CAM Simulation looks good but machined parts is getting gouging in one tool path

[chris.lN44N4]

So I did the CAM on this part and did a test run (see image 0263), it came out great. I adjusted the CAM and changed the tool height IRL and then after rerun I am getting tool gouging (see photos 75, 76). I've measured and remeasured the tool in real life, the tool length in the control (Mach3) tool table is correct within +/- .0005" which is the resolution of my measurement tooling. I even reprogrammed the tool number in Fusion and then entered the tool as new in the tool table thinking I had to have messed up the phsical tool height and that's where the issue was, after running a second part with this new tool # I got exactly the same result. 

All I can think of is that my maximum fine step down is too large on my adaptive clearing operation and as a result material is left behind. That said, I am using a horizontal operation on the flat surfaces before I run the contour with the ball end mill that is gouging. So the material heights should be good. The simulation in the screencast doesn't show the gouging I am getting in real life. With the horizontal op I moved the Top Height (stock top) down .015" as to avoid the the very top of the part so I can reference it in later operations.

I've been messing with this for a week now and am not making any progress, some fresh eyes I think will help.

[chris.lN44N4]

I’ll take any ideas I can get. I know this is a bit of shot in the dark, the fact that this occurred after adjusting the CAM should be a clue, but I’m missing something.

I might be able to roll the CAM file back to the version that worked and pick through it, I’ll have to fork the build though. The model has changed ever so slightly since then. 

[LibertyMachine]

Could you export and share your part?

File > Export > Save to local folder. Return to thread and attach the .f3d file in your reply

[chris.lN44N4]

Yes, sir.

[LibertyMachine]

What is your machine and spindle capabilities? Coolant?

[chris.lN44N4]

Grizzly G0704. 1.5 HP, 6000 RPM generally. I've run much faster and much slower F/S's. It doesn't seem to be a factor. I can overdrive or underdrive the spindle, either getting 8k or 4k rpm with obvious impacts on torque.

[chris.lN44N4]

Also, yes, coolant. I have a pretty good flood system that generates enough pressure to flush paths and pockets.

[HughesTooling]

One thing looking at your 3d contour tool path is you have the ramp set to 90 in/min and your cut is only 30 in/pm. The green lines between each level will be done at 90 in/min so might be a problem. Not really sure what your looking at from from the pictures. Also use the photo option to embed images so wee don't have to download, I generally don't bother with thread if the pictures are attached and you will get more help if they're embedded. Wish AutoDesk would hurry up and fix the forum!

Mark

[chris.lN44N4]

This is my second posting here, so I to be honest I missed the embed option all together.

If you compare to the photo below you see that the contour with the ball end doesn't dip below the horizontal surface. In the photo above it goes to a significantly lower height. So either the horizontal op is high or the contour is low.

[HughesTooling]

Checking the toolpath in NC Corrector it shows the horizontal toolpath in that area is at -1.3 and the 3d contour is also -1.3 so the problem's not in the g code. You can download NC Corrector here and check. Do the tools still match the library when you've finished cutting?

Mark

[chris.lN44N4]

Yes, the tool lengths were verified before and after run. The tool heights were within the +/- .0005". 

Since the tools are extended fairly far in their tool holders I had considered that perhaps they migrated, but that doesn't appear to be the case.  

I also check to see if I was loosing steps. Z0 before and after run was consistent. I even turned down maximum velocity on Z to avoid step loss, but no change. 

[chris.lN44N4]

I only adjust tool lengths in CAM to avoid collision. So when I see collision I adjust in CAM until the collision goes away and then I adjust IRL to get close to how i CAM'd. This is actually where the problems started. I wanted to run the contour deeper, so I extended the tool in CAM, then I extend IRL, remeasured the tool and entered into the tool table in my controller accordingly. Ever since then I have gotten the gouging. 

Also, NC Corrector is a really useful tool. I've only been using it for three minutes and I can see I'll installing it on my main work station.

[HughesTooling]

---

@chris.lN44N4 wrote:

I only adjust tool lengths in CAM to avoid collision. So when I see collision I adjust in CAM until the collision goes away and then I adjust IRL to get close to how i CAM'd. 

Doing this does not affect the g code that's produced, all the levels will be the same as it's at the tool tip. The flat where you have the problem produces a Z of -1.3, if you make the tool longer in Fusion it will still post -1.3. I don't know where the error is coming from, it sound like you've checked the obvious one's but it is not Fusion or the G code.

Mark

[chris.lN44N4 · Accepted solution]

Thanks, Mark.

When I'm back in the shop I'm going to try three things. 

1 - Calibrate my touch probe length, this might seem unrelated but my entire tool table is relative to it. 

2 - Remeasure tools 52, 39/40

3 - Check the actual tool table file stored on the controller to make sure it is updating when I try to update it through GUI

My thinking is that that the tool tool length offset isn't being updated correctly on back end.