I lost count of my scopes. Now I just want mobility. I came, I saw, I bought some interesting accessories, and put names to faces: NEAF 2012, ASAE 2012, SWAP 2013, ASAE 2013.
------------------ Mert 42º49"N 1º38"W 3" Polarex refractor,6" F12 SW MCT,Moonlight "Stepperized" EQ6 + EQMOD,SPC900NC/DFK21AU04.AS/ASI120MC My web-page Still working to make it decent!
Quote:This however requires that you have an additional, absolute indexer - a normal PEC indexer is a single slot, but a worm rotates once every 4-10 minutes. Since the RA shaft only rotates once in 24 hours, you can't determine the PEC index with a single slot: you need an absolute encoder.
Quote:You can't calibrate the disk by spinning the RA shaft at a fixed speed. Because how can you ensure a fixed speed when any motor or geartrain you'll use to spin the axis would also have its own periodic errors.
Quote:Many commercial encoder manufacturers calibrate their higher-end encoders using the concept you've suggested.. but they use a laser interferometer with sub-arcsecond resolution, or a high-resolution master encoder, to provide the "known good" readings. By mapping the "known good" angular readings with the (distorted) output of the encoder under test, they can construct an error map specific to that encoder.
Quote:The problem is... the error map only takes into account the eccentricity of the encoder "stand alone." Once that encoder is installed, bearing wander, axle swash, etc. will induce its own periodic error which the "canned" map in the encoder can't compensate for. So for the really high-end positioning systems like IC etching machines, they do calibration with an interferometer on the completed system.
Quote:Here are some more encoder references I've coughed up in the course of this project.. (most are available online, but some require IEEE membership and payment)
Quote:The optical disk would have to be accurately mounted. Otherwise, what happens if you jump tracks? So, it isn't quite so easy since the encoder has to be internally consistent. Now, if you could trick a disk burner into printing a perfectly radial set of alignment data, then you could have one which would be able to survive jumping tracks, and then you could conceivably make one out of blu-ray player parts.
Quote:On the issue of indexing the PEC of the encoder, I had an idea: Provide multiple unique indexing points. So, for example, you could have one every 45°, and the mount would be assured to run over various ones periodically just in the alignment routine. Notionally, there could be a pinhole in a disk, and then have 8 detectors, so at the moment it runs over any of the detectors, the alignment of the disk is known.
Quote:An even better way is to have a non repeating code all the way round on the index track. This way you can tell the absolute position when you have moved a few counts in either direction.
Quote:I don't think it's as simple as changing the pattern on the index track. You would need an index track with a large number of unique gray code patterns.
Quote:I believe "all" the high-tic absolute encoders are really pseudo-absolute - they have a high-tic incremental grating, and an absolute index track with a much lower tic count.
Quote:The trouble is you cannot print 32 million tics on a 3" diameter disk. It's trouble enough to print 5000 tics.
Quote:Yes, you need a large number of unique codes, but not necessarily grey codes, and why isn't that simple? It's still only one track. You just replace the single index bit with a more intelligent pattern. No extra hardware is required.
Quote:Why can't the index track have the same resolution as the quadrature track? You should be able to read both tracks whith the same resolution.
Quote:A 3" track will have a length of 239,4mm. With a bit width of 130nm we have a pulse spacing of 260nm and a quadrature resolution of 920.728 pulses on the track. This is a quadrature resolution of 3.682.913 ppr corresponding to an angular resolution of 0,35". Not too shabby for a low cost solution.
Quote:What will that pattern be?
Quote:Remember that the track is binary. So to have say 256 different gray codes, you need 8 tracks ( = 8 bits ). This is why high-bit absolute encoders are very expensive.
Quote:Because the incremental track doesn't have slots. It has a diffraction grating. That's how the interpolation works.
Quote:I don't know where that bit width came from. Normally encoders only have 5-20um physical resolution. Everything else is interpolated.
Quote:These optical drives don't care about angular position. They just need the bits to come off the track at the correct rate. Variations in disk rotational speed are perfectly acceptable, because there's a big memory buffer where the bits are held before they are played out.With an encoder, you don't have that option.
Quote:I think you are conflating the bit width of CDROM, DVD, or bluray with encoders. They are not the same.These optical drives don't care about angular position. They just need the bits to come off the track at the correct rate. Variations in disk rotational speed are perfectly acceptable, because there's a big memory buffer where the bits are held before they are played out.