[camtd]

Thats sounds amazing and fun. Did you have to do a lot of hand editing? Did you have a reference text or manual you used? Was there a NC verifier that you used to check the program?

[aa8vs]

Not a lot of hand editing at all in fact that was a company direction was to generate the tool path using UG and ICAM in order to create a file that you could put on the CNC machine. We used UG extensively to verify what the tool tip was doing with respect to the part in terms of travel. the tool verification if available was when tool approached the fixture and left to go back to toolchange position.

In order to verify this you had to build the model of machine in the system or... you could step through program on CNC itself. Which was most reliable, you would step through at reduced speed to assure general clearances and then once at full speed to assure the timing did not change. Rapid to finished surface stopped about 3 mm short then feed and cast surface 5mm then engage feed.

Biggest drawback we ran into at first was in a single feed stroke change speed and feed three times. With multistage tooling this was required based cross sectional areas of tool. At first it was hard to explain that to the salesfolks and more embarassing when you pointed out that APT could do that. But the later NX systems had finally gotten that one right.

We used the machine manual as a reference on what the various cycles did. In terms of drilling and boring we did that point to point rather than relying on the G8X cycles. It was found that each time the cycle line was read in it took 180 milli-seconds. Using G0 / G1 eliminated this but we could not use the G0 / G1 when rigid tapping.