[Charles2]

Hi Tom,

Trying to configure KMotionCNC. I have the Init3Analog.c program in Thread 2 of the KMotion C Program screen. With it open, from Config & Flash screen, I hit the 'Export All to Open C Program' button.

I then opened KMotionCNC and went to Tool Setup, User Buttons, and set the INIT button for Thread 2, and selected KStep\InitKStep3Axis.c.

From the Tool/Setup Files tab, I selected 'Basic Lathe 2 Axis'

From the Trajectory Planner tab, I set Axis Parameters for X and Z at 80000 Cnts/inch, Vel at 1 in/sec, Accel at 5 in/sec2.
Under Lathe Options, I checked Lathe and X Positive Front.

From the KMotionCNC operator screen, X will jog with the arrow buttons, Z will not. If I click on the Keyboard Jog button, Z will move with the Page Up or Down button and X will move with Left or Right buttons.

If I click the INIT button, nothing moves again until I power everything off and restart.

I'm not too concerned about the buttons yet. I can probably fix that in the screen editor.

I'm guessing I have problems with the Initiation program.

In my case, is the spindle considered the third axis?

I see that the axis' are defined at the end of the Unit file. Is that correct?

Thanks

[Charles2]

See attached INIT file.

File did not attach. Will try again.

[Charles2]

See INIT file

[Charles2]

I found some of my own errors. First, I had the INIT1 button pointing to the wrong place (KStep\InitKStep3Axis.c) instead of (Init3Analog01.c). I thought that looked wrong when I did it but didn't look for the correct directory at the time. Next I had the ' DefineCoordinateSystem' set to (0,1,2,-1) instead of (0,-1,1,-1). This being a lathe, it uses X and Z for axis'. I'm guessing I will change it to (0,-1,1,2) to utilize the spindle as an axis. Changed the square box Gcode program to use X and Z and it ran fine.

[TomKerekes]

Hi Charles,

That all sounds correct.

I'm guessing I will change it to (0,-1,1,2) to utilize the spindle as an axis

Yes. A common practice is to create 2 MCodes to include the Spindle into the coordinate system to control it like an axis or remove it to be controlled as a traditional spindle.

[Charles2]

Do you have an example of the code? On a lathe, is it more common to set it as an axis or as a traditional spindle? Or is it better to set it as an axis in order to thread or use CSS?. How much tuning do I need to do for the spindle as an axis? As an axis, it has some very slight oscillation (about 6 counts on the axis screen) that I can't seem to tune out. When I move x or z, the spindle oscillates more (about 150 counts when I click the jog arrow, then -150 when I let off the button), then goes back to very slight oscillation. Maybe it needs to be disabled at the drive when not called to run.

[Charles2]

Plots and Data for spindle as axis

[TomKerekes]

Hi Charles,

On a lathe, is it more common to set it as an axis or as a traditional spindle?

Its more common to operate as a traditional Spindle. Basically the only time it needs to be configured in the motion system is when it needs to indexed to specific angles.

As a traditional Spindle you can do CSS, single point threading and so forth.

The accuracy doesn't have to be high as the other axes will follow its motion based on Spindle encoder feedback. Probably better to have some following error if that make Spindle motion smoother and more stable.

Since your Spindle works as an axis you can use the Spindle Programs that command an axis to move with standard axis Jog commands with little modifications. See here.

[Charles2]

Hi Tom,

I was able to get the axis limits and homing working.

The spindle, when it is off, does a lot of oscillating, especially when I jog the x or z axis. I don't understand why jogging the axis would affect the spindle. If I hit the INIT button, it settles down, but still moves a little back and forth. When I hit spindle on cw or ccw, is not smooth, but acts like it is hunting for the correct speed. Lower speeds are worse. Do the plots and data above look anywhere near acceptable? If I jog the axis while the spindle is running, the spindle goes nuts, stopping, reversing, I would say completely unstable.

Thanks

[TomKerekes]

Hi Charles,

The spindle, when it is off, does a lot of oscillating

I'm assumeing you mean not commanded to rotate, not off.

I don't understand why jogging the axis would affect the spindle.

Correct it shouldn't. The only thing I can think of is a noise or grounding issue. XZ Motor noise might be injected into the analog command signal to the Spindle. Or motor currents might be taking an unexpected path back to the motor supply causing GND levels to shift.

You might Jog X Z while doing a spindle Step Response Plot to help understand what is going on.

Do the plots and data above look anywhere near acceptable?

They look stable and not oscillating much when stopped. But you are only testing one speed and it is only at speed for ~0.1 seconds so hard to tell. You might try a much longer move.

As you stated before you might turn off the Spindle when it isn't running to avoid oscillation issues while stopped.

Or since accuracy isn't critical you might use Dead band to not make corrections when errors are small.

[Charles2]

Hi Tom,

Yes, I meant not commanded to rotate.

I pulled the spindle drive command cable away from the other cabinet wiring. Ran it outside the cabinet to the Kanalog. Separated it as much as posible from the axis cables on the Kanalog. X, Z, Spindle are connected to 0,1,2 outputs. There is no proximity between the spindle command cable and any motor wires. Checked ground from machine ground bar to spindle drive ground test point ( 2 mV ) with everything powered up. Still have erratic spindle motion. Spindle jerks back and forth when axis is jogged.

Removed spindle command cable from Kanalog. Axis jogging does not affect spindle at all. Connected 1.5 volt battery to spindle command cable. Spindle runs smoothly at 686 RPM FWD or 710 RPM REV if battery is reversed. Spindle drifts without battery about 10 RPM. While spindle is running on battery, jogging does not affect spindle.

Reconnected spindle command cable. Disconnected spindle encoder shield at KFlop. Ran Step Response tests. Jogging does affect spindle. Reconnected shield.

Ran DAC spindle moves. DAC2=500 = 1194 RPM. Slow jog axis moves caused spindle RPM to rise to 1234. Fast jogs similar.

Ran Step Response tests. Retuned with somewhat better results. See Plots and Data. While spindle was reversing , I hit the X jog. Spindle continued to run until I stopped it. See Plot 125. Next run, I hit X jog while spindle was stopped between FWD and REV. Spindle remained stopped. See plot 126. Sometimes when jog was hit, the spindle violently jerked back and forth.

Started spindle with OnCW button (M03). Spindle rotated at 250RPM (jerky). I have noticed that slower RPM means more jerky motion. When I hit axis jog, the spindle goes nuts, switching from FWD to REV violently.

Since the spindle runs its best when the command cable is disconnected from Kanalog and is connected to a battery, I suspect the problem is somewhere in my settings. I suspect I have a conflict somewhere causing a cross connection from the axis (both) to the spindle. It could be a grounding issue, but I have not been able to find it. From the plots, the command goes out of whack when I jog, and the position is just trying to follow.

Please explain how Threads in the C Program page are used. It seems to me the same threads get used over and over with different programs. Is it possible I'm doing something wrong there causing wrong information to get placed somewhere? Should my spindle be something other than ch2, to separate it from the X (ch0) and Z (ch1) axis terminals on Kanalog?

I've read a lot of your responses to problems for different people with apparently all different skill sets. You are a very patient man.
Thanks for all your help.

Charles

[Charles2]

See Data and Plots

[Charles2]

Plots and Data 127 are without jogging.

[Charles2]

Hi Tom.

Maybe I have corrupted KFLOP somehow. Since I was using version 434, I thought I should update to 435f.

Moved 434 to a different folder. Loaded 435f and started all over. Hoping to start with a clean slate. Configured spindle only. Spindle works pretty good, considering the tuning could probably be better. Lower speeds (under about 300 RPM) sound bumpy or a little erratic. Starts and stops good. Probably usable as is.

Configured X axis. Back to same problems. Spindle turns or shutters when I jog axis. If I run the spindle, then jog, the spindle switches back and forth violently. The battery test mentioned earlier seems to indicate the problem is forward of the spindle command cable connection at Kanalog. Any thoughts on what I may try next?

Thanks
Charles

[TomKerekes]

Hi Charles,

Assuming this is noise the first step would be to determine if it is noise in the encoder or the DAC. Move the Spindle to a known physical position and record the position. Then after a lot of erratic behavior move back to that physical position and check if the encoder reads the same as before.

If not the encoder then check the DAC. So the battery test doesn't show the erratic behavior, correct? A battery has an isolated GND. You might try connecting the battery GND to Kanalog GND at Kanalog. This should show erratic behavior if GNDs are different between Kanalog and the Drive. Next you might try a DAC test where the Axis is disabled but a constant DAC voltage is applied which should behave like the battery.

HTH

[Charles2]

It will be a few days before I get back to the machine. I will try your suggestions then.

Thanks

[Charles2]

Hi Tom,

Following your test suggestions:

[QUOTE Move the Spindle to a known physical position and record the position. Then after a lot of erratic behavior move back to that physical position and check if the encoder reads the same as before.][/QUOTE]

When disabled, the spindle idles about 10 RPM. I have adjusted the drive pot to balance this, but it eventually returns. When enabled, the spindle vibrates or oscillates about 10 counts on the axis screen. I enabled the spindle axis (2) and used the Move command (Move2=0), then marked the spindle on the OD and the adjacent stationary surface. Jogged X axis, spindle jerked back and forth a few times. Hit spindle ON CW, spindle rotated at 450 RPM. Jogged X axis, spindle violently jerked back and forth a few times, then stopped. Performed Move2=0, spindle rotated about 400-500 RPM (did not notice RPM on screen), then stopped near mark, and slowly vibrated it's way back to the "0" mark.

Repeated the test. This time when Move2=0 was performed, the spindle rotated for about 5 seconds before stopping near the mark. It took about 3-4 minutes for it to vibrate it's way back to "0".

Moved spindle 2000 counts, then back to 0. It went straight to "0" this time.

If not the encoder then check the DAC. So the battery test doesn't show the erratic behavior, correct?

Correct

You might try connecting the battery GND to Kanalog GND at Kanalog. This should show erratic behavior if GNDs are different between Kanalog and the Drive.

Disconnected spindle command (+) wire from JP11, left spindle command ground on Kanalog. Touched battery + to spindle command + wire, touched battery - to Kanalog ground terminal near JP11. Spindle rotated smoothly. Reconnected spindle command wire to JP11.

Next you might try a DAC test where the Axis is disabled but a constant DAC voltage is applied which should behave like the battery.

Disabled spindle, applied DAC voltage (Dac2=500, spindle ran smoothly at 1182 RPM). Dac2=100 =243 RPM. Dac2=0, spindle stops (idling at about 10 RPM). Dac2=100 =243 RPM, jogged X axis, RPM rose to 260, running smooth.

Performed Move2=0, Axis screen showed 5-12 counts position, but spindle was about 190 degrees away from the marked position. Made a new "0" mark, then after random moves it would repeatedly find the new "0" mark. Note: While doing the battery tests, the spindle encoder cable shield ground was disconnected and reconnected ( no good reason why).

I looked back through all my notes trying to determine how I decided to connect the spindle encoder wires, but didn't find anything. The encoder is a Data Technologies DT25 1024-50/5-SR-03-210. Seems to be obsolete. For some reason that probably made sense at the time I connected the cable as follows:

Shield to GND

BLK to GND
BLK to GND

BLK to A2+
BRN to A2-

BLK to B2+
YEL to B2-

BLK to A5+
GRN to A5-

RED to +5V
WHT to +5V

Each group is a twisted pair.

[Charles2]

Found the logic behind how the spindle encoder was wired to Kanalog. I had mapped the cable shielded/twisted pairs by color to the original control connector. I have a set of electrical prints from a different Warner/Swasey CNC lathe with the same control that came on this one, that has been fairly reliable. From the prints I was able to determine the connections.

Just to be thorough, I opened the encoder Canon style connector to verify the pin connections to my "fairly reliable" prints. Found that chA and chB were reversed from the print pinout. I switched chA and chB at Kanalog and powered up. About 2-3 seconds after initializing, the spindle took off wide open. Switched wires back, now back to same interference as before. Disconnected marker wires, no change.

[TomKerekes]

Jogged X axis, spindle jerked back and forth a few times. Hit spindle ON CW, spindle rotated at 450 RPM. Jogged X axis, spindle violently jerked back and forth a few times, then stopped. Performed Move2=0, spindle rotated about 400-500 RPM (did not notice RPM on screen), then stopped near mark, and slowly vibrated it's way back to the "0" mark.

That would tend to indicate the encoder is counting correctly

Disconnected spindle command (+) wire from JP11, left spindle command ground on Kanalog. Touched battery + to spindle command + wire, touched battery - to Kanalog ground terminal near JP11. Spindle rotated smoothly. Reconnected spindle command wire to JP11.

Well the idea was to Jog XY and see if this caused erratic Spindle motion.

Disabled spindle, applied DAC voltage (Dac2=500, spindle ran smoothly at 1182 RPM). Dac2=100 =243 RPM. Dac2=0, spindle stops (idling at about 10 RPM). Dac2=100 =243 RPM, jogged X axis, RPM rose to 260, running smooth.

This indicates a small amount of interaction. Did the RPM return to 243 after you stopped jogging?

Performed Move2=0, Axis screen showed 5-12 counts position, but spindle was about 190 degrees away from the marked position.

This would indicate the encoder miscounted.

While doing the battery tests, the spindle encoder cable shield ground was disconnected and reconnected ( no good reason why).

You might repeat the above test without disconnecting the encoder shield and see if it still fails.

For some reason that probably made sense at the time I connected the cable as follows:

Shield to GND

BLK to GND
BLK to GND

BLK to A2+
BRN to A2-

BLK to B2+
YEL to B2-

BLK to A5+
GRN to A5-

RED to +5V
WHT to +5V

This seems strange having many black wires. If they are all connected together and/or to GND that might be the problem. There needs to be separate + and - complementary signals. You might check them all with a voltmeter that they all switch from ~ 0.4v to ~4V.

Just to be thorough, I opened the encoder Canon style connector to verify the pin connections to my "fairly reliable" prints. Found that chA and chB were reversed from the print pinout. I switched chA and chB at Kanalog and powered up. About 2-3 seconds after initializing, the spindle took off wide open. Switched wires back, now back to same interference as before. Disconnected marker wires, no change.

That makes sense. Swapping A and B will reverse the direction the encoder counts. So as soon as the servo tries to make a correction it will drive the wrong way, increasing the error, and runaway.


You might post a 3.5 second plot of the Spindle with move size 0 while Jogging XY for a second or two and having "erratic" spindle behavior to see what the plot shows.

[Charles2]

Hi Tom,

Connected the battery again, negative to Kanalog GND near JP11, positive to disconnected positive spindle command wire. Spindle ran smoothly at 730 RPM. Jogged X and Z slow and fast, with no effect on spindle. Removed battery and reconnected positive wire.

Disabled spindle, Dac2=100 yielded 234 RPM, jogged X axis minus, RPM went to 254 then back to 247 when jogging stopped.
X positive took RPM to 252 then back to 243. Z minus to 295 then back to 262, Z positive to 292 then back to 290.

Move2=0 took spindle to the second mark from yesterday. After jogging axis, would sometimes go to mark and sometimes not.

The black wires in the encoder cable are shielded/twisted pairs as grouped above. The black pair connected to 0V are combined and the red/white pair are combined. The others are not combined. The other black wires are the positive compliment to the paired colored wire and are connected to A2+, B2+, and A5+. A5+ and A5- are the marker.

Tried to test voltage on the encoder wires to ground, but even at .5 RPM, my voltmeter may not be fast enough to keep up. Showed about .8 - 2.8 volts.

Ran some plots as you suggested. Oscillation starts when the move button is hit then decreases until jog starts. Plot 128 is without jogging. 129 is with jogging.

Plots to follow.