Spindle and encoder

Updated specifications of the OD and ID of the pulley have been inserted into Masso's code wheel generator, printed and affixed to the pulley face.  www.masso.com.au/optical-encoder-pattern-generator.php .  The spindle encoder is configured to 7 PPR which is 7 black squares per revolution.  MDI, g97 900 m3 but the portable tachometer read 600 RPM.  Can the commanded RPM be calibrated to match the tachometer reading and how?  The GUI RPM is still shuffling.  Is it velocity, is it convertible to RPM and how?  The Hal Scope of channels A and B might of improved but the 90 degrees out still isn't as close as expected.  Can the analog signal be improved anymore with the existing encoder and how? 

The A,B signals are still not really quadrature so I would not expect
the velocity (or more importantly, the spindle count/position) to be
correct.

Also the noise in some of the pulses (also indexes)
indicates either that the encoder is simple not working
well or maybe does not work reliably with the 7I96's TTL input
threshold.

It might require an actual oscilloscope to determine signal level
/signal integrity issues.
 

I don't think you should bother too much about the speed for now. The thing is that you've set your PID values  at 0, except FF0 at 1.0. This means that there's actually no PID correction at all. The speed is then for now solely controlled by the FF0. If you want an easy way to increase the actual speed you may simply slightly increase the FF0 value, maybe start with 1.2. This should give better correlation between the commanded speed and the actual speed, but still you will not have any pid correction.

This being said,I strongly suggest that you first focus on the encoder parts, which is better now but still not good enough to me.

Main problem is that you still have little 'spikes' here and there instead of a perfectly uniform train of pulses. This could be caused by a slight misaligment of the sensor or maybe the signal itself.

Other than that I would like to see a better view of the hal scope. It is going to be much easier to see if the quadrature is right if you extend  the width of the pulse on the scope. A perfect quadrature means that A lead (or lags) B of 1/4 cycle. We really need just a few full cycle to see that.

 

... @pcw seems that I was writing just about the same as you at about the very same time..

I'd be willing to try another encoder. Sherline did have a rotary encoder with a pulley and considerable to use. Are you knowledgeable about it and is compatible with 7i96s? Is suggestible encoder, what brand and model to use with the 7i96s and the Chucker Lathe?

Preferably a single-ended encoder due to an 8 pin female circular connector at control box. Included is a wire with an end that’s an 8 pin circular connector and other end a DB9 which maybe female. Can a differential encoder be configured single-ended?

8 wires is enough for a differential encoder
(which is preferable for noise immunity)

Differential encoders can be used single ended

I don’t mind the differential and probably would like to try it for a change but the circular connector includes P1, P2, P3, A+, B+, Z+, V+ and Gnd. I think the differential would include three more terminals A-, B- and Z-. The brush motor has it own 115VAC plug in receptacle. I could put a new connector on the control box but rather not if possible because how meticulous soldering is. I might want to get a differential encoder anyway just in case the quality is required. However if the differential could get configured to single-end it’d be convenient.

I wonder if the code wheel lines cord width could be incorrect regardless of them being printed from the Masso generator. Maybe the Masso encoder is working okay regardless not as well as expected. Should the encoder actually be encoding the motor spindle? Should then the encoding from the headstock spindle and headstock and motor pulleys ratio configure to the motor spindle for accurate encoding?

With the current noise issues it would not work for threading.

For threading you can use the counter mode (A only) but it would
make the resolution only 1/4 of what quadrature would give you

The noise makes me think either there is some issue with the
disk (low reflectivity of white or too high reflectivity of black)
or the signal does not reliably swing close to ground