Millport VMC1660 Centroid Upgrade (Leadshine EL7)

[KreiderMachine]

Update to previous description of issue: One of the brakes was occurred while jogging, and not during a probing cycle. (in the z direction).

The detail about the machine stopping (on its own, but late) during the probing cycle break is unconfirmed. The operator referring this to me isn't the person who was there when it happened.

[cnckeith]

Update to previous description of issue: One of the brakes was occurred while jogging, and not during a probing cycle. (in the z direction).

The detail about the machine stopping (on its own, but late) during the probing cycle break is unconfirmed. The operator referring this to me isn't the person who was there when it happened.

lots of "455 Spindle axis(8) encoder quadrature error" messages in the log, what's going on there?

[cnckeith]

Sorry for the delay on this, Keith. I've been busy with the lathe, and as long as the guys have been able to work around the issue, I've been leaving it be. The probing cycle would have been the work-zero measurement. If we just put the probe in, it detects it (at least it indcates that the spindle is disabled) but then the z-measurement is based on the tool that was supposed to be there, rather than the probe (if my memory serves.)

I'll revisit this issue at some point and try to get some fresh observations.

In the mean time, we have had a recurring issue when using intercon with an imported .dxf. We had an issue with a windows error coming up that says: "A fatal exeception has been detected. The application will be terminated." And then crashes CNC12. It isn't the same file each time, but it was the same project.

thanks for posting the files, just seeing this post now. looking into it.

[cnckeith]

A new issue with the probing on this machine.

The guys have been breaking probe tips, because the machine doesn't seem to be responding in time to the probe trigger. They say it is intermittent (most of the time it works fine) but twice in the last two days, it has broken a probe tip when probing when the probe triggers, but the machine continues to jog. I am told that the light on the probe is lit when the breaking of the tip occurs, and that the machine does stop, it just stops too late.
Jogging speed at time of breakage is 10ipm.

Any insights on this?

I'll follow up this post with a fresh report, and I plan to verify all connections to the probe port as soon as they have a pause in the use of the machine.

-H

something is really wrong if the probe tripped is not stopping cnc12.
start with basics and review/watch the cnc12 trigger input (virtual LED Alt-i screen) and trigger by hand to observe and verify the input being made consistently. then jog axis slowly with probe in the air and trigger to make sure it stops every time, then do the same thing with a simple probing cycle (edge).

(we have hickory on a machining center in our shop that probes every day and uses TT.)

[KreiderMachine]

We never figured that out. I made some efforts to isolate the signal cable, thinking that it was due to noise, but it's the same encoder and cable routing as before the upgrade, and the spindle encoder feedback seems intelligible to the Hickory, so we just disabled the fault and have been using it like this. It's the same db9 plug that plugged into the old Servo4.

Update to previous description of issue: One of the brakes was occurred while jogging, and not during a probing cycle. (in the z direction).

The detail about the machine stopping (on its own, but late) during the probing cycle break is unconfirmed. The operator referring this to me isn't the person who was there when it happened.

lots of "455 Spindle axis(8) encoder quadrature error" messages in the log, what's going on there?

[KreiderMachine]

Yes, and it seems to be intermittent. Everytime I have investigated, it works fine. Next time I am in the shop and the machine is idle, I plan to inspect all connectors and splices in the probe port cable, and ohm out all pins in the connector. If it's intermittent, though, I don't know how hopeful to be that this will turn up anything.

-H

A new issue with the probing on this machine.

The guys have been breaking probe tips, because the machine doesn't seem to be responding in time to the probe trigger. They say it is intermittent (most of the time it works fine) but twice in the last two days, it has broken a probe tip when probing when the probe triggers, but the machine continues to jog. I am told that the light on the probe is lit when the breaking of the tip occurs, and that the machine does stop, it just stops too late.
Jogging speed at time of breakage is 10ipm.

Any insights on this?

I'll follow up this post with a fresh report, and I plan to verify all connections to the probe port as soon as they have a pause in the use of the machine.

-H

something is really wrong if the probe tripped is not stopping cnc12.
start with basics and review/watch the cnc12 trigger input (virtual LED Alt-i screen) and trigger by hand to observe and verify the input being made consistently. then jog axis slowly with probe in the air and trigger to make sure it stops every time, then do the same thing with a simple probing cycle (edge).

(we have hickory on a machining center in our shop that probes every day and uses TT.)

[KreiderMachine]

Well, now we've got a fresh problem, which may or may not be related to the spindle encoder. Three times today, the VFD has thrown a spindle fault, and kicked out of the program. This has happened before in Frequency Mode, but never in Vector Mode (We can switch between them, by using a pair of buttons on the VCP/Operators panel), so something new has begun to happen.

I am speculating, at this point, that the spindle encoder quadrature error may be related (i.e. whatever hiccup the Hickory is flagging, has now gotten bad enough that it triggers a control loop error in VFD itself).

The spindle encoder is a Sumtak LHE-408-1024 encoder, and seems to show smooth operation when observing its feedback in the Encoder menu. The Encoder Quadrature error occurs dependably anytime the spindle is commanded to power up, either manually or automatically.

This is my best guess, but I'm facing the possibility that this machine may be unuseable, again, after just putting it back into service after months of down-time for the Hickory upgrade.

Do you have any advice on how I might test this encoder, without having an oscilloscope?

[cncsnw]

Do you have any advice on how I might test this encoder, without having an oscilloscope?

An imperfect test, but one that does not require an oscilloscope, is to watch the behavior of the spindle encoder index pulse and A/B counts, on the PID Configuration screen.

First, set P334 = 1 so that a quadrature error on the spindle encoder does not trigger a fault.

Then go to the PID Configuration screen and watch the "Abs Pos" value and the asterisk next to the 'N', as you turn the spindle around by hand while in emergency stop. Put a mark on the spindle where the index pulse comes around.

At the MDI prompt (which you can call up from the PID screen and other sub-menus with Alt-M), enter "M150", press Cycle Start, and continue turning the spindle until the you pass the index pulse and the "M150" that you typed disappears from the screen. M150 will zero the spindle encoder count at the index pulse.

Turning the spindle by hand, verify that the index pulse comes around (the asterisk flashes) every 4096 counts (assuming that the numbers in your 3/27 report were correct), and always at the position you marked.

Then run the spindle under power for a while at low speed (e.g. "M3 S120"). At 120 RPM, you should see the asterisk flash twice per second.

If the asterisk is on more or less continuously at 60 - 120 RPM, that is a sign of noise on the encoder lines.

After it has run for a while, stop the motor, turn the spindle by hand back to the index pulse, and see if the "Abs Pos" count is still at a whole multiple of 4096. It could reasonably be off by two or three counts. If it is off by dozens of counts, that is a sign that noise causing extra counts or dropped counts.

[KreiderMachine]

Sounds Good. I'll do this.

In other news, because we have encoder feedback to the VFD, which the encoder for the Hickory just piggybacks onto, I was able to figure out that A channel and B channel are reversed, at least between the orientation control board on the VFD and the DB9 that plugs into the Hickory encoder port.

At the terminals, I tried the following:

Switch A and B channels for the Hickory only: - Quadrature Encoder Error Repeats.
Switch A and B Channels from encoder (returning DB9 pigtail to original terminals: - VFD does not like... spindle spins at 30rpm max, and shudders when commanded to stop, as though it had been fighting itself or something.

Switch A and B channels on VFD, but retain original configuration between db9 pigtail and wires to encoder: Same as previous.

I assume that there are VFD parameters to let the VFD know to expect reverse encoder feedback, but none of the three configurations eliminated the encoder quadrature error, so no luck.

I verified the presence of pulses for A+, A/, B+ and B\ (so no completely broken connections.)

-H

[KreiderMachine]

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?