Sounds like a digital tach. or a marker pulse.
I bought this little stepper to experiment with thinking it had an encoder but now after getting it I can see it only has 3 wires. I hooked it upto my breadboard to test it thinking maybe I would get lucky and it would be a hall effect not that I would have wanted that but it would be better than a 1 PPR tach... for all but 1/8 turn of the shaft the voltage is high on the signal wire but then for that 1/8 turn the voltage drops to 0.
Sounds like a digital tach. or a marker pulse.
CNC, Mechatronics Integration and Custom Machine Design (Skype Avail).
“Logic will get you from A to B. Imagination will take you everywhere.”
It properly can be called an encoder, it wouldn't work as a tach. People are using encoders as tachs, but you need more than one pulse per rotation to do a half decent job of that.
Curious what you were going to do with an encoder. Stepper drive with encoders is not done very often
Once you get a regular pulse wave out of an encoder, you could inded feed that signal into a V-F converter (ala National Semiconductor's LM2917 IC). At that point, one could READILY turn the encoder signal into a tachometer signal suitable for determining RPM.
Effectively, this is what Bridgeport seems to have done with the PWMINT board on their TorqCut and other systems so equipped. Can't verify it as nobody seems to have a schematic of the PWMINT card but I'll bet they're doing something like that on it.
A few paragrpahs of "thought" from Smert about why I bought this little motor and why I am getting a different motor and drilling it myself so I can mount that us-digital encoder I bought 2 months ago before entering an idle time for this project... ($$ earnage has to come first)
I am learning electronics with the intent of creating a loss less stepper that can easily be integrated into existing stepper systems or be made stand alone with its own stepper drivers. I believe I have the math/programming skills required to link my board and some other stepper driver together using encoders. Problem is I have had to start with very little almost no electronics experience. Not really a problem I guess just pricey, timely and a learning curve but the more I read the closer I get to understanding the fundamental electronics circuits I didn't understand before I started so its just a matter of time.
Biggest "next" choice now will be which chip to use... If I interpret the graycode from the encoder myself using a program in a chip I limit the rate at which transitions can be interpreted and if I use something like microchips QEI I increase chip cost by about $5 but the limitation disappears. I want to keep price as low as possible but to make an encoder board that can keep up with some of the really fast drivers QEI is a must have. I do think that for pulse rates at about 45khz I could interpret transitions for 3-4 encoders on a $1 chip and just have the results sent to another "master chip" for each axis of motion that also costs $1 as opposed to using 3-4 $6 ones. All depends on how much I want a board to be able to do I guess. For a more modular approach I would use 2 $1 chips for each axis rather than sharing an encoder transition interpreter with 3-4 axis driver chips.
Would be great to be the seller of a low cost "closed loop" driver board. If I can be satisfied with it I know others will. I am months away from a prototype at least although I'll have something put together within days of getting all my equipment... spent over $700 and alot of time so far on this project and have yet to do anything more than test a bogus encoder on a breadboard with a volt meter. At least nothing even remotely resembling a circuit. Mostly its just thinking and pondering the what/ifs right now so once those questions are answered I'll be ready for some design work with the breadboards.
I know alot of users claim feedback for steppers is not necessary but using my TAIG for hole drilling in aluminum with a .1875 drill at slow feeds I see thousandths of drift accumulating (I think other users dont even check I seriously do believe they don't check). Use servos?... nice but I am thinking of the hobbyist without a scope. Closed loop steppers will be "ready to run" without tuning like alot of the drivers available but lossless. Alot of low speed torque and enough high speed torque for positioning moves at a fair speed especially when encoders are used as there will be no lost steps on those rapids.
Food for thought: the position error mentioned in your last paragraph could be due to mechanical anomalies, not electronic or computer anomalies.
A 'backlash' search on this website should uncover a lot of info as to the root cause of cumulative errors like that you mention. Backlash causes more issues than most people are willing to recognize.
I've done a lot of studying into a particular computer integrated servo drive system for a 2D, 3D and optional 4D mill/VMC. When you see what's involved in linking the outside world into the digital world, it becomes very easy to see why drift, postion control and ultimate machine accuracy are SO difficult to integrate. In its day, it was a REAL pricey system. Today, you could probably duplicate it for much less with the PLC's that are available.
On the machine I"ve been working with, the iron was in very good shape. We had to assume that the digital side was reasonably well developed as it was a commercial system. Even so, accuracy left a lot to be desired, especially if you did a lot of direction reversals.
When it was all said and done, it was NOT the electronics or computer side that was the weak point - it was the mechanicals that caused the problems. Sadly, it turned out that some creative computer tricks were being exploited to try to make up for the use of less than optimum mechanicals - tricks that simply could not fix bad part integration practices. In that case, I contend that you could have a super computer and STILL not be able to compensate completely for sloppy mechanicals.
I have some information that may be of assistance to you in your endeavors. It would give you some insight into what's going on in a well developed albeit dated closed loop F/B system. Please contact me via P/M if you are interested.
Four encoders can be read from a single 8bit port.
To insure a good read 4x sample - 5.5us processing loop time
assuming 25 cycles per channel of decode time gives you a processor clock of 18MHz - easily done with a SX microcontroller - but you'll still need to pump up the clock in order to get the info off-chip!
You probably won't find a $1 chip to do this sitting on a shelf though - the Altera Max3000 CPLDs will get you close, but I don't think four interfaces will fit on one chip.
there are people here on this forum that check their systems obsessively. I believe a well tuned stepper system will not accumulate errors, or else they wouldn't still be around. Performance is what drives people to servos, not accumulated errors.
I'm not sure there is a significant enough difference in price between new steppers and servos to really drive a market.
Rather than buying a stepper with an encoder used and risking the chance of another bad item I took the 4th axis stepper for my Xylotex system and drilled holes in it for a us-digital 300-E4P encoder. Went well the encoder seems to work after the install I have checked it with LEDs and my scope.