Incorrect 5x feederates on Siemens 840d

[Manofwar]

Like the title says, I'm seeing issues with feedrates on a DMG Mori DMU-40 with a Siemens 840d control. Issue is during a 5x path (simple pocket routine wrapped on a cylinder and locked to 4th axis.) the machine feed during Y (or X) only moves is moving drastically faster then when C is commanded as well. The machine is running G94, I did already try FGROUP commands, but with little success. Time estimation in CAM for cycle is around 5 minutes, on machine it takes about an hour. Uncertain where the issue is. Any help would be appreciated.

[Greg Williams]

A Mastercam zip2go would help, the more info you have the easier the solution is to find

[?Mark]

FGROUP(....) function should do the trick.

Which axis did you insert in the group? Try inserting only the axis used in the toolpath...(XYZC)?

[Slick]

Manofwar, I have a background with the Siemens 840D, although not on a DMG Mori.

I am interested if the machine has a TRAORI function?

[Manofwar]

Sorry I missed the responses to this. I did get it working.

@Greg Williams Program isn't built with Mcam, so I can't share a file.

@?Mark FGroup alone wasn't fixing the issue. I did every combination I could think of for this. FGroup (X,Y,Z,C) still made the C axis very slow during C only moves (probably around 20% of what it should have been moving).

@Slick The machine does have TRAORI and it is commanded during machining.

Fix I found for this with Mark's help was to add FGref[C]=1.000 where 1 is the radial distance from center of the cylinder. Of note that both FGroup and FGref needed to be below the CYCLE832 line in order to function.

[Slick]

(Sorry for such a delayed follow up)

I was curious how close the true cycle time was to what Mastercam reported?

You mentioned Mastercam was reporting 5 minutes.

[Manofwar]

@Slick Not using MC for this. Running another CAM software at the moment. Time to run was about 6 minutes. I know from previous experience that Mastercam was pretty good with times. If you removed tool changes and other ancillary actions from the machine, the time would be about +-5% of the predicted time. (At least for the machines and posts I was using.) This would change significantly for the 5 axis paths if you were running something not COR as the machine would have much further to travel for each move.