On 5/25/2011 9:26 PM, Jon Elson wrote: > Dave wrote: > >> How would that SSI to quadrature encoder work? The nice thing about >> SSI is that when the power goes on, the motion system knows exactly >> where it is without homing or moving. >> >> > That only works if a memory is kept alive to remember the rotation count > (assuming shaft > encoders). If a linear encoder, the number of bits becomes unwieldy > pretty quickly, and then you > go back to some kind of cyclical count plus a multiple counter. > >> I'm trying to find an encoder now. The best I have found far on the >> surplus market is a 360 count per rev unit which is not very good. I'd >> like to find at least a 4096 count per rev unit or better yet a linear >> unit that I could actually use to mockup with a hydraulic cylinder. >> >> > A 4096 count/rev unit would need 12 optical sensors, also the higher the > count, the more likely you > would start to get artifacts due to alignment of the sensors. I would > guess these units might get > real expensive at the higher resolution. > > The Fanuc aA64 encoder has an absolute position with 1024 > counts/quadrant for commutation, and > a position count with 16 bits resolution on angle plus a 16-bit rotation > count. It uses a quadrature > optical disc with interpolation. There is also a 15-bit angle version > and a 20-bit version. > This seems to be the sanest way to make such an encoder. > > Jon >
I've been working on another project where the control system is blowing SSI encoders for some unknown reason.. Heidenhain is being called in with their test gear as four have failed so far. Two went out initially due to wiring errors (don't hook 24 volts to an incremental channel and expect it to survive - the encoders have both incremental and SSI data outputs). But then two other ones went out over a period of time for some unknown reason, so I have been studying up on encoders - more than I would prefer. The diagrams/drawings that I have seen for high count SSI encoders appear to use a multiple tracks on a single disk. The encoder part number says it has 512 counts per rev but it actually puts out 4096 positions per rev. The diagrams they use to depict the disk do not show binary looking graduations but uneven, gray code type encoding methods (or something different). Not at all like a quadrature encoder where the increments are simply 90 degrees out of phase. I didn't dig into the actual encoding method but it did not seem overly complex. These encoders are intelligent. When the encoders was damaged, they were smart enough to send out an error message to the controller. Unfortunately the controller is too dumb to interpret the error messages so it simply declared the encoders faulty. In my quest to get replacement encoders, Heidenhain told me that SSI encoders are becoming scarce as Endat has become a much more popular interface for encoders (or at least the ones they sell). Dave ------------------------------------------------------------------------------ vRanger cuts backup time in half-while increasing security. With the market-leading solution for virtual backup and recovery, you get blazing-fast, flexible, and affordable data protection. Download your free trial now. http://p.sf.net/sfu/quest-d2dcopy1 _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
