I'm thinking about buying a few of those $10 or $20 encoders on eBay. The specs say they are 600 or 1000 lines and 'NPN Open Colector". I would worry a little about the signal when spinning them fast. At the high end I'd have 50KHz square waves but with square waves I'd have considerable power up to at least 250KHz, (that is the 5th harmonic) because they are square waves.
Open collector transistors are likely not the best cable drivers. and also the sensor is a physical device where a physical dark object slides over an aperture. This makes for a raising edge that depends of shaft rotational velocity. The best designs I've seen use a comparer with a bit of feedback applied to create hysteresis. So that each edges is latched. I think you really do need hysteresis, at the very least a 74ls14 at the sensor end of the cable On Wed, Nov 8, 2017 at 8:51 AM, Jon Elson <[email protected]> wrote: > On 11/07/2017 10:54 PM, Gene Heskett wrote: >> >> >> They are visible on my 100 mhz hitachi dual trace, but not to the extent >> shown by the pix I just sent, Jon. Quantization errors keep adding up. > > Well, this may actually be informative. What are the voltage levels coming > out of the encoder? > Are they a good match for the logic thresholds of the Mesa board? Maybe a > pull-up or terminating resistor might help clean up what the Mesa sees. If > the A and B HAL signals are AFTER the digital filter, then these width > variations may be caused by the Mesa board rejecting noise on the inputs. > I've seen a lot of problems with encoders where cable capacitance causes > both signals to transition at the same time, and this illegal transition is > rejected by the digital filter. The transition is not recognized until the > crosstalk-caused transition corrects itself. With weak pull-up, that can > take some time. >> >> I'm looking at the Omron 2500 line encoder for a touch over $40, and at >> least a month to get across the big pond. >> >> But while it looks to have ball bearings, they are probably vxd's. Shitly >> stuff thats not really fit for roller skate wheels. >> >> No clue how hard it might be to put decent bearings in it. > > Many encoders are not meant to be repaired. The glass disks are plenty > fragile, and getting the read head off and the shaft pressed out without > breaking them is quite tricky. Even the ones with etched steel discs are > fragile. >> >> Can you pretend to advise? > > There are also "kit encoders" where you get a disk and read head, and maybe > an enclosure, and fit it to an existing shaft. You could make up a bearing > nest and shaft and fit that to it. Avago, US Digital, Renco and others make > these, and they are not terribly expensive. You can get a wide variety of > the Avago (formerly HP HEDS-series) from Digi-Key and Avnet. You should > have no problem getting one of these kits for $40, but of course that does > NOT have the bearings in it. I think US Digital has complete encoders close > to that price, with the etched steel disk. I have used the Avago HEDS-type > encoders to mount on motors with rear shafts. Quite easy to do. So, if you > can manage to fit a rear shaft, that would be the best way to go. > > I'm not sure a higher resolution encoder will really solve your problem. As > you raise encoder resolution, the timing information you get from the edges > gets smaller and smaller. >> >> As in is the idea worth pursuing? Or should I >> make another of those disks, but first get my A axis moving, which >> should remove any variations in slot timing by elimination the xy >> backlash's incomplete compensations. But I'd still be stuck with a low >> slot count unless I can locate some mills with only 0.010 tip diameters. > > No, I would not try to do any big milling job with .010" end mills. Got any > hair? You won't after experiencing that! > >> IIRC I used several mills out of a ten pack of .028" diameter mills to >> make that one, and don't have any more of those left. >> >> That also begs the question about how small a slot is practical with >> these honeywell interrupters too. I should tour yard sales looking for >> dead $5 printers I suppose... They have very fine pitched wheels and >> matching encoders in /some/ of them. >> > Well, I would not use slot-type sensors for an encoder. There are sensors > specifically made for encoder use. > > Jon > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > _______________________________________________ > Emc-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-users -- Chris Albertson Redondo Beach, California ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
