the ULTIMATE encoder… if you can get it stable.

from an engineering college, one DIY version.

An easier to digest version

Now, the theory is as follows:

We send a laser through a beam splitter. The two beams now hit two mirrors. They bounce back and the beam splitter now mixes the two signals making interference patterns (rings ideally) on a target down the fourth leg. As the rings show up at 1/4 wavelength intervals, it's possible to create sub micrometer linear encoder with this. At 532nm (green laser) that's 108nm per change. Once you have a solid interference pattern, it would be relatively easy to put two photodiodes at the rige places, convert their signals into square waves, and feed an encoder input for an incremental encoder. You're not going to get an absolute position sensor out of this.

Why is a green laser pointer good? Unlike a red pointer, it actually uses a crystal to frequency double 1064nm light (remember the frequency relationship is 1/wavelength). This naturally filters out a lot of variation in the light and creates a more coherent beam. while a red laser pointer will work, it may not have as good cavity control and can have its wavelength wander, making accurate measurements more difficult. I do NOT expect to get the same results I'd get off a single mode HeNe laser with any laser pointer, though.

Now, all Michelson interometers I've seen illustrated move the target that's directly across from the beam splitter. I guess that's because it may be a 10% beamsplitter so to not overload the mixing .

Making one of these will also be difficult. Not because it's difficult conceptually, but because it needs to be still and dimensionally stable so flexing of the interferometer structure itself doesn't change the values. Now, I'm not worried too much about that, as if I used this for encoder feedback on a project, I'm NOT going to attempt 100um resolution on a hobbyist budget.

So, using 100um. That's 10,000 counts per mm, or 2500 electrical cycles (A-B quadrature). So, that's 2.5MHz/m/s. Many industrial encoder inputs max out at 1MHz, so 0.4m/s would be the maximum move rate they could record. I see higher end ones capable of 10MHz, so taht would be 4m/s.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License