Gib Adjustment

[wwendorf]

Yes, I will be doing that. My goal is .0002-.0003", they are sure they can do that or better. I'm going to have them actually check the repeat-ability of the ball screws before I send them off as well. I don't want to spend 5K if I only need to spend 3K.

Wade

Hi Wade - Can you specify with them that all the axis must have an agreed maximum of (eg 0.0003") backlash via a dial indicator on direction reverse checked by you or them via Skype - before you pay or return with no charge.

Keen

[flick]

And may I assume that you realize whatever backlash is left can be adjusted out with parameters in Mach? I assume the same is possible in PathPilot, if that's what you're working with...

[wwendorf]

Yes, I'm aware of that. I did enable it initially, but because the backlash was so great, it didn't do anything except exacerbate the problem.

Adjusting out via the controller doesn't help me in my case. The resulting finish is still affected with little bumps or ridges. Getting it down mechanically is my need in this case.

Wade

[flick]

Yeah, I admit the implementation of backlash compensation doesn't seem to be great in MACH, which is the only one I'm familiar with. Certainly the big machine tool builders can control the inevitable lost motion in their machines utterly seamlessly. I wonder if it's a more difficult problem than it seems, or if there just hasn't been much attention paid to solving it with the little PC based controls.

[wwendorf]

The big machines can actually adjust the backlash via the position on the leadscrew.

Our little controllers don't have that functionality --- yet...

Wade

[flick]

I'm not sure what you mean "adjust the backlash via the position on the leadscrew".

When I was playing with Mach it seemed to simply move the axis an extra (configurable) distance every time the direction reversed. The big problem I had was that playing with the "acceleration" associated with that distance resulted in a weird balance of non-linear/non-circular motion vs. a heavy "bump" every time the axis reversed. How bad the effects were also depended on the feed rate in play at the time. It seems that a smooth, effective implementation of backlash compensation would have to anticipate axis vector transitions and integrate the compensation into the transition acceleration profile. Instead it seems like the software attempts to deal with it independently of the current state of motion, and I believe that is where the ugliness comes from.

[flick]

Keen, I'm afraid that's mostly utter nonsense. Properly implemented backlash compensation on well-built machines of any size or design will improve both geometric and dimensional accuracy.

[wwendorf]

Going for a ride.

[Advil]

Wwendorf, please keep us posted on how this turns out. Controversey or no about what you are spending to get the machine trued up, you are breaking some ground most of us havn't seen done to a Tormach. I for one am very, very interested both in how the pro rebuild turns out and if they have anything to say about what the issues are that they find.

Thanks!

[CadRhino]

Take a video!

I second that ! Keen

[wwendorf]

Hi Guys, I spoke to the guys who are rebuilding my machine. They are buttoning it up today! I'm going to be picking it up tomorrow, but they want me to spend some time with them going over the machine and checking tolerances. They said that it was basically riding on 4 points as the ways and gibs were very poorly scraped. The ball screws were tested and appear to be just fine, so at least I didn't need to send them in for re-manning.

I'll post a report here when I get a chance to play with the mill after I get it back and setup.

Possibly even do a video. I didn't ask them to do any photos as I didn't want to annoy them and cost myself anymore money, but I will be asking them for an itemized list of stuff they had to do to it so I can both report here and send a copy to Tormach (who is also interested in what I am doing).

Wade

[wwendorf]

Hi Guys,

Got my mill back home and the enclosure installed back on it. I have to level it up again and am waiting on a machinst level.

The crux of the problem with the mill was that the table was basically attached by 3 points on each axis. They had to scrape the ways and the gibbs. After doing that, they also tested the ballscrews which they declared fine, so I didn't need to wind up sending them in (saved $2k there).

When I went to pick it up, they actually spent about 3 hours with me after we hooked up the control. I've got her down to .0006" on X and Y and .0004" on the Z. I'm WAY good with that. They weren't able to get any better on the backlash due to the way the table was manufactured, and more specifically, the way the ball nuts were attached. (I didn't ask details on that. It made sense so I didn't get to into the brush.)

They said that the scraping done from the factory was very minimal, definitely not a "job well done". The table runs perfectly flat now as well.

So there it is. Didn't get the .0001" they were talking about, but the construction of the table precluded that.

I'm happy and hope to be cutting some chips this weekend.

Cheers!
Wade

[CadRhino]

Thanks for the report. Let us know how she cuts.

Hi Guys,

Got my mill back home and the enclosure installed back on it. I have to level it up again and am waiting on a machinst level.

The crux of the problem with the mill was that the table was basically attached by 3 points on each axis. They had to scrape the ways and the gibbs. After doing that, they also tested the ballscrews which they declared fine, so I didn't need to wind up sending them in (saved $2k there).

When I went to pick it up, they actually spent about 3 hours with me after we hooked up the control. I've got her down to .0006" on X and Y and .0004" on the Z. I'm WAY good with that. They weren't able to get any better on the backlash due to the way the table was manufactured, and more specifically, the way the ball nuts were attached. (I didn't ask details on that. It made sense so I didn't get to into the brush.)

They said that the scraping done from the factory was very minimal, definitely not a "job well done". The table runs perfectly flat now as well.

So there it is. Didn't get the .0001" they were talking about, but the construction of the table precluded that.

I'm happy and hope to be cutting some chips this weekend.

Cheers!
Wade

[TurboStep]

I've got her down to .0006" on X and Y and .0004" on the Z. I'm WAY good with that.

Hi Wade
Good to hear you're pleased with the results! I don't recall anyone on the zone boasting less than .0004 (I know I'm not). It's interesting that your Y is now one of the best axes and that Z is also extremely good - I found Z to be the trickiest.
Step

[CountrySmith]

Wade, I'm glad it has worked out well for you.

At the Tormach Open House last week, I had a long talk with one of their technical support specialists. He had confided with me that, although they have several QC personnel at the manufacturing and assembly points, there are times when stuff slips through the cracks. Engineering changes will be made only to have manufacturing revert back to the old design. Unfortunately, when QC is checking a final assembled machine, they cannot determine if some process or assembly was done to spec. An example of a hidden defect was my discovering on the cabinet fan motor and the coolant pump motor, the wires were just twisted and taped.

I can't help wondering how many other machines have the "three point contact" problem. It might explain some of the problems with my 770.

R J

[flick]

Nice!

[wwendorf]

I'm happy to report success! I have eliiminated the tick marks at 12,3,6,and 9 oclock positions. Roundness of a 1" dia interpolated hole is .0005 to .001.

The machine cuts like butter now, better than it ever did before.

Plus side is that the tech who reworked my machine said he'd be happy to come out in 2 months and teach me to adjust gibs the way they do it.

Tonight was my first cut with it since I got it back. My day job kicked my ass and so did my weekend.

Anyway, I'm up and running again.

Wade

[kstrauss]

How did you measure the roundness of an interpolated hole? I'm familiar with the V-anvil micrometers used to determine the roundness of a rod but I've never seen a similar instrument for testing holes.

[wwendorf]

Just a digital caliper. It works to give me a good idea of where it's at.

Wade

[MichaelHenry]

How did you measure the roundness of an interpolated hole? I'm familiar with the V-anvil micrometers used to determine the roundness of a rod but I've never seen a similar instrument for testing holes.

Would that be a tri-mic? They measure a hole diameter using 3 point contacts.

Tesa 00910006 Tri-O-Bor Internal Micrometer 20-25mm 0.002mm - H Roberts