Spindle speed feedback

Sure will, but it'll be tonight before I can get to. Thought about posting it last night but ran out of time.

motion.spindle-revs is an input, and counts the number of revolutions. I assume gmoccapy.spindle_feedback_bar is an input also, so that won't work.

This would be more correct, but encoder velocity output is in units/s and you probably want your display in RPM. Using a scale component can multiply the value by 60. That should be in your HAL already if you used stepconf or pncconf.

Note that in the first line we assign the encoder velocity value to motion.spindle-speed-in. In doing so, the signal of spindle-velocity holds that value, so that you can assign it to other inputs. (pins can't connect directly to other pins. signals are the interfaces between them)
encoder.00.velocity is an output, as documented here: linuxcnc.org/docs/html/man/man9/hostmot2.9.html#encoder
motion.spindle-speed-in is an input, as documented here: linuxcnc.org/docs/html/man/man9/motion.9.html#PINS

and I'm not sure if software encoder uses encoder.0.velocity or encoder.00.velocity, like the hardware encoder counter.

Is it a PNP or an NPN prox?

NPN prox can be fairly-easily connected to 5V inputs. PNP are a little harder

Basically an NPN prox can be considered as a switch to the 0V line, and a PNP is a switch to the +24V line.

www.mesanet.com/pdf/parallel/7i77man.pdf

Says that the encoder inputs are OK to 12V. Most prox switches need at least 6V, so there is an ovelap there.

Just got in, here is the requested HAL file. I'll dig through the other information after dinner.

andypugh wrote: Is it a PNP or an NPN prox?

NPN prox can be fairly-easily connected to 5V inputs. PNP are a little harder

Basically an NPN prox can be considered as a switch to the 0V line, and a PNP is a switch to the +24V line.

www.mesanet.com/pdf/parallel/7i77man.pdf

Says that the encoder inputs are OK to 12V. Most prox switches need at least 6V, so there is an ovelap there.

Its an NPN.

For NPN and a Mesa 7i77 I think that you would need a pull-up resistor to the encoder 5V supply, and then allow the Prox to pull that pin down to ground for index. I think I would start with a 1k.
(You might need to invert the input too)

Thank for the suggestion. That should be a lot better than the software encoder. Once I can confirm I am getting reasonably correct spindle feedback I've got to implement a servo to adjust the spindle speed. That machine originally had an AC servo turning the handle where you normally adjust the spindle speed on a manual mill. You guys have been an invaluable resource, thank you, my machine wouldn't be running without the help I've received here.

EDIT: I moved the gmoccapy line from the main hal file to the postgui hal file and the errors went away. everything starts up and I can start hal show and see the spindle speed change there, it just doesn't update in the gui for some reason. I suspect that a gmoccapy specific problem to pursue. As long as I can see the value updating in hal show I should be able to continue on with setting up the servo for the closed-loop control, correct?

Well, I got some time to play work on the machine this afternoon. Got the spindle feedback bar working in gmoccapy and it matches what I see in hal show. The problem is the values are way off. Put my scope on it and saw that the signal was in the kilohertz range so I switched from using the software encoder to the hardware encoder 05 on the 7i77 because I figured the frequency was too fast. This hal stuff is slowly starting to make sense, I got it working correctly, however, the pulse rate is way off by the same amount. Setting the encoder scale to 1 results in the display reading roughly half the actual spindle speed. the problem is I can clearly count approximately 34-36 teeth on the spindle looking through the sensor hole. Its worth noting as well that the scope shows a nice stable signal, I'm stumped on this one.

biqut2 wrote: however, the pulse rate is way off by the same amount. Setting the encoder scale to 1 results in the display reading roughly half the actual spindle speed. the problem is I can clearly count approximately 34-36 teeth on the spindle looking through the sensor hole. Its worth noting as well that the scope shows a nice stable signal, I'm stumped on this one.

Is it possible that the readout is showing 34/60 of the true speed?

It sounds a lot like a confusion between encoder velocity (RPS) and spindle speed (RPM)