Hardinge HC Bandit Retrofit

After a big interruption on this project, I have bee able to spend a little time on it again.

I completed most of the wiring, and started working on the Hal files. The resolvers worked the first time I turned it on! I had to spend some time getting the scale right, but now they match the axis movement. I have not energized the servo motors yet, I wanted to sort out some of the other controls first.

Now I have some more problems and questions.

1. There is a section of GPIO that does not respond to the switches. These switches are wired the same as the ones that do respond. They are called out in the Hal file the same way:

net ro-incr-a <= hm2_5i24.0.gpio.043.in

The pins show up with halshow pins, but if I show parameters, there is nothing listed for gpio 42 through 47.
Am I missing something? Is my7I42 board bad?

2. Is there a way to call up a 4, 8, or 12 element MUX-Generic with s32? For bits or floating it looks like it is b12 or f12. For s32 is it s3212? That doesn't seem to work.

3. My spindle has a mechanical speed adjustment that can be sped up or slowed down by activating two relays. Can I use the PID to control this? It would be on/ off output instead of proportional signal. Maybe have a fairly large dead band so it does not hunt. I know that many oven controllers use PID to cycle the heater on and off without a proportional output.

The spindle motor also has two speeds, but I can probably treat that like a gear selection.

Thanks,
Mike

1. There is a section of GPIO that does not respond to the switches. These switches are wired the same as the ones that do respond. They are called out in the Hal file the same way:

net ro-incr-a <= hm2_5i24.0.gpio.043.in

The pins show up with halshow pins, but if I show parameters, there is nothing listed for gpio 42 through 47.
Am I missing something? Is my7I42 board bad?

HostMot2 firmware pins can only be fuil GPIO pins with all parameters
if they are not claimed by other functions (PWM gens, stepgens, encoders etc)
I suspect that the pins that are missing parameters have been claimed by another function.
The way to fix this is to check the firmwares pinout and then only enable the required number
of special functions (in the hostmot2 command line)

Note that if you don't specify the number of instances a specific function,
_all_ instances will be enabled.

Thanks PCW!

I had num_encoders=2, but I added the resolvers, and PWM generators.
Now they show up on the list.

Edited. found mistake.

I figured out my mux-generic issue. I have a twelve position switch, so I was trying loadrt mux_generic config="ss12" which would give an error. Using ss16 worked.

It is a coded switch with 4 binary outputs.

Ok, I attempted getting the motors going, and now I am really lost.

First off, the DRO will track the axis movements. To get the DRO to match the actual movements, I had to put in a resolver scale of 0.05. Mechanically the motor makes 10 revolutions per inch, so that would mean the resolver makes 0.005 pulses per revolution. That does not make sense, so either resolvers are treated differently than encoders, or I have the set up wrong.

When the amplifiers are energized the motors both turn slowly. Linuxcnc makes no attempt at stopping them, and eventually gets a follow error. For now I have the amplifier enable independent of Linuxcnc, and the motor speed does not change when Linuxcnc disables. I was able to get the motors to stop by turning the "B" pot on the amps which I guessed to be bias.

I am now able to move the x axis. It moves fairly smooth in the (-) direction, by shutters violently with (+) moves. When off, moving the axis by hand there is no noticeable difference in resistance for either direction. The z axis is unpredictable.
There are tachometer inputs to the amps. Could the tachometer be wired backwards and fighting control?
Should all the tuning on the amp be disabled? Should the tachs be disabled?

I have no documentation for the amps. They are brushed DC and based on Glentek GA 4559-1.
There are 6 pots marked CL, G, C, B, T, and S. Other than B for bias, I have no clue what the rest do.
I did run the motors/amps with the original Bandit control. The x axis is the same, but the z amp was replaced with another used one because the original smoked when it was disabled. It still worked though.

How do I determine if the amps require velocity, position, or torque control? How do I select the different modes in Linuxcnc?

Any help would be appreciated.

If you have tachometers, they are velocity mode drives
(slow drifting if enabled but with no feedback from the controller is also suggestive of velocity mode drives)

Getting the drive enables to work and controlled by LinuxCNC should be the first step (for safety reasons)

Typically if the tachometers are wired backwards, you will get an immediate runaway, the fact that you have slow drift is a good sign.

The resolver scale is probably OK, how are the resolvers driven? (if a 1speed resolver was directly drive from the ball screw I would expect a scale of 0.100")