Paul Frederick Can you see how my Z axis occasionally goes lower than it

diycnc-discuss
gplus
(George Allen) #1

@Paul_Frederick Can you see how my Z axis occasionally goes lower than it should? What would cause that? Mechanical or controller issue?

(John Bump) #2

I bet mechanical. Controllers, if they drift, don’t usually drift back unless they’re servos. They drift and then stay there (or are overdriven and lose steps.) I’d suspect varying loads resulting in machine deformation. You could put a dial indicator on the spindle and do a bunch of z cuts at a fixed depth through different materials and see if you can catch it moving.

(George Allen) #3

@John_Bump Thanks John. That helps. And that sounds a little like what it did (drift). Though it did stay down to its lowest height each time I ran the job. But, the (drift) occurred at different places of the cut; indicating to me that it is likely a mechanical issue…
I did raise both the spindle rpms and the feedrate following a boneheaded mistake I made not realizing the spindle was rotating backwards. Now the cutter practically glides across pine.

(Paul Frederick) #4

I hate to guess based on too little information. But I can tell you some things you need to observe in order to suss out the root cause. Like what direction your machine is cutting in. Which way it is going in relation to how the spindle is turning does make a difference. The two are called climb and conventional milling. You are going to get a different cut doing either too. Your work could even be springing. Just bending under the pressure of the cutter passing over it. For all I know now you might be running your spindle in the wrong direction for the tool you have in it. It could be a lot of things. I know one thing for sure now. It’ll be the last thing you think of that is why. It is always the last thing.

(Paul Frederick) #5

@John_Bump the downside of open loop is you never make up lost motion. Once it is gone you’re off from there on out. Or until you home again. You can check actual measured vs calculated position. Then see if there’s a discrepancy. Then to remedy that you have to drive less hard. You’ve exceeded your machine’s capabilities. Which is something open loop should never do. That’s why the big boys run servos with encoders.

(John Bump) #6

Yeah, open loop is essentially a random drift process and the only question is how wide a drift you get. If you’re lucky and careful you don’t get any. I suppose with most machining strategies all you’ll get is less Z depth than you intended, since you’re always driving Z down into a load, never up into a load. But X and Y could go either way and slowly accumulate error. (Or quickly: I’ve had sudden accumulations of several MM of error when a chip welded itself solidly to a tool and the tool stopped cutting but kept trying to move.)

(George Allen) #7

@Paul_Frederick Mostly it’s a climb direction (not Conventional). However, these 3D cuts do parallel cuts and go both ways. Since I changed the rotation of the spindle, today was the first time I’ve had this issue. I used the same style cut earlier this week and it didn’t happen. Your comment about bending under load is an interesting one, and I’ll check that because it’s possible, as I’ve wondered about that at times. One note about the current issue: I had just been trying some cuts on aluminum just prior to this cut and, in order to limit any likelihood of that bending, I positioned the vise just above the supporting beam under the table. Unfortunately, for the aluminum cut, I still didn’t quite have a fully sufficient work holding strategy which ruined its cut. But, I don’t know if switching from wood to aluminum for one cut had any effect on the machine.

(George Allen) #8

@John_Bump So, the fact that I machined some aluminum before this cut (thus the greater load) may have caused the current (drifting) you are talking about?

(John Bump) #9

No. Lost steps are from an overload during a cut, when the stepper just can’t move the bit through the workpiece. Once you rehome, for the next job (or even within a job) you get your position back. Some people rehome after every toolchange (as part of assigning the tool length and making sure it’s in the collet right) and some people even rehome after every heavy cut.

(George Allen) #10

@John_Bump Got ya. After you mentioned drifting, I did a little research and I think it may be the result of a loosened pulley on my z axis. Obviously, I can’t be sure, but that makes some sense as I’ve been having issues with that pulley. The fact I put such a heavy load on it may have loosened it up.

(Paul Frederick) #11

@George_Allen I have a milling machine so you don’t have to tell me about inadequate work holding. Been there done that. Had to buy new tools too. Because usually when work slips it catches in the tool and ends up breaking it. Something always has to give. Then I have to pay.

(Paul Frederick) #12

@George_Allen climb cutting is the more problematic of the two if there are any rigidity issues. When climb works its great. When climb walks on you then there’s problems. For a while the X axis nut in my mill was loose and sometimes it’d skip on me climb cutting. I never liked that whenever it happened. I finally tightened that nut up and it’s gotten better. But it’s also gotten me in the habit of conventional milling more than climb milling. I’d rather crank slower and harder than skip along.

(George Allen) #13

@Paul_Frederick I just like climb milling because it’ has always been smoother.

(John Bump) #14

My suggestion is conventional mill everything out as a roughing pass, leaving some very small finishing depth, then climb-mill the whole works in a single pass. That way you get the smooth finish, but your depth of cut isn’t high so it reduces the strain on the machine, reduces tool deflection, and generally gives a better finish than a heavy climb mill cut (and way better than conventional.)
I don’t notice as much difference in wood, because the wood doesn’t weld to itself if there are chips, during conventional milling.

(Nathaniel Stenzel) #15

“Since I changed the rotation of the spindle”? Did you reverse it or just change the speed or did the angle of it change?

(George Allen) #16

@NathanielStenzel The 2 hot wires were reversed when I re wired the 220v outlet for my VFD. I didn’t notice that the spindle was rotating CCW instead of CW.

(Nathaniel Stenzel) #17

@George_Allen ah. So now it is rotating the right way. Did the bit get damaged any from rotating the wrong way?

(George Allen) #18

@NathanielStenzel Good question! Surprisingly, none broke as the cause of it. I think because it was rotating the wrong way, the collet chuck kept loosening and before they would snap I would notice something wrong and stop the machine. As for damage, I don’t know. One thing I can say, that bit chews up pine and the cuts are as smooth as glass. Watching it yesterday, was like watching a hot knife through butter. With it cutting now, I can raise the feed & speed and it cuts like a beast! I still need to get a proper bushing or coupler on one of the pulleys on my z axis, but thankfully, most of the adjustments seem sufficient and complete. I may also need to upgrade my drivers on my y axis, but I’m pretty satisfied. Although, my x axis motor got pretty hot last night when I was pushing it like I was. It was a lot to ask from an average Nema 23 motor, but it did the job and didn’t miss a step.

(Nathaniel Stenzel) #19

Is your Z a screw drive? Does it move well when manually moving without apparent snags? Did you accidentally loosen one of your 4 Z stepper motor wires while rewiring the spindle? If the Z is a screw drive, is it zinc plated or iron and is there rust on it?

(Paul Frederick) #20

@George_Allen reversing single phase alternating current leads changing the direction a motor spins is pretty weird. Normally that is not the case. So what are you doing that’s so abnormal? A collet would have to be pretty loose for spin to loosen it more. But a tool spinning the wrong way is going to cut poorly enough to cause issues. There’s going to be added pressures then as material is not removed properly.