroup\$
    – lyndon
    Jul 24, 2013 at 23:37
2
\$\begingroup\$

For a really good accurate system try a Linear Variable differential transformer. Here is a link.

Here's the blurb from the supplier above: -

LVDT's are inductive displacement transducers that have proven themselves as very robust and particularly suitable for use in installations where high vibrations, high accelerations and/or very long cycle lifes are to be expected. Even oscillations with small sweeps, and measurements under 1mm can be achieved. Displacement ranges from 2mm to 200mm are available. Typical LVDT sensors are energized with AC voltage, produce an AC output and thereby require external electronics for amplification and processing. Our LVDT products all incorporate internal electronics allowing the transducer to be supplied with either ±15VDC or 24VDC and provide an output signal of 0 - 5V, 0 - 10V, 0 - 20mA or 4-20mA. This feature allows for simple field integration.

They basically operate like a 3-coil metal detector. Centre coil applies an alternating magnetic field to the outer coils which are wound antiphase. A piece of metal (usually an iron compound) is moved across the three coils by displacement. At midpoint balance there is no net signal; either side there is a signal whose amplitude and phase tell you direction from centre-point and distance.

EDIT Here is a website that shows you how to build one and make the electronics too.

\$\endgroup\$
6
  • \$\begingroup\$ +1, interesting find, and so obvious once one hears of it! \$\endgroup\$ Jul 24, 2013 at 9:44
  • \$\begingroup\$ You can get really big ones (Used on flight simulator motion systems) and rotary ones. Better than 1mm on 2m stroke if I remember correctly. \$\endgroup\$
    – Spoon
    Jul 24, 2013 at 11:21
  • \$\begingroup\$ LVDTs have insane resolution, but the signal processing can be a bit hairy and the sensors (transformers) themselves are generally well out of "hobby" price range unless you DIY. I have a couple of industrial LVDT instruments that were around $5k each new. If you don't need resolution down to the micrometer, then a linear magnetic sensor like the ones from Austria Microsystems or an off the shelf rotary encoder coupled to a string or rack & pinion is probably good enough and a couple orders of magnitude cheaper. \$\endgroup\$
    – lyndon
    Jul 24, 2013 at 12:54
  • \$\begingroup\$ @lyndon Read the question LOL "What is the best way" etc.. Best probably does equate the LVDTs. Why don't you post your comment as an answer? \$\endgroup\$
    – Andy aka
    Jul 24, 2013 at 13:23
  • 1
    \$\begingroup\$ @Andyaka, if accuracy and precision are the goal, and cost is no object, there are systems out there that can achieve below 1 nanometer. But that's not the "best" solution. The best solution is the lowest cost system that meets the system requirements. \$\endgroup\$
    – The Photon
    Jul 26, 2013 at 15:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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