[CNC Viking]

Today my Y-axis AC-servo started to make loud noice and vibrations at random positions during Y-axis travel.

The noice is a strong humming noice and it vibrates through the whole machine.

I first encountered it during programmed machining, but it appears during jogging and manual cranking as well. When operating the Y-axis there is a constant background rattle and the random strong noices I mentioned, but every 3mm (1/4 servo-rev.) it is completely silent. As I continue, the rattle starts again and soon I run into those noice / vibration spots again.

No alarms are shown on the display, and the part I was cutting came out fine, with correct dimensions.

The X-axis has a weak background rattle and has had so as long as I have owned the machine. The Z-axis is completely silent.

The machine only has 4500 machine-on-hours with far less working hours so I am a bit puzzled.

Anyone that can share some information so I can fix this? I don't dare to run again it before I have fixed this.

[Docter Don]

My experience with that kind of sudden-onset vibration is that the coupler from the encoder to the motor or the ball screw is loose or broken, or else that the encoder has a bad spot, so check the mechanical coupling, and then replace the encoder.

Hope this helps.

[Tom Antrim]

Do you have an extra encoder? If not does one of the other drives have the same encoder? You might exchange Y with one of the other ones and see if you chase the problem.

[CNC Viking]

Thanks for your input Doc.,

This morning I started by switching the X- and Y- servo amplifiers. The Y-axis run smoothly but the X-axis started a constant high pich howling immediately. It moved when I cranked the manual dial, but I had to shut off the control because of the loud noise. I switched back the amplifiers and started up the machine and both X- and Y- axis were fine for a couple of hours, then Y started getting worse again, and after 5 hours I had to give up since it was as bad as yesterday.

I removed the Y-axis waycover and started checking things. The flexible coupling had 1 screw that was very easy to loosen and the flange coupling towards the ballscrew had 1 screw of 4 that was completely loose and the other 3 could be additionally tightened. When I had tightened everything up, the servo motor started the same constant high pitch howling as the X-motor did earlier this morning and it also vibrated.

I removed the servo motor and then the cover of the encoder housing. Now I don't dare to go any further before I know exactly what to do.

Mitsubishi servo HA80NT-E33, built 1996.
Sorry no camera with me today, so I have to explain.

The first thing you see is a circuit board screwed to 3 columns of the metal base plate. The screw heads are sealed with paint. Underneath the circuit board is a rotating glass disc with some delicate coupling at the centre. Underneath the disc is the baseplate which is secured to the motor housing by 4 screws. If I loosen those screws, I can rotate the baseplate but not easily lift it off the motor end.

Waiting for further advise before I remove it from the motor...

EDIT: Sorry Tom, you just answered while I was writng mine, but I think that my answer works for you as well... I will bring home the motor, take some pics of the encoder and post later tonight.

[Tom Antrim]

Are the servos inside the control cabinet the same? If they are swap them and see if it chases the problem

[mactec54]

CNC Viking

You really should not mess with the encoder, you are dealing with a very precisely installed encoder, on the Misusbishi motor

What has happened is now the machine has had some running, it is now moving easier, than when new, you need to,Tune the drive, start with a little less gain & this will solve your problem

There may be just some filters that are in your control that need ajusting,if not you have to do it in the drives

[Al_The_Man]

On most Mitsubishi AC servo's the encoder is keyed on to the shaft in order to correctly register the commutation angle, I had a problem on some Mitsubishi's of the encoder disc coming loose causing an error on the commutation and usually resulted in excess current.
Al.

[Brian Polley]

If you determine that the servo motor needs service, maybe the encoder has gone defective--which is common, or the axis card is bad, I have a source for repair in Canada-Oakville, Ontario- Accu Electric Accu Electric Motors - Industrial Repair Services | Call: 1-888-932-9183. They have a New York office that US customers can ship motors into. You can contact me at [email protected] for details.

[onthebumper]

Try running the motor while it is disconnected from the ball screw. If it still vibrates then you know the mechanics of the machine are good. You can try to reduce the gains and sometimes this will help as some manufacturers crank the gains as high as they can but this causes problems down the road when the machine gets a little looser mechanicaly. What type of control do you have so we can tell you what servo parameter to change?

[CNC Viking]

Tom Antrim:
The X- and Y- servo amplifiers inside the control cabinet are the same MDS-A-SVJ-10, but for the Z-axis there is a slightly different one MDS-A-SVJ-20 with an internal fan as well. As for testing, I have been told to stay away from the Z-axis amplifier. Don't mess with that one!!, he said.

I would like to move the X-axis motor to the Y-axis position and try again for some hours with the original Y-axis amplifier. Maybe it does not scream there.
If it does, should I suspect the Y-axis amplifier then?

Can this be done without any motor present on the X-amplifier while that amplifier is still being powered?

mactec54/Al_The_Man:
This is really over my head. I will have to consult some experienced CNC-technician here in Sweden when it comes to that.

OK, I may have done something stupid by loosening the 4 corner screws for the encoder baseplate and thereby messing up the rotational tuning against the motor. If this is no good, how can I retune it again? However I have not used any force attempting to pull it off the motor.

The servo amplifiers. From left to right, X, Y, Z.


X- and Y-axis servo motor labels.



The encoder inside. There is an inner plastic cover as well, not shown here. Note the separate white connector for the 2 wires snaking into the motor.

A closer look at the encoder's reading department.


The rotating glass disc with its special hub.


When rotating the shaft, everything feels OK. Firm bearings, smooth rotation and no axial play. I don't think anything is bad mechanically with the motor.

[Tom Antrim]

You'll have to write down all the wiring numbers then remove the servo amp swap them and wire them back the proper places. Put the connectors in as they were before.
I worked as a service engineer for Mazak but I wouldn't have taken apart the encoder. If you haven't damaged it it should go back together.

[CNC Viking]

onthebumper:

I will test-run the motor asap, disconnected mechanically from the ballscrew. The Y-axis is really smooth , I can easily move the table by rotating the ballscrew by hand.

My control is a Mitsubishi Meldas 520AMR built 1996.

I welcome any info re. where to find the relevant parameters in the menu's and what values to go for. First I will look it up and then report back the existing values for your advise.

[CNC Viking]

Tom Antrim:

It is OK, I did it this morning. I only swap wires U, V, W and ground + swap the connectors from the encoders between X- and Y-amplifiers.

What about only testing Y-axis amplifier with X-axis motor on it and leaving the X-axis amplifier without any motor during the test?

As for the encoder, I don't think I have damaged anything, just loosened the baseplate's corner screws and as the holes in the baseplate are larger than the screws, this allows a certain rotational displacement. It worries me a bit that I may have ruined any tuning, since several people here talk about tuning it against the motor and all.

What is the proper way to replace an encoder anyway?

[Tom Antrim]

I never have tried that. I alway disconnected everything and rewired after changing their position. Sometimes if the amp is not under load it may appear to operate fine.

[mariner]

I do not think the encoder is the root of your problem. The problem is the servo drive itself. It is 1996 model, so it at the end of its life. You have to service it urgently (ball park figure A$1000) before it dies completely. We could help but we are in Australia, so you have to find a local electronic workshop with some expertise in repairing servos down to component level.

[CNC Viking]

Tom Antrim:
It is only the power wires R, S, T, the brake resistors, and the bus wires from the CPU connecting all amps in a chain left. Don't think it will cause any problem to leave them as they are. Or maybe?

mariner:
What is the most common failure when an amp "dies"? What is normally done when servicing the amps?

[Tom Antrim]

They will normally alarm out because it can't take the current required to function. If you haven't change any of them I would say it just has quit from old age. It is one of the expenses of operating a CNC machine. I always recommended having a spare one around because the down time can eat up the price of one. If you have a reputable electronic repair company around they might fix it but I would probably contact Mitsubishi to see if you can get one. I would get a extra encoder too.

[Al_The_Man]

I integrated one machine with a Mitsubishi system about 8 yrs ago, after 18months they started going through motor/encoders about every 12 months.
We eventually pinned it down to the encoder disc coming free from the centre adhesive.
These were OSE253S encoders, we went to a different encoder number and so far have been trouble free.
These encoder can be removed quite easily as they have a double ear key to drive with, they usually have an arrow or black mark to orient the double key, although it should not matter as they are 8 pole motors.
Al.

[Rait]

I think you have to change the servo amplifier in electrical cabinet and encoder as well. We have same control and we had same symptons in one axis. For testing swap in electrical cabinet X-axis servo wires to the Y- axis servo and Y-axis servo wires to X-axis servo. When same symptoms are in other axis then you need to change your servo amplifier.
If you opened encoders screws which are holding the green plate and are locked with paint then you have to change encoder also. I made this mistake once.
In sweden you could ask for help from Highpeak.se - Start
Good luck.

[canwong]

Sometimes it is machanical parts cause the noise, like coupling connection or axis.