Do-More & H2-CTRIO2 ???

[scott.lawrence]

I just changed out a DL260 PLC & two H2-CTRIOs to this new set-up...Do-More PLC with three H2-CTRIO modules

Existing : three STP-DRV-80100s, and three STP-MTR-34127s.

I separated each axis because of the homing issues I've been experiencing. The machine will home perfectly, but intermittently, an issue will arise where either the axis doesn't change direction when the prox is made, or will stop at the leading edge of prox instead of trailing edge... operators will think the machine homed, but really is up to 1/4" off.

 I have tried both the CTAXLMT & the CTRUNPOS and still experience this phenom. Oddly enough, this only happens when using a two-step instruction, i.e., a CTAXLMT that runs to limit A, then changes direction to limit B. I have one axis (vertical) that doesn't need to compensate for backlash (gravity + weight), and this axis only runs to limit A. The other two? It's random enough for it to be frustrating. Completing 50 or so home routines, then out of nowhere an issue!

This is an ongoing problem, and others have reported issues with homing using other instructions as well. This got me thinking...  would homing reverse/forward (backlash comp) be better accomplished using two homing instructions that were single move?

Oh yeah, I did update the H2-CTRIO2s with the firmware Greg sent.... did not resolve

[BobO]

The CTAXLIMT in the reverse direction was very much broken and the firmware change fixed it. The nature of the problem would not have allowed it to work at all in certain configurations...nothing random there.

On the working for a while, and then not, I wonder if you are getting some noise or bounce that is being mistaken for the target edge. The CTRIO2 has programmable filtering, you might try to play with the filter settings.

[scott.lawrence]

Bob, thanks for replying... I have been in contact via email with Greg over this for a few weeks now, but the latest WIN7 update has knocked out my OFFICE 2007, so I can't email

The randomness was referring to the issue of not changing direction and/or stopping at the leading edge of the home prox... either way, the instruction would not complete, say, one out of 40-60 times, then would 'hiccup' for about 3-7 times consecutively, then another 30-60 times without issues.

On a positive note, I have managed to overcome this by using two single-action instructions instead of a single two-step instruction. Using either two CTRUNPOS or two CTAXLMT have proved successful, where the two-step instructions have given me fits...

[BobO]

That sounds like a bug then. We'll put together a more extended test and see if we can duplicate it.