# Micron-Scale Linear Actuation

I don't know a lot about linear actuation, but I need a precise device that I can control easily (using some kind of hobby board) that is able to perform linear positioning with a resolution of at least 1 micron. I've seen piezo actuators, stepping actuators, ultrasonic actuators but I'm not sure which is the simplest. Is this possible? What would such a device be called?

Thanks.

• What is the range of motion required? – jms Feb 8 '16 at 23:42
• @jms Anywhere from 10-100mm would be great, although I'd be interested in any actuators capable of operating with this precision. – Freddy Tuxworth Feb 8 '16 at 23:42
• That is rather huge for the resolution you want. Piezos sprang right to mind (quartz changes shape under applied electric field) until I hit that comment. – Ecnerwal Feb 8 '16 at 23:53
• @Ecnerwal I see. What would be a more reasonable expectation? – Freddy Tuxworth Feb 8 '16 at 23:54
• Also, the Piezo LEGS Linear 6N by PiezoMotor seems to have a significantly larger range, 80 mm, but unsure on price. – a sandwhich Feb 8 '16 at 23:57

One micron resolution in a 10mm distance is one part in 10k.

I am presently building a micro-leveller for an inclinometer which will have something like those specifications, for a few GBP.

It starts with a stainless steel M5 threaded rod (studding), driven by a geared stepper motor (28BYJ-48, pocket money on eBay), stepped by an Arduino, running between two ball races, with a brass nut. The rod has a 0.8mm or an 800 micron pitch.

The motor claims it has 2k steps per revolution, so you could theoretically achieve sub micron resolution with just that setup. Well, maybe if you only stepped in one direction. I have measured the backlash and 80 per revolution is a more conservative repeatable accuracy for it.

Turning the rod by 1/80th turn gets 10um resolution at the nut. I gear that down with a lever. In my case it's a compound lever to get 100:1 reduction, giving 100nm resolution. You would only have to gear down 10:1, or take your chances with no levers and fine adjustment in one direction only.

It's very slow, relies for linearity on a cheap bit of threaded rod, and I suspect wear with use will be quite high. But it comes in at the right price.

• Thanks for your answer. I wonder if the bearing between the nut and the studding would be loose, and able to move by a few microns? Also, would this be better with a leadscrew, which is still relatively cheap? – Freddy Tuxworth Feb 9 '16 at 8:30
• Yes, yes and yes, but then I have short arm and deep pockets, and a lever gearing that gets me to the resolution I want. The point of the answer is to show it can be done, but you have to be a bit handy with mechanics, and do some careful analysis and measurement to see whether it would work for you. Bear in mind that the levers I use are connected by flexure links, not bearings. – Neil_UK Feb 9 '16 at 10:11

What you're talking about is a high-resolution linear stage. You can get what you want, but it's not remotely in your price range if you're concerned with using a hobby board for control. Newport, for instance, carries a series of linear motorized stages which will give resolution of 1 nm and accuracies better than 1 um, and drive intervals up to 350 mm. Of course, a 100 mm throw unit will cost you 8k+, but they are available.

If you're willing to take chances, eBay currently has a stage with 1 um resolution and +/- 5um accuracy for about 1/3 that.

• Thanks for this. I'll have to look into another solution. – Freddy Tuxworth Feb 9 '16 at 0:09
• @FreddyTuxworth Newport are somewhat known for their high prices. They sell mainly into the scientific market for people who want a reliable pre-engineered solution and are not very sensitive to cost. Other companies such as Parker Motion will have cheaper offerings. What you basically need is an anti-backlash lead screw, stepper motor, and linear and/or rotary encoder. It will not be cheap for good quality components, but they are available for significantly less than one can expect to pay at Newport. You might be looking at $500 rather than$5000. – Oleksandr R. Feb 9 '16 at 15:41