Millport VMC1660 Centroid Upgrade (Leadshine EL7)

[KreiderMachine]

Well... It remains to be seen how many of our remaining issues will be affected by this, but I did find something in investigating the encoder quadrature error that seems to resolve that issue:

On our A520 inverter orientation control pcb (encoder inputs) there were extra jumpered terminals that had never been disconnected when integrated with the cnc controller. By removing the jumpers, I was able to eliminate the quadrature error. It remains to be seen whether that will also affect the spindle fault issue we were having today, but a win is a win.

[centroid467]

Ok, it looks like the asterisk is basically solid when running at low speed. I can detect the slightest flicker in it down around 10-20 rpm, but nothing like a pulse you could observe at 120rpm, or even 60. Just looks solid.

It didn't lose counts, really. Gained one count in 490 revolutions if my math is right...

Do I have any options here, other than trying for a different encoder/cable routing?

Is it possible to invert the index/Z pulse output in the VFD settings? This description sounds like it is high most of the time and goes low when the pulse is reached.

[cncsnw]

I think that he confirmed normal index pulse behavior (one brief flash per turn, always in the same place) when he was turning the spindle by hand.

I think he is saying that the nearly-solid asterisk only appears when the VFD is powering the motor.

[KreiderMachine]

Correct, yes. When turning by hand, the pulse is in a repeatable place and only shows up for a single count.

[centroid467]

I missed that, sorry. Please disregard my comments and questions there.

[KreiderMachine]

All is not well, after all.

Now, periodically, we are getting error #9066 - Orient Timeout, and the quadrature error still shows up from time to time.

Other things I have noticed:

The spindle speed seems to fluctuate. Might be normal, but my experience isn't broad enough to know if this is noteworthy or not.

The index pulse still isn't regular. It isn't constant anymore, but it does seem to have an extra stutter to it.

I'll experiment this morning with selectively replacing those jumpers I had removed.

Any input/advice is greatly appreciated.

-h

[KreiderMachine]

This video shows the issues referenced earlier:

[KreiderMachine]

Do any of you know whether the encoder channel inputs have signal terminator resistance designed in? I don't understand fully, but it seems that a ~100ohm resistance is key to dampening channel signal, and should only be found at the most distant termination of the signal chain in a situation like this, where the signal is also read by the VFD. It seems that with the resistance left in place at the VFD, the signal is overdamped, but with it removed, it seems to be a bit unstable.

I have changed the following parameters in the VFD:

Pr. 351 - Orientation speed: changed from 4hz to 2hz
Pr. 355 - DC Injection Brake Start Position: - Increased from 12 to 16
Pr. 365 - Orientation Time Limit - Set to 7s (Had been disabled, the implications of which I am unsure of)

These changes were arrived at after giving all relevant manuals, as well as my list of parameters as currently configured, to ChatGPT and describing the situation. It seems to me that Pr. 366 (Recheck time) and pr. 362 (Position loop gain) could apply, but they remain at factory defaults for now.

With these changes made, I re-ran autotune on the VFD, and it hasn't triggered the #9066 orient timeout fault in the 6-10 toolchanges I commanded manually following the reboot.

Furthur insights welcomed.

[centroid467]

There is a termination network on the Hickory encoder inputs with 150 ohm resistors. I do not know how that would interact with the network on the FR-A5AP encoder card.

[KreiderMachine]

For reference:

We have been having a a problem with the carousel crashing during automatic tool changes following any manual command to rotate.

Finally got around to addressing this this morning by editing the M50 and M51 macros to match the M6 macro by changing

"M100 /70301 ; Wait for DoingIndex_M to reset"
to
"M101 /70311 ;Wait for carousel at Requested Tool Position"

Seems to have resolved the issue. If there is a reason why this is inadviseable, let me know.