Strange axis feeding issue

[wmpy]

I am having a strange issue with the feeding of the Z axis on a lathe. Maybe one of you has an idea what this could be. Below is a picture of a threaded part I made on this machine. It's a 5/16-18 thread. If you look at the crests of the thread, you can see how they alternate between a skinny one and a fat one. The part goes between gages, but measuring with a pitch mic, the pitch is all over the place- up and down a few thou over the length of the thread. I'm suspecting that the feed rate of the Z axis is not constant, but rather oscillating, which causes the funky thread. I don't know how that would be possible. Could it be a servo tuning issue? The axis doesn't sound great when rapiding, but it's never been an issue. It just has a little bit of a vibration to it. I haven't completely ruled out a mechanical issue with the axis, or even a spindle issue, but those seem less likely to me.

The machine is an old Star KNC-20 Swiss style lathe with a Fanuc 0T control. This is on the sub-spindle which I mostly use for spotting, drilling, and rotary broaching. I haven't noticed any issues with those operations.

 

[Donkey Hotey]

What is the material? Have you tried running a few parts in softer material (aluminum) to see if it still happens? I'd be just as inclined to wonder if the bars have hard spots or poor distribution of alloying materials that are causing the material to walk around in X as you're threading.

 

[guythatbrews]

Be sure the tool has enough clearance to cut the somewhat higher helix angle. They won't all do it.

 

[wmpy]

What is the material? Have you tried running a few parts in softer material (aluminum) to see if it still happens? I'd be just as inclined to wonder if the bars have hard spots or poor distribution of alloying materials that are causing the material to walk around in X as you're threading.

It's annealed 416. The other end of the part has a 1/4"-20 thread put on it by the main spindle with no problems. I can try some softer material, but I don't expect it to be any different. I remember having this same problem the last time I ran this job in this machine. It's a crude part, so I just lived with it, but I won't always be able to accept parts like this.

 

[wmpy]

Be sure the tool has enough clearance to cut the somewhat higher helix angle. They won't all do it.

So, it's the sub side of a Swiss, so my tool holding is limited. I'm using a left-handed internal threading bar with an external 16ER18UN insert. The bar has a fixed helix angle of 1.5 degrees, which is not ideal. But really, at that small diameter, I don't think it has a lot of effect. The thread "falls away" from the tool radially before the flanks of the insert have any chance to rub.

 

[Northman Ind.]

Growls in rapid, could be a dying encoder, or perhaps loose connection (wiring mumbo jumbo), or the ball screw is out of adjustment/bad bearings?
I'm not a swiss guy, but if you can jog the sub (thats a thing right?) set an indicator up touch off to zero, then back it off, see if indicator reads what you backed it off, do that in a couple of spots along the travel, should tell you if the ball screw is out of adjustment, you'll need a .0001" indicator to be accurate, now I only now how to test this, not how to fix it... but if the needle bounces a bunch thats a pretty good sign too (fixing involves a lot of parameter changing assuming the ball screw and encoder are still good)

 

[chip_maker]

Looks like maybe an encoder issue. How is the encoder set up on this machine? Is it belt driven or hall affect sensors?

I had a similar problem on one of my lathes that uses hall affect sensors. There was some congealed coolant or something on the rotating part. I cleaned it up good and that fixed the problem.

 

[guythatbrews]

So, it's the sub side of a Swiss, so my tool holding is limited. I'm using a left-handed internal threading bar with an external 16ER18UN insert. The bar has a fixed helix angle of 1.5 degrees, which is not ideal. But really, at that small diameter, I don't think it has a lot of effect. The thread "falls away" from the tool radially before the flanks of the insert have any chance to rub.

The thread helix angle is about 3.75*. I think the tool is your problem. I use a bar that has different angle shim seats so you can get close. For 5/16-18 it's a 4* shim.

If you don't have trouble with other thread of slower helix angle this is something to look into. And it's the easiest thing to look into.

 

[wmpy]

Looks like maybe an encoder issue. How is the encoder set up on this machine? Is it belt driven or hall affect sensors?

I had a similar problem on one of my lathes that uses hall affect sensors. There was some congealed coolant or something on the rotating part. I cleaned it up good and that fixed the problem.

There is a Fanuc red cap servo motor that drives the ball screw through a timing belt. In addition to the encoder on the servo, there is another encoder on the end of the ball screw. I don't know how this works with two encoders. Does the machine only look at the one on the screw? See the picture below for reference. The pulley on the bottom left is attached to the servo motor. The pulse coder on the top right is mounted on the end of the ball screw.

 

[chip_maker]

There is a Fanuc red cap servo motor that drives the ball screw through a timing belt. In addition to the encoder on the servo, there is another encoder on the end of the ball screw. I don't know how this works with two encoders. Does the machine only look at the one on the screw? See the picture below for reference. The pulley on the bottom left is attached to the servo motor. The pulse coder on the top right is mounted on the end of the ball screw.

View attachment 433042

I'm no Fanuc expert by any means, but I think the 2 encoders "compare notes" and will throw an alarm if they're on different pages. How's the other pulley and belt look?

I also don't know if removing the encoder will disturb anything? it looks like it can be clocked?

 

[Mtndew]

How fast are you spinning the part?
How far away from the end of the part are you starting?
I only ask is on some machines if you start too close and/or feed too fast, it can't position itself correctly in time.

Something to think about before messing with encoders.

 

[wmpy]

How fast are you spinning the part?
How far away from the end of the part are you starting?
I only ask is on some machines if you start too close and/or feed too fast, it can't position itself correctly in time.

Something to think about before messing with encoders.

2100 rpm
That's about 117 ipm, which shouldn't be too fast for threading.

I'm starting .250" away from the face of the part, which is my default for anything 16 TPI or finer.

 

[Vancbiker]

There is a Fanuc red cap servo motor that drives the ball screw through a timing belt. In addition to the encoder on the servo, there is another encoder on the end of the ball screw. I don't know how this works with two encoders. Does the machine only look at the one on the screw?…….

In a dual encoder system the velocity loop is controlled by the encoder on the motor and the position loop is controlled by the encoder on the screw. This is the same when linear scales are applied to an axis.

 

[wmpy]

In a dual encoder system the velocity loop is controlled by the encoder on the motor and the position loop is controlled by the encoder on the screw. This is the same when linear scales are applied to an axis.

So in a single-point threading situation, which one is being used? Both? Seeing the issue I'm having, which would you suspect to be the culprit? I have another machine of the same model that I might be able to swap parts between.

 

[Northman Ind.]

2100 rpm
That's about 117 ipm, which shouldn't be too fast for threading.

I'm starting .250" away from the face of the part, which is my default for anything 16 TPI or finer.

its easy enough to try running it slower.
Again I'm not a swiss guy, but never had much luck running any thread over 1500, machines would rattle and shake, probably fine, but I didn't like it.

 

[Vancbiker]

So in a single-point threading situation, which one is being used? Both? Seeing the issue I'm having, which would you suspect to be the culprit? I have another machine of the same model that I might be able to swap parts between.

Both. The velocity and position of the axis are critical for a proper thread. The spindle performance also is critical. Typically there is an encoder driven by the spindle that provides an index pulse (one revolution pulse) for starting each thread pass as well as a pulse train that the spindle drive uses to control the spindle speed.

On a machine like this it's easy to view encoder signals with an oscilloscope. Personally, that's what I would do to begin troubleshooting the control system. I'd only do that after confirming that there were no tooling or mechanical issues though. Belt driven dual encoder feed back systems can be troublesome if there is lost motion (loose or bad belt, worn pulleys) in the belt or pulleys. If the velocity loop and position loop are not pretty closely in sync, the control and drive can get into an oscillation.

 

[couch]

Any chance your sub collet isn’t tight enough and the part is slipping?

 

[sinha]

If the velocity loop and position loop are not pretty closely in sync, the control and drive can get into an oscillation

The error in Z feed appears to be periodic.

 

[Mtndew]

2100 rpm
That's about 117 ipm, which shouldn't be too fast for threading.

I'm starting .250" away from the face of the part, which is my default for anything 16 TPI or finer.

Those numbers look fine to me.
Do you have a dummy piece where you can run it about 1000 rpm to see if there is a change?

 

[Ox]

I was gunna say to check collet pressure as well, but then I went back and looked at the pic for a while aggin, and - nah - not collet.

By the same regards, I'm gunna rule out thrusts as well.

I would say that your ball nut could possibly be a thing?
Maybe you have a nut 1/2 full of bad balls, but I would think that you'd see other issues as well.

I'd start by slowing it down.
And then I start swapping encoders.

But since this thread is a few days old, I'm guessing that you've done something by now?


-----------------------

I am Ox and I approve this here post!