Surface Finish Issues on New Machine

[PegroProX440]

I know its obvious but double check the level of the machine and make sure all feet (depending on manufacturers spec) are supporting the appropriate amount of weight.

 

[Rcgiovannani]

I know its obvious but double check the level of the machine and make sure all feet (depending on manufacturers spec) are supporting the appropriate amount of weight.

They did re-level the machine. It was out by a lot and the surface finish did improve when they did that.

They keep getting it a little better, but it's still not 100%.

 

[Omega]

It almost looks like the head is out of tram, you said 3d surfacing looks good, which would make sense if using a ballmill, maybe try the pocketing with a small flat endmill to see if any difference in the scallop.

 

[rklopp]

They did re-level the machine. It was out by a lot and the surface finish did improve when they did that.
They keep getting it a little better, but it's still not 100%.

Was the machine out of level from the initial installation, or because the floor moved afterwards?

 

[pcasanova]

did they run a ball bar test?

 

[michiganbuck]

I wonder if one could set the machine on solid Pads perhaps 24 x 24" then level and balance each foot and make it better/Ok.

I thought Tram error or the spindle head is not affixed properly to the machine. If it was a bad ball screw then one feed axis might be good and one bad.

Guess one should treat it like an old used machine and check out everything from gibs to spilled beer.

It is really a crappy finish for a new machine, poor for an old clunker.

 

[guythatbrews]

If you are still having trouble maybe there is a way to rule out the floor once and for all.

If you core drill a hole in the floor near the machine you can inspect the sub grade and detect if you have a void. Maybe drive a post in and put a DTI on it. Epoxy a little target to the floor and zero the indicator. Run the part and watch the indicator. You can check vertical and horizontal movement.

If there is a void I'll bet one of those driveway guys can foam it.

 

[pcasanova]

also I would check ALL the electrical connections, drives, contactors, plc ect... My factory tech (although lathes with Siemens) on install goes around and tightens everything after an install in the cabinet. He's had issues on new machine where something strange was happening and after re-tightening everything cleared up the problem. and does the machine have scales?

 

[Rcgiovannani]

Was the machine out of level from the initial installation, or because the floor moved afterwards?

Honestly, I'm not sure.

If you are still having trouble maybe there is a way to rule out the floor once and for all.

If you core drill a hole in the floor near the machine you can inspect the sub grade and detect if you have a void. Maybe drive a post in and put a DTI on it. Epoxy a little target to the floor and zero the indicator. Run the part and watch the indicator. You can check vertical and horizontal movement.

If there is a void I'll bet one of those driveway guys can foam it.

I have a friend dropping off a hammer drill tomorrow morning so I can drill a few holes and check things out.

also I would check ALL the electrical connections, drives, contactors, plc ect... My factory tech (although lathes with Siemens) on install goes around and tightens everything after an install in the cabinet. He's had issues on new machine where something strange was happening and after re-tightening everything cleared up the problem. and does the machine have scales?

The machine doesn't have scales. They ran trace programs on every axis so I think that might have shown up for them.

 

[EndlessWaltz]

If you slow it down.....and it does this effect AFTER it has already done a direction change and is still traveling diagonally in both X an Y it is the servo tuning. If you overshoot on direction change and get chatter that is coupling, ball nut, or a parameter typically. If you are getting this while traveling fast and doing hog cuts then I could maybe blame the floor or rigidity in the table/fixturing.

This effect as someone else pointed could be the head out of tram perhaps?

My issues were with a Siemens control.

 

[Rcgiovannani]

So I did some more testing this weekend. I drilled a hole in my floor and the concrete was about 5in thick. There's soil underneath too so I don't believe there is a void.

A few weeks ago when I first starting working on this issue, I checked the tram of the machine and it was out by 0.0015" from front to back. At first I thought this was definitely the issue so the MTB came back out to re-level the machine. So I just trusted that they did it and didn't check it. My mistake.

I checked the tram this weekend and it was still out 0.0015". I checked the level of the machine with a Starrett Precision level and it was good. I remembered that my vices weren't removed from my machine when the MTB did it. I assume they leveled the machine and didn't check the tram... None of the jam nuts on the feet were tight either.

I couldn't do this next test very scientifically but I decided to try and lift up on the head of the machine while the indicator was on the table. I could move the head almost .001" with one hand. With two hands I could move it multiple thou. I did the one hand test on my Fryer and it only moved 0.0003". One hand on a Haas and it only moved 0.0004". Got the head to move about a thou on both the Haas and the Fryer with two hands.

I'm tired of this back and forth with the MTB. I feel as though these things should have been checked but instead they are looking very hard to find something wrong other than the machine.

Can machine tools be returned?

 

[gkoenig]

Did you sign it off yet?

 

[Rcgiovannani]

Did you sign it off yet?

Yes... I've already made payments on it.

First time buyer mistake I guess...

 

[PegroProX440]

Time to start name dropping the builder MTB and see if they start taking you seriously.

 

[rklopp]

So apparently something is "squishy" in the vertical direction. I would expect the weight of the spindle head to take out any mechanical slack in the Z direction, such that your hand could not move any of that. Maybe my expectations are off. I would expect your hands to be able to affect a poorly tightened or non-flat column to base connection (I'm assuming it has a separate column). Otherwise, I'd go back to servo tuning, as if there's a big dead band in the allowed position error feedback control. I know that I can make typical servos fight back against my hand (their noise changes as they work to hold position). What happens if you push on X or Y instead of Z?

On edit: Do you have a right angle head? If so, use it to run circle-diamond-square tests in the XZ or YZ plane or both, at "normal" or slightly aggressive feeds and speeds.

 

[mmurray70]

I couldn't do this next test very scientifically but I decided to try and lift up on the head of the machine while the indicator was on the table. I could move the head almost .001" with one hand. With two hands I could move it multiple thou. I did the one hand test on my Fryer and it only moved 0.0003". One hand on a Haas and it only moved 0.0004". Got the head to move about a thou on both the Haas and the Fryer with two hands.

I've noticed this before too. Amazing you can see flex in a machine so easily. This does seem more then normal.

I was going to post a while back suggesting this could be your problem. The whole head bouncing up and down. Maybe resonating somehow because of some vibration in the spindle. Try an indicator between table and head and turn the spindle on, see if it starts moving.

 

[Vancbiker]

If the machine is box ways on Z you may have gibs on the keeper plates that need adjusting. Some machines with box ways don’t use gibs and the clearance is set by step grinding the keeper plates.

In any case those steps are up to the builder on a new machine.

If this has linear ways then no adjustment is possible. It’s a design, spec, or assembly error.

 

[Rcgiovannani]

A little update.

I put an indicator in the spindle and jogged the machine in the x axis. The indicator barely moved. Very smooth.

I did the same thing on the Y axis and...well, not so good. The needle was jumping all over the place and vibrating like crazy on the initial moves, and then would smooth out after it was in motion for a little while.

This machine has linear roller bearings, btw.

X-axis smooth motion:

https://share.icloud.com/photos/0e50PI4G_phDNPf05l2HnfhoQ

Y-axis jitters:

https://share.icloud.com/photos/00cfgfCqOVcA2NjUgBFidGr0A

running a 2d pocket operation with the indicator zero'ed. Please excuse the poor camera work:

https://share.icloud.com/photos/04aDom604CDqzVmbfaQHdzQCQ

So maybe something loose in the Y axis? I didn't think to try and push on one side of the table to see if I could "skew" the table. I will try that next. Doesn't really explain the z axis head movement I can measure but who knows...

MTB is spending the day with me tomorrow. They're trying, but I need this machine making money!

 

[rklopp]

A little update.

I put an indicator in the spindle and jogged the machine in the x axis. The indicator barely moved. Very smooth.

I did the same thing on the Y axis and...well, not so good. The needle was jumping all over the place and vibrating like crazy on the initial moves, and then would smooth out after it was in motion for a little while.

This machine has linear roller bearings, btw.

X-axis smooth motion:

https://share.icloud.com/photos/0e50PI4G_phDNPf05l2HnfhoQ

Y-axis jitters:

https://share.icloud.com/photos/00cfgfCqOVcA2NjUgBFidGr0A

running a 2d pocket operation with the indicator zero'ed. Please excuse the poor camera work:

https://share.icloud.com/photos/04aDom604CDqzVmbfaQHdzQCQ

So maybe something loose in the Y axis? I didn't think to try and push on one side of the table to see if I could "skew" the table. I will try that next. Doesn't really explain the z axis head movement I can measure but who knows...

MTB is spending the day with me tomorrow. They're trying, but I need this machine making money!

In the pocketing op it's interesting that the effect from Y motion "sticks" when switching back to X motion. Is it possible the Y way covers have a lot of drag and are pulling and pushing on the column?

 

[EndlessWaltz]

I am happy this thread exists and problem solving has been thorough. A great learning device for us here. 1) Don't pay till proven. 2) Drunk guy built on a Monday.