We are experiencing a rather strange phenomenon.
When we mill flats using the G77, the first 3-4 parts will end up with the flats not being parallel. after these scrap parts the machine will run good parts. (?)
does anyone else have this problem?

We always use the milling cycle - g112 if I recall correctly.

For some reason your C-axis isn't locking (the brake isn't engaging fully) for the first 3-4 parts. Does the finish look acceptable on the parts with the nonparallel flats? Does the machine "hunt" for the proper C-axis angle before it settles down, or does it engage promptly?

The c-axis seems to engage properly and the finish is ok. it appears as if the c-axis does not reference properly. this is an intermittant problem. we can run the machine for several weeks without a problem and then it will occur over a period of several days. maybe we should reference the c-axis first by sending it to 0 degrees before we start?

try a G4 P1. after you orient the spindle before going in and cut.

could be a problem with the encoder.

This may be a bit far out and I hope I can explain myself clearly.

The G77 command has an optional I value that specifies the angle of the first flat. Are you using an I value or do you just omit the I? If you are just leaving it out I suggest you try putting in I 0.

The reason I suggest this is that I once had a funny problem on a mill using the Bolt Circle canned cycle. My bolt circle came out rotated even though I did not use the I value in the command line that set up the conditions for the canned cycle. What I finally discovered was that the canned cycle was reading an I value that had been entered into a memory register by a previous command line. Just omitting the I from the Bolt Circle command did not zero this register and the bolt circle canned cycle used the existing. entry. I solved the problem by putting in the explicit I0. to set up the canned cycle.

It is possible you have something similar.

Our Haas rep. told us not to bother with the flatting cycle & just program it for the milling cycle. We never had a problem after that.

thanks for the suggestions.
I realize that the haas machines are rather sparse in their software, but what is there should at least work properly ;-)
Will give the above mentioned ideas a try. if not then we will just have to program manually.

Sending C axis to zero first is always a good idea. We were milling some 45* flats on rectangular stock in our Daewoo Puma lathe and the vibration from the stock hanging out coupled with our feeds and speeds and material all helped sound the death knell for the live tool bevel gears. So...I guess where I'm going with this, is that there are always going to be some work-arounds you may have to deal with, unless of course you buy a $$$million dollar machine, in which case, I want the damn thing to load, unload and program itself, before tucking me into bed at night. :D