AXBB + Clearpath = Intermittent ENABLE signal

[grecobene]

Currently having some issues with an AXBB + Clearpath servos (used previously UB300ETH + Clearpath with great success).

I have two servos hooked up to the AXBB and one of them has an intermittent “ENABLE” issue. Steps will be lost (servo is de-enabled) at random times, for a random amount of time (from a quick glitch to a few seconds). Only occurs on one servo, on the EN signal (STEP and DIR are unaffected).

On the UCCNC software side, the Diagnostics window shows a constant “EN” signal. On the Clearpath MSP software I can see that the servo input “Enable” is lost at the random times stated above.

On the hardware side the initial wiring was done according to the CNCdrive doc (all negative STEP/DIR/EN pins to 5V0 and the positive pins to the AXBB non-isolated terminals), I always measure a steady 4.730V at the EN pin, even when the servo loses the enable signal (power supply is set at 4,95V). When in series with the EN wire, I measure a steady 6-7mA, so it’s not a loose wire.

When I tried to probe the signal with a scope yesterday, I grounded the 5V0 pin to the scope’s ref and the 5V faulted on the AXBB (The EN pin started drawing 47mA when that happened). I unplugged everything and restarted and the problem went away, could not reproduce it. The issue came back today: tried to switch the control wires of both servos and the problem went away again. I then tried to switch the wiring to “sinking” (all positive STEP/DIR/EN pins to 5V+ and their negative counterparts to the AXBB terminals) and the problem is still there: intermittent enable state of the servo.

I am suspecting that I have the same issue then the fella here:
https://www.forum.cncdrive.com/viewtopic.php?f=4&t=2150
But the guy never came back with the fix he did…

I am tempted to simply ground the 5V0 rail to “earth” to make sure there are no floating references, but this was not recommended by CNCdrive (at least, not link 5V0 and 24V0 to keep the isolated outputs separate).

This setup drives an expensive and dangerous machine… I cannot afford to lose steps here and there and cause a crash. Any idea on where to go from here to fix this?

[cncdrive]

What do you mean by "5Volt faults"?
The 5Volts on the non isolated output or a non isolated output of the EN signal on the AXBB-E.

[cncdrive]

I would try to wire the EN signal directly to the 5Volts powersupply rails to cross out a possible drive problem.

[grecobene]

Sorry for the late reply, before the holydays I could not reproduce the problem. But now it surfaced again so I could do some testing.

What do you mean by "5Volt faults"?

The red light marked "Power Failure" above the 5V+/5V0 inputs turned on and the CNCdrive controller shut off. Unplug and re-plug fixed the issue.

I would try to wire the EN signal directly to the 5Volts powersupply rails to cross out a possible drive problem.

I didn't get to test this as the last test I did "fixed" the issue (swapping back and forth the control cables, see below). Will try next time the problem occurs.

After posting the issue, I switched the logic of the drive: I hooked up all the "+" of the Step/Dir/En to the 5V+, connected the "-" servo pins to the non-isolated CNCdrive inputs and set the logic to "Active Low" in the software. The servo kept de-enabling for a while but the problem eventually went away: I couldn't reproduce the bug since then.
Yesterday I added the 3rd axis to the machine so I re-wired everything to make things cleaner. Out of curiosity, I restored the initial logic (grounded all the "-" of Step/Dir/En and hooked the "+" pins to the CNCdrive inputs) and the issue is back, with the same servo (not the other 2), only with the "Enable" signal. I swapped the "Dir" and "Enable" pins to see if the issue would follow the pin, but the intermittent "Enable" kept happening and the "Dir" signal never switched off. I then re-tested to swap the servo control cables and the problem disappeared. Even when swapping back the control cables, the problem is no longer there.

I have shut down the machine for an hour and turned it on again: no issue. I expect to have it happen again tomorrow morning...

[grecobene]

Ok, the culprit is most likely the servo.

I ran Teknic's "battery test" (connect a 9V battery directly to the ENABLE terminals of the servo) and the problem was still there. I measured a steady 14mA drawn through the motor even when the EN was lost.