Thread: possessed mori seki !

We have a truely possed Mori Seki ME junior.
Heres a description and some of the things we have done to isolate the problem without much success. The Y-axis servo will on occasion start making a resonant singing I guessing about 1000hz. If the feed rate is cut to zero it still resonates for nearly a minute. We can usually induce this if we run very high feedrates for a while. Heres what we have done to try to isolate the problem.
1) swap X and Y servo drives
2) swap X and Y servo motors
3) Run X axis with Y output (crazy but were desperate)

We have even swapped servo drive inputs between X and Y and run in manual mode.

In all cases the resonate problem remained in the Y-axis. Note that when the resonance occurs it remains even when slides are stationary. In fact. the machine can cut parts while the Y-axis is singing along. (but we dont let it continue).
Yasnac has no idea what the problem is over the phone and they wont guarantee they can fix if we shell out and fly them onsite. So were turning to this list in hopes someone may have some insight.

Thank You All Kindly, Tom Phillips.

Could it possibly be something mechanical and not the servo making noise. Or maybe something mechanical causing the servo to make noise. We had a moore #2 jig borer at a place I worked and it alway made a humming sound in y direction. I kept telling them something was wrong. They finally called the moore guy out to shut me up and he discovered the y way luber was plugged. Fixed it and the noise went away.

Does it make any difference if the spindle is running or not?

Like smalllshop said check other areas that can transmit the noise to the Y axis.

Check the end bearings of the Y ball screw as well as the ball screw nut. Could be just one or more loose bolts in the Y axis assembly.

If all else fails, then take 1/4 stick, uh no, it's a Mori take a 1/2 stick of dynamite and see what shacks loose. [img]smile.gif[/img]

How about a wire issue (like a bad wire going to the encoder), on the Y axis? That would seem to be a good way to get the problem to stay at the Y axis when you swap the motor or drive.

If you have scales that would be a place to look too.

Theee are all good ideas. The noise can be produced in jog mode spindle off. We pulled all oil lines and they are clear. We also ran the X axis with the Y axis cables (crazy huh) to eliminate a cable problem. So that leaves the theory that some defect in the Y mechanicals is inducing this resonance. We already replaced the Y-axis lead screw end bearings but have not yet pulled the nut. Turing Y-axis screw by hand is easy enough. So now we are wondering if the balls in the nut are skating and setting up a harmonic resonance in the Y-axis servo system. There appears to be some evidence that the machine was crashed in Y before we got it wich lends credence to damaged ball screw idea. Im just having a hard time believing the servo system has this sensitivity to a mechanical problem though.

Have you checked the Y axis for binding or lack of free movement especially when stationary?
M.

This is a long shot, but ..

The Mori Junior has a shorter (i.e. lower inertia) ballscrew on the Y axis than it does on the X axis. This would make high frequency occilations more likely in Y than in X.

I've seen Fanuc motors on low inertia systems "buzz" due to a worn oldham coupling between the pulse coder and the motor shaft. To find this coupling, you'll have to take the back cover off the motor and remove the pulse coder. When you remove the pulse coder, you'll find a small bronze coupling between the pulse coder shaft and the motor shaft. This coupling can't have ANY radial play in it. It must slide easily to prevent stress on the pulse coder bearings, but it must not be loose at all. Sometimes, just a tiny smear of high temperature grease on the coupling surfaces will quiet things down.

My guess is that you have worn couplings in BOTH motors, so swapping the motors doesn't make any difference. The loose coupling will only cause a buzz if the inertia is very low, so it only happens on the Y axis.

These pulse coders are used for position feedback, but they're also used to create a "Velocity" signal for the servo amplifier. The frequency of the pulses is fed into an F/V (Frequency to Velocity) conversion circuit, and that output is used as a tach signal by the servo. If the motor can move a tiny bit without generating any pulses, the motor will buzz like crazy.

We had this problem a lot on Trumph punch presses, where a low-inertial Fanuc motor was driving a low-inertial rack & pinion drive system. Those motors would buzz at the drop of a hat ... ANY hat.

Dan that makes a lot of sense. We had a rebuilt motor installed but Im guessing normal testing would not detect the pulse coder problem your describing. I did not realize they used an oldham couplings either. Ill check this out and post and update.

got an o'scope?

jeffe

Yes, got an old two channel tektronix that works.

We found something! Get this. The Y-axis lead screw seesm to be the problem. Heres why we think so. This lead screw is supported by two bearing at each end wich are normally in tension. We were checking the tension and found that.. loosening the tension so the bearing is not tight seems to GREATLY REDUCE the noise! After pondering this for a while we believe the nut is somewhat worn allowing radial vibratoin to occur with excitaton source the motor. The reason it diminishes when loosing the support bearing is it allows the screw to drop a bit and rest on the nut thus damping the vibrations. So, working on that theory we wrapped the plane section of the y-axis screw with heavy rubber strap (think brake drum turning). This totally killed the vibration!!!! We tried every possible test to induce it and couldnt! We have a Y-axis screw on order from Mori-Seki. the odd thing is the screw is well within tollerance for backlash and shows no signs of distress visually.
This is a new one for me.
-Tom Phillips

Wow, that's a new one to me!. Most ballscrews are not in tension like that. Usually one bearing set takes all the axial thrust, while any bearings at the other end just support the screw radially.

I wonder what happens to all that preload tension when the ballscrew changes temperature ??

The preload tension is to reduce thermal error. Your higher end machines should all have a nut on each end of the screw and pull at least 0.001 in length to tension it properly. a loose ballnut may be your problem, especially if you have scale feedback. scale feedback and loose screws = a motor that can't get happy with position, always hunting. Loosening the screw however, in my opinion is not the answer, just a piece of the puzzle to the solution. Check your alignment, check your lash, and leave that screw in tension.