[HuFlungDung]

I thought I was being safe I have the setting turned on to have the Haas reread the program whenever I start somewhere other than the beginning. However, on my 1996 machine's software version, there is still a potential for a problem.

What seems to be the case, is that the machine normally moves to the 'recovery position' just before the line that I elected to restart on. That seemed clear to me from past experience. However, I had a circumstance where I happened to be using a positive Z value in a work offset. Apparently, when the machine makes the move to this 'recovery position' this does not include a regard for a positive Z in the work offset. Instead, it seems to just move to Z0 in that work offset (initially).

This happened to be a lower position than I expected (it was not the actual Z coordinate of the recovery position as written in the program), and fortunately, the tool was almost clear of the part, but it did move down and graze the side of the part. I was at 5% rapid, because I was being cautious on a complex setup, and was jealous of my multi-angle head's safety Nonetheless, this movement did not make any sense to me when I saw it happening.

I don't know if anyone else has noticed anything like this, or not?

[Geof]

I have found the Haas Restart is not foolproof; not that I am implying either you or I are fools .

I have not encountered what you describe but I did have a lulu of a crash a few years ago with a Restart. I was using the Mirror command, mirroring X and Y which is equivalent to a rotation. I did a Restart that had the last command position ahead of the Mirror command so the machine went to that position but it did not read the Mirror so it went to the Non-Mirrored position which was not yet machined and tried to go to the bottom of a non-existent hole.

Since then I have standardized as far as possible on Restarting at a location in the program where the last command position was a tool change, i.e. Z0.0 in machine coordinates.

[Kool Parts]

HuFlungDung,
Restarts are tricky for sure...defenatly not the time for 100% rapids(nuts) ...it wants to move to the last z level on the line before restart. I found a way to trick the control by restarting on a z move, and adding a extra z move to start on....
.
G3 X1.1345 Y0.8514 I0.0114 J0.2097
G2 X2.1525 Y0 I0.153 J-0.8514
G1 G40 X2.19
G1 Z-0.25 F15.0
G1 Z-0.25 F15.0
G41 Y0.0375 F55.0 D12
G3 X2.1525 Y0 I0 J-0.0375 F15.0
G2 X1.1345 Y-0.8514 I-0.865 J0 F55.0

On this one if you started on the second z move down instead of the first one...it would (read) restart on the first z line...instead of the G40 line...make sense????

[HuFlungDung]

Yes, that makes sense. In fact, that is what I ended up doing to actually get restarted, was to insert a dummy move in front of the line I had to begin on.

[1ctoolfool]

it seems the newer machines are better at this, my 99 VFE I never trusted to re-start correctly, but the 06 minimill seems to do an excellent job.

I don't use mirroring, scaling, or G18, G19 which I think could be potential problems - but never thought about a positive Z offset, have to watch for that.

thanks
joev