Calibration question

[David Falkner]

Until my cheapo HF magnetic base broke I got about 0.0015" backlash on Y and 0.001" on X. I used MDI and went in 0.10" increments back to zero and then to 0.750" in one large step and back to zero. I saw 0.003" one time but the magnetic base was starting to slip right before it broke. The first few readings seemed repeatable. Nate is sending me some new spindles and belts so I'll get those on soon and test it again. In the meantime I need to either buy another magnetic base or go into DIY mode, likely the latter.

David

[Gary Campbell]

Mark...
Once the indicator is preloaded and in place, I like to give a "push/pull" on the gantry or XZ car to see what happens with some force on it. I am looking for ~25# of force, which is at or even less than can be generated with a 1/4" bit

[Mark Bolton]

Until my cheapo HF magnetic base broke I got about 0.0015" backlash on Y and 0.001" on X. I used MDI and went in 0.10" increments back to zero and then to 0.750" in one large step and back to zero. I saw 0.003" one time but the magnetic base was starting to slip right before it broke. The first few readings seemed repeatable. Nate is sending me some new spindles and belts so I'll get those on soon and test it again. In the meantime I need to either buy another magnetic base or go into DIY mode, likely the latter.

David

Interesting. Be a real brain teaser to see how those correlate to the cut numbers your seeing.

[Brian Lamb]

As for mag bases, Noga makes very good and useful ones, a bit pricey but if you use one much, well worth it.

[Mark Bolton]

Mark...
Once the indicator is preloaded and in place, I like to give a "push/pull" on the gantry or XZ car to see what happens with some force on it. I am looking for ~25# of force, which is at or even less than can be generated with a 1/4" bit

Had it still sitting next to the machine. I can only guess at the #25 but Im a big guy so perhaps more than I thought. But I could see about .001 to .0015 shoving up as high as I could on the gantry (most leverage) in Y.

Hopefully David gets some resolution. While its great to be as spot on as possible thankfully the bulk of our work never has me looking at this stuff but its definitely interesting.

[Bill George]

Its a router with rack and pinion drive not ball screws as a CNC mill would have. What are you working on that requires .001 +/- accuracy ?

[Mark Bolton]

Its a router with rack and pinion drive not ball screws as a CNC mill would have. What are you working on that requires .0001 +/- accuracy ?

I didnt get the impression he was looking for one tenth accuracy but perhaps a couple thou?

[Bill George]

Yeah, I had too many zeros.... my bad. Maybe his router does have ball screws. I was going to website to check when I saw this. From what I could tell he does have ball screws.

[Mark Bolton]

Yeah, I had too many zeros.... my bad. Maybe his router does have ball screws. I was going to website to check when I saw this. From what I could tell he does have ball screws.

No, he did say in his reply #6 that he's rack and pinion. Seems the issue in the original was more of a discrepancy between some inside and outside cuts and about a 15-25 thou tolerance issue he was trying to chase. Was a pretty intensive build to watch as he went along with it. It would seem if anyone will lick it, he will. The build was slick all the way through.

[Bill George]

So is it repeatable, so every time you cut a size its the same? Or is it random? David did a much better and more detailed build than I did. I wish I had his patience.

[David Falkner]

Its a router with rack and pinion drive not ball screws as a CNC mill would have. What are you working on that requires .001 +/- accuracy ?

Things like this, Bill -

Plaque I did in Walnut and Maple a couple of weeks ago. I now have another one to do and this one took 2-3 hours of hand work to get the inserts to fit properly. There's no money in it if I have to do it by hand anyway. Or they won't pay enough for the extra hours so my expectation is that I cut a cavity/pocket and then an inlay/insert and it fits with minimal tweaking. I'll be doing several of these and have another to do today. I also built the guitar and want to do some inlay work on the next ones and for other people. These have to fit precisely.
037 - Command Chief Stripes, size comparison.jpg

This is the bridge I cut for the guitar above, at least I cut the back curve, saddle slots, and bridge pin holes. The saddle slots were supposed to be 0.126" but they came out about 0.122". Closer would have been better.
028 - Build #1, bridge.jpg

This is an inlay I did entirely by hand in 1985 and still have the guitar. I'd like to do these and others like this on the CNC. Again, the pieces have to fit because the expectation is very high for tight fitting Abalone and pearl.
'85 Headstock torch inlay.jpg

If I was just carving names in signs then the machine is plenty accurate enough. But I rarely do that. Many of the pieces I cut are more toward the engineering side where slots, pockets, steps, etc. are expected to be the size specified, or at least much closer than I'm getting now even though I work primarily in Walnut, Maple, and the exotics for guitars. When I cut a neck block with mortise for a guitar neck the related tenon in the neck needs to fit. I have several people wanting electric guitar bodies and necks with radiused fingerboards and fret slots; those have to be very exact or the guitar won't play in tune.

Anyway, that's the sort of thing I'm doing with the CNC that requires a bit more accuracy than ±0.025". I'd probably make more money and have more jobs if I did signs full time, though! LOL!

David

[David Falkner]

So is it repeatable, so every time you cut a size its the same? Or is it random? David did a much better and more detailed build than I did. I wish I had his patience.

Thanks, Bill and everyone for helping with this!

Yes sir, it is repeatable and that's what makes me think it's something I'm doing or not doing in Fusion 360. Yesterday I invested some time in recalibrating all axes. Z and X were spot on but for some reason Y was out 1/16" in 44". So I corrected that and ran the files again with conventional cuts for final and got everything far, far closer. The inside cuts are still out 0.007" to 0.009" but that's close enough for me to do a negative value on 'Stock to leave'. I didn't want to compensate 0.025", though.

David

[Bill George]

I am just curious in case I need to do it, how did you adjust the Y?

I see by the video you posted your putting grease or lube of some kind on the racks. I am debating on weather to do the same. One train of thought the grease or lube might get full of sawdust or shavings? I do have some spray on dry lube, I might use that.

[Richard Gonzalez]

Yesterday I invested some time in recalibrating all axes.
David

One more thing to check while you are calibrating is checking if gantry is square. Do the biggest 3-4-5 triangle you can fit. I just use a v-bit and put a small dimple on the spoilboard in 3 places and measure the diagonal side. This one is often forgotten.

[David Falkner]

I am just curious in case I need to do it, how did you adjust the Y?

I see by the video you posted you're putting grease or lube of some kind on the racks. I am debating on whether to do the same. One train of thought the grease or lube might get full of sawdust or shavings? I do have some spray on dry lube, I might use that.

For both X and Y, Bill, I used a tape on the spoilboard and a 60° V bit for a sharp point. I set the tape where I could start on the 2" mark and then used a command in the MDI window to move the bit 44" at 150 ipm. I moved the bit down to where it was almost touching the tape to remove any parallax factor and set the bit directly on the 2" mark, then raised the bit before using the command to move 44" and then lowered the bit again to see if it was on 46". The first time I tested it came up 1/16" short. I made some changes and ran it again 3-4 more times until I was satisfied it was as close and I could get it with the measurement tools I have. Btw, I have about 6 tapes and I put them all side by side to see if there was any discrepancy and to my surprise they all matched up. They were probably laid out on a CNC machine that's been calibrated - LOL!

Here's the Y setup except that when I did the test I taped one edge of the tape down to the spoilboard which rocked the other side up, making it easier to line up the bit with the 2" and 46" marks. I did the same setup for X.
Y axis calibration setup.jpg

For Z I clamped a Starrett 24" rule to a block so it would stand vertical and set a pencil on top of the spindle mount. The pencil has been cut away to half its diameter so it make a good pointer (I use the pencil to lay on top of guitar frets to mark the curvature of the fingerboard onto the nut and saddle so it was already on hand). I tested 8" of travel even though I have more available, figured that was enough. Same with Y; I have 50" of travel available but only tested for 44".
Z axis calibration setup.jpg

When I was setting the machine up the first time I used dry lube on the R&P and on the rails. But it just didn't 'feel' right to me. The recommended grease is NLGI#2 and I'm using Royal Purple for that. It's messier, for sure, but also very reassuring that there are no spots without lube. I have yet to notice any issue with sawdust or chips.

One more thing to check while you are calibrating is checking if gantry is square. Do the biggest 3-4-5 triangle you can fit. I just use a v-bit and put a small dimple on the spoilboard in 3 places and measure the diagonal side. This one is often forgotten.

It seems like I did something similar when I set everything up as I was building it, Richard, but haven't done this triangle test. I may do that tomorrow if I can work it in.

Thanks, Guys!
David