Expected part length accuracy when transferring with 8 axis lathe

I have a Nakamura Tome WT150-II that has ATS true length S-20 collet chucks. Just finished a run of a couple hundred bushings that were +/-.002 on length. Just a basic drill, turn, bore on the main spindle then part off and transfer to the sub (counter) spindle, finish the second side and eject. I would get 20 or so parts that were within tenths of each other then it would kick off a few that were -.002/-.003 then go right back to holding size within tenths. It didn't seem to matter if I was mid bar, end of bar, or beginning of the stock. My infeed onto the part was 50ipm so I cant imagine that being a variable as far as positioning goes.

A sudden jump .002-.003" is a process control issue, ie chips in the jaws or something like that. A decent warm up on machines like this make a big difference in making tight tolerance parts and maintaining it throughout the day.

When it comes to Z length on a machine of this configuration you have 4 places of thermal growth that can cause problems, growth through spindle 1 and spindle 2, and Z axis ball screw growth for your upper cross slide, and for your lower cross slide.

Dave Cross said:

A sudden jump .002-.003" is a process control issue, ie chips in the jaws or something like that. A decent warm up on machines like this make a big difference in making tight tolerance parts and maintaining it throughout the day.

When it comes to Z length on a machine of this configuration you have 4 places of thermal growth that can cause problems, growth through spindle 1 and spindle 2, and Z axis ball screw growth for your upper cross slide, and for your lower cross slide.

Click to expand...

But this would not explain the jump in size, then back to making great parts again. Thermal expansion would be constant and predictable. Is your sub-spindle collet a through collet or is it counter bored. If c'bored, could you use torque skip to ensure proper seating in the collet? Not sure if this machine has that.

LockNut said:

But this would not explain the jump in size, then back to making great parts again. Thermal expansion would be constant and predictable. Is your sub-spindle collet a through collet or is it counter bored. If c'bored, could you use torque skip to ensure proper seating in the collet? Not sure if this machine has that.

Click to expand...

I didn't suggest that the thermal growth was the issue. I suggested it was something else. A chip doesn't need to be against the part and jaw/collet to cause a problem, it can be between the chuck and jaw/collet etc or it can be something completely different, like a cutting tool issue, ie deflection.

If there's any wear or backlash in the ball screw or thrust bearings, perhaps purposely advancing faster with the second spindle, then having a rapid stop/reverse of travel by a small amount at the desired position, then clamp would "seat" everything if the cutting loads are not great.

Or if forces in the thrust direction are significant on the parted end can you clamp with the second spindle, then further advance the axis by a couple thou to seat any axis slop, then withdraw to the cut zone? It might give you a more consistent 2nd op length.

Just trying to "thought experiment" this, as I don't have experience on this style of lathe. If hooey, please ignore...

LockNut said:

But this would not explain the jump in size, then back to making great parts again. Thermal expansion would be constant and predictable. Is your sub-spindle collet a through collet or is it counter bored. If c'bored, could you use torque skip to ensure proper seating in the collet? Not sure if this machine has that.

Click to expand...

Im using a through collet. Now you have me thinking though, my pull is 200ipm and collet pressure is on the lower side. Since the parts always are on the smaller side I wonder if I am getting slight slippage as I pull, whether its collet slack or slight interference in my spindle liner. I think im going to try to slow that down the next time i run this type of a job.

I finally figured out what is/was happening with my part length inconsistency. I narrowed it down to this happening only when I was using E collets. The gauge diameter seems to different between the hardened standard Hardinge collet pads and the emergency collet pads. My collet was closing just enough to hold the part when using the E collet pads but since it was at the end of its stroke the part could shift very easily when clamped. My only solution since the master collet was as deep as it would go was to shorten the master collet on the threaded end. I took about .05 off both the main and sub masters and that seemed to do the trick. Haven't had as much as .0005" deviation since.