Re: [CAD_CAM_EDM_DRO] Ref: Laser Metrology/ Encoders

stan wrote:

> So not much chance of an el-cheapo interferometer type thing, it's super
> accurate lab type stuff all the way.
> Thanks very much for the reply, it cleared up my confusion. I was thinking of
> waves as the surface of the water going up and down which seems like a bit of
> a misconception regarding light.....and there was some quantum too.
> So all in all the 2 beams are necessary as the phasing of the beams is what
> makes the wavelength measurable.
> And a laser pointer isn't really a usable source as the wavelengths are
> messed up (not sure if its the lengths or the alignment or if I'm barking up
> the wrong tree altogether).
> Besides all that, something would be needed to count all the phases and there
> are around 4 million of them a meter

4 million/meter would equate to a wavelength of 250 nm, in the
UV. Oh, maybe you are multiplying by 4 for quadrature counting,
then that would be a wavelength of about 1 um. 1067 nm lasers
are available in Nd/YAG, but they are high power pulsed devices.
The standard HeNe laser is at 632.8 nm (red visible). That
would give 6.32 million counts/meter in quadrature.
if the sensor is beside the laser and

> the beam is being reflected back from whatever is being measured. Actually,
> that bit isn't too bad and 4000 divisions per mm is a usefull number.
> Could you give me a better idea on laser pointers or cd drive lasers as a
> light source?

None of these will work, their coherence length is a couple mm.
Any interferometer longer than that will just read "mush".
You need a laser with a coherence length several times the
longest length to be measured to get good, sharp interference
patterns that can be detected. HeNe gas lasers do a lot better.
One problem is those run hot, and the thermal expansion causes
"mode hopping", where laser power output can jump rapidly
between two or more levels, often as much as a 2:1 step change
in intensity.
That makes the detector circuit a little harder to do right.
The concept doesn't seem to tricky but something to hold it all

> together on a vibrating machine looks difficult, it seems like it would need
> 3 or more side by side to tell which direction the thing is moving. Oh,
> and 'if you need to ask, you can't afford it' prisms, but it would be nice
> even to see it working on a steady surface.

The only expensive part is the retarder plate that sorts out the
interference for the detector. These run something like $400
at Edmund Scientific, you can probably get them cheaper from the
actual manufacturer. What this does is turn plane polarized
light into circularly polarized. There may be another
interferometer scheme that doesn't need the retarder plate.

> Looking over that, I'm assuming the phases stay where they are if the
> distance stays the same and aren't going in and out at some speed related to
> the speed of light. is that right?

Assuming there are no air currents, no temperature change, no
vibration, etc. In general, it will be bouncing over a range of
counts, and your counter needs to keep up with the flicker of
the waves to maintain the position reading. All that flicker
will be in the um range, so you won't even bother to display it,
but the counter needs to maintain the count.

Jon