Surface finish banding

[homemachinist]

Hello all,

I'm fairly new to machining and am in the process of converting my lathe (Optimum TU2506 clone) to CNC. I notice banding at regular intervals (seems to match the Z leadscrew pitch of 3mm) as if the tool is being pushed towards the stock but only when both X and Z axis are in motion at the same time (like turning tapers). Simple turning operations where Z and X are moved separately but not at the same time comes out perfectly fine as seen in the picture.

I initially suspected a bent leadscrew but if that's the case expect the banding artefacts to appear in the straight turning operations as well. I've tightened the saddle gibs at the back (there's none at the front).

Any thoughts or suggestions ?

Thanks

 

[ddmckee54]

Simple minded question, because it's always the simple things that bite you in the butt. Is your power supply big enough to handle the load from both motors at the same time? A marginal power supply might be able to handle one motor just fine but significantly drop the voltage when the load of both motors hits it.

Don

 

[homemachinist]

Thanks Don, that's a good point. I did measure the peak use independently, but will test if the voltage drops during the taper turning cycle. It's an odd coincidence if that happens every 3mm matching the lead screw pitch.

 

[RM-MN]

Does your leadscrew turn easily or is there a tight spot?

 

[homemachinist]

I pulled off the lead screw covers and took a look a the lead screw as it rotates. I think there's a slight bend, or incorrectly rolled leadscrew which is barely noticeable - relative position of the thread root seems to move in (towards bed) and out while the crest seems stationary. Weird, I'm not an expert I would've expected to see that result in artefacts in the turning operation along Z.

No tight spots, I disengaged the stepper on Z and the transmission gears. It rotates pretty freely. Shear pin looks fine as well, on the outside at least.

I'm sourcing a 2nd 24V 5A power supply to try the taper turning as Don suggested. I'll see how that goes and post an update. Thanks for the pointers.

 

[homemachinist]

> I'm sourcing a 2nd 24V 5A power supply to try the taper turning as Don suggested.

Did that and to be extra certain also changed the stepper driver and motor. Same results, clean finish on the straight Z movement for turning but stepped / banding finish with taper turning. It's most likely something mechanical, but I can't explain the discrepancy. It's driving me nuts.

Next is to put a dial indicator somewhere and see if I can see a pause or jitter in the X axis movement

 

[Charles Lamont]

From here, the finish looks very good anyway. Is it simply a question of which axis takes the 'first' step? That might well vary cyclically at low frequency.

 

[homemachinist]

Is it simply a question of which axis takes the 'first' step?

Good point Charles. The algorithm for linear motion is pretty similar to Brenseham's where it's a sequence of steps in the dominant axis followed by steps in the other, this repeats until a line for the taper is traced. I'm not sure it's related to that though since the individual steps are very small (0.005mm), unless there are skipped steps in the X direction.

I'm going to do some measurements and pull the leadscrew out to double check. I want to use this lathe for manual work and CNC, otherwise I'd replace the leadscrew with a ballscrew.

Cheers

 

[ddmckee54]

deepfryd:

Maybe it's a combination of problems with both X and Z? You've observed questionble behavior on the Z leadscrew, what if that was combining with a small problem on the X axis? It could be that neither issue by itself is enough to show up as an artifact. But the 2 problems together is enough to give you the artifacts?

Skipped steps would show up as not being able to repeat a measured position, so that would be easy to check.

Don

 

[ajoeiam]

Hello all,

I'm fairly new to machining and am in the process of converting my lathe (Optimum TU2506 clone) to CNC. I notice banding at regular intervals (seems to match the Z leadscrew pitch of 3mm) as if the tool is being pushed towards the stock but only when both X and Z axis are in motion at the same time (like turning tapers). Simple turning operations where Z and X are moved separately but not at the same time comes out perfectly fine as seen in the picture.

I initially suspected a bent leadscrew but if that's the case expect the banding artefacts to appear in the straight turning operations as well. I've tightened the saddle gibs at the back (there's none at the front).

Any thoughts or suggestions ?

Thanks

Hmmmmmmmmmmm - - - - taper turning (which I've only done manually) is fun at its best!

You are using only only the longtitudinal and cross slide for this - - - yes?

Your lead screw - - - its likely an acme thread - - - - wondering - - - - is that thread form accurate?
(Read through the previous on the thread again and you mention that there is some movement, in out' on your leadscrew. This would very definitely be a factor imo. )
If your steppers are 'pulsing' correctly you shouldn't have what you got.
Maybe you are getting some kind of 'resonance' in your system.
(That's where at a specific rate of pulses something moves or shifts in some kind of regular pattern.)
What are your electronic specs? (x pulses per rotation is one that I'm thinking of)

 

[Chiptosser]

The reason Ground Ball screws are used is for precision, rolled acme thread is not.

 

[Ken I]

I had that problem on a previous lathe and never got to the bottom of it.

Have a look at the following old thread on this exact problem - some more suggestions and tips there.

Banding Finish

The link will take you to page 2 of the thread and my photo is in post #35

Regards, Ken

 

[Steamchick]

Hi Ken 1: Having read quite a lot of the linked thread, I recall that I once had the change wheels "too tightly set" - seemingly to minimise play when thread cutting. I did get a lot of noise from the change wheels, and found I got better results when I backed off "the tinniest amount" so they stopped grinding my brain to insanity!
Cheap "imperfect" change wheels are a result of "quality" versus money...
K2

 

[Guzphile]

Hello deepfryed, I recognize these patterns. See pic attached, the uppermost plate was faced on my lathe and has the same type of markings, coincident with the cross slide leadscrew pitch. The problem was a bent leadscrew and looser-than-they-should-be gibs on the cross slide. So I guess runout of the leadscrew is able to push the cross slide around if the leadscrew is not true and the gibs properly snug (I bent the leadscrew by crashing the lathe cross slide while under power feed - oops!) I managed to straighten the leadscrew and of course tightened the gibs, so good results now. If you are using the compound to produce the taper, it might be the compound leadscrew/gibs. Good luck.

 

[johwen]

Hello all,

I'm fairly new to machining and am in the process of converting my lathe (Optimum TU2506 clone) to CNC. I notice banding at regular intervals (seems to match the Z leadscrew pitch of 3mm) as if the tool is being pushed towards the stock but only when both X and Z axis are in motion at the same time (like turning tapers). Simple turning operations where Z and X are moved separately but not at the same time comes out perfectly fine as seen in the picture.

I initially suspected a bent leadscrew but if that's the case expect the banding artefacts to appear in the straight turning operations as well. I've tightened the saddle gibs at the back (there's none at the front).

Any thoughts or suggestions ?

Thanks

One of the things we found on Chinese lathes with belt drive was the belts had hard spots in the rubber and this caused banding. Changing the belt fixed the problem. Otherwise the other mentioned areas covered any one of which can be the culprit. Best to have a slight drag on the slides. Just tighten the gibs to give this drag. Cheers

 

[homemachinist]

Lots of nice pointers, thanks heaps everyone. That's an especially helpful thread @Ken I

Updates:

1. It is an ACME / Metric leadscrew. The thread form is fairly decent, but I suspect it's rolled not precisely ground (not expecting more than 0.01mm accuracy with this). It's good enough for hobby / home machine shop I suspect.
2. The thread crest / root positions do vary as the screw rotates which is why I suspected a bent lead screw, not by a lot but perhaps just enough to cause issues.
3. I'm testing this with a custom control unit I'm building before I wire everything up to Linux CNC or Masso. I'm able to dump the pulse count per axis in real time and the pulses are being generated as expected without any jitter. That mostly rules out software issues.
4. I placed a dial indicator on Z and X axes to check the total travel and they end up spot on target +/1 0.01mm. So ruling out stepper skipping steps, the Z axis uses closed loop while X uses open loop - I did consider missed steps first but in that case the target measurements with dial indicator would have been off.
5. The spindle is belt driven with relatively new v-belt. I don't feel any hard spots as the spindle rotates while driven by the belt and not sure if a the nearly perfect finish with straight turning would be possible with such a belt. The spindle has tapered roller bearings with reasonably decent preload. I did pop a dial indicator against a ground gauge pin held with a collet chuck and there was no jitter at low rpm.

To Investigate:

1. Cross slide screw pitch is 2mm, while the banding spacing is 3mm but my eyes could be deceiving me. I'll try to take the cross slide part, check the leadscrew there and put it back with a more snug gib. I'll be damned if it's the cross slide!
2. Going to snug up the saddle gibs one more time, replace the carriage lock with brass bushing / plate to act as a replacement gib on the front of the saddle.
3. I've also ordered replacement leadscrew, but that's 3-4 months away. Germany -> Australia via container freight :S

Again, thanks every one. All good comments and pointers.

Cheers

 

[homemachinist]

The leadscrew acts as a plane/wedge and raises the nut(s) open slightly as the carriage encounters resistance. Then, the halfnuts close up tight until the leadscrew opens them again. This causes the carriage to move with a jerky, but consistant pattern. The pattern translates into the bands as the carriage jumps forward.

This is a quite promising lead, it could possibly explain the banding if Z move is a bit jerky. It might be possible to measure this as pointed out in the post.

 

[homemachinist]

A few updates (tl;dr still an issue)

1. The leadscrew is bent, 0.2mm radial runout at the middle, which would be pushing the saddle up and down, front and back by a fraction of that.
2. It's hard to tell on straight turning of small stock, but more obvious with steep taper turning because of the errors compounding as the X axis moves - this is my best guess.
3. I've tried loosening the saddle screws holding the apron and cross slide, and the tail end of the leadscrew bearing clamp to allow leadscrew movements in X axis to be absorbed but no luck with that

I'm going to suck it up for now until I get a replacement leadscrew early next year and see how that goes. It's pretty annoying, but I've run out of ideas.

Thanks for the pointers.

Cheers

 

[djswain1]

A few updates (tl;dr still an issue)

1. The leadscrew is bent, 0.2mm radial runout at the middle, which would be pushing the saddle up and down, front and back by a fraction of that.
2. It's hard to tell on straight turning of small stock, but more obvious with steep taper turning because of the errors compounding as the X axis moves - this is my best guess.
3. I've tried loosening the saddle screws holding the apron and cross slide, and the tail end of the leadscrew bearing clamp to allow leadscrew movements in X axis to be absorbed but no luck with that

I'm going to suck it up for now until I get a replacement leadscrew early next year and see how that goes. It's pretty annoying, but I've run out of ideas.

Thanks for the pointers.

Cheers

If there is a marine propeller shaft manufacturer nearby they might be able straighten it for you as they will be used to straightening prop shafts. Some marine engineering workshops or other engineering workshops may also be able to do it for you.
Cheers,
Dave.

 

[Ken I]

Straightening long slender shafts is a tricky business - I spent years in the shockabsorber business as well as using the rod straightening facilities for straightening (lightly) bent motorcycle fork legs bent in racing mishaps. (don't try fixing badly bent fork legs.)
Using a hydraulic bench or arbor press can only get you so far - you reach a point where it will flex and accomplish nothing and then going just a couple of thou further will bend it the other way - often worse than it was.
Ultimately you have to hit it with a soft hammer or lead thumper - use aluminum supports or covers to strike points if using a steel hammer.
Start over the full length - then half and half - go back to full length - repeat until you get 0.001" TIR
Easy if you have the skill - practice on a similar diameter round barstock first just to get a feel for how hard you need to smite it.

Regards, Ken