Schaublin 125-CNC retrofit.

it's nice for a tool change to be able to back up, especially with long boring tools. in most lathes, yeah, small lathes with small crosslides and a few tools  certainly makes sense for soft limits. 

Thanks for your experience on that Andy. I am really happy with my soft limit set up. I put so effort into getting not wasting travel but still having enough space to decelerate from rapids.

SMC.Collins. Good point about the retract. Lathe tooling offers so many different ways to crash things.
Mark

RotarySMP wrote:

SMC.Collins. Good point about the retract. Lathe tooling offers so many different ways to crash things.
Mark
 

OMG you aren't kidding.  As a (still) fairly novice turret lathe operator, I still have pants-filling moments proving out a program.

DTG and single-block are our friends here.

spumco wrote:

RotarySMP wrote:

SMC.Collins. Good point about the retract. Lathe tooling offers so many different ways to crash things.
Mark
 

OMG you aren't kidding.  As a (still) fairly novice turret lathe operator, I still have pants-filling moments proving out a program.

DTG and single-block are our friends here.

I've taken to dry running almost ALL of my programs, becuase, you know when it's going to crash, before it does it. Also the machine sim options in most cad/cam software do a great job of showing you if you're stupid. I know my cam software proves to me how stupid I am all the damn time

My problem is impatience when I have hour + tool path. Do you have a way to more quickly dry run it?

Fusion360 has machine sim, and you can set the run speed, even has collision detection

(sorrry for a long-winded hijack)

SMC is right - CAM preview (I use F360) does a nice job of crash detection for a particular tool.  You can run it as fast as you want and if you're doing single-tool tool-paths it's great. Just make sure to model the tool holder and chuck.

My problem is when I've got a turret full of tools and they all seem to want to eat the spindle, collet, part, or the back sheetmetal.  It's a pretty small lathe and there's not much room - wind up having to hang the bar stock way out some times.

I don't know of a financially viable simulator that can do crash detection for a whole turret.

If it helps...my crash-avoidance workflow is thus (for a Fanuc 21TB):

• Check each toolpath in CAM simulation. If nothing stupid, proceed...
• Estimate 'work shift' and enter an initial value (part stick-out)
  • note - 'work shift' is the Fanuc method of setting the part length sticking out of the chuck.  Like G54.  All tools are referenced from the new work shift.
• Load each tool and set tool offsets.
  • Parting/grooving tools tend to be really close to the turret, so the turret is really close to the spindle/chuck when those are in the cut
  • Which means...long tools (drills, floating tap holder) go opposite the parting or grooving tools
• Load the program and write down the smallest Z-minus number for each tool.
  • This is usually the cut-off or parting tool, but not always.
• With no part sticking out of the collet...Call up each tool so the offset is active and jog the tool to the closest point to the chuck/spindle
  • Check for clearance on ALL the tools.
  • If there's anything really close, jog the active tool down to Xnnn (smallest dia in pgm) to make sure an adjacent tool isn't going to crash through the whole tool-path range of motion.
  • Parting tools can be a pain, especially when I've got a long-a$$ tap in the turret.  My lathe control can't rigid-tap, so I'm stuck with a tension-compression holder...which you can't buy for VDI16 so it's home-made and sticks way out.
• Modify the work shift (G54 in your case) to minimize the part stickout but leave enough clearance for all tools so you don't poo yourself when the lathe is running full-boogie.
• Bonus - to save tool change time during bar-feed/pull repetitive jobs, you can jog the turret just far enough away from the part to clear the longest tool and note the G53 work coordinates.  Once the tool path is acceptable edit your program so the turret moves to that clearance position instead of the usual G28 home position.

I tend not to use dry run, as there's nothing to stop it from plowing a tool in to the back sheetmetal (other than having the feeds at 1%, of course).  On the other hand, I'm new at this and am probably doing some dumb stuff.

I find that checking CAM simulation & jogging to the closest Z-value for each tool is a sure way to check for crashes - and takes less time than running through a program.  And of course I single-block, watch DTG like a hawk, and have the override down.

PS - the Schaublin's slotted cross-slide is perfect for mounting a bar-puller.  You've got a power draw tube and could probably get 24" of stock safely in the spindle.  Automatic parts-making servant...
 

My thoughts are, it's best to limit long tools to the opposite side as much as you can. Anywhere you can use ID tools put them opposite OD tools. and try to limit the total number. Surprisingly a lot of the 2 flute insert drills, can also be used as boring bars. Which again, saves on loading lots of those kinds of long stick out tools and reduces the likely hood of collision as well.

for linux cnc, work shift is just, Touch off, so just most the carriage + to good clearance "touch off" , and check

Great inputs. Thanks guys,