Has anyone ever tore into a servo motor

[Kevin54]

The "Y" axis, as I have mentioned before on my CNC has a hard time finding position at times and will shake back and forth until it finally finds it's point. So in working on it and hiring a guy a few times to work on it, everything pointed to the encoder.

So last night I was programming it to make a Turners Cube to see how it ran. It finds it's position now, but there is a rattle inside of the servo when it shifts from forward to reverse or when it moves from a dead stop. If it has a load on it though, there is no noise, If it makes any difference, it is a permanent magnet servo made by Powertron. I believe at one time the motor had been rebuilt. If I can fix it myself, I will tear into it, IF I have an idea as to what it might be. I was looking on Fleabay and holy **** A new servo is $1800 and a rebuilt servo around $700.

All of this time, I believe the motor has been the problem and not the encoder which the new encoder was a $500 part. I could check with Powertron and see what it cost for a rebuild, but that is sending it in, letting them evaluate it, then they have you by the short hairs.

So I thought that if anyone has ever torn into one and knows what's inside, I may just work on it myself. I'm great at tearing **** apart.

 

[rockwithjason]

nothing that you and a 12 pack of beer can't fix!

 

[Kevin54]

I did find this, but it doesn't look like anything that can come loose inside from what I can tell unless one of the magnets is busted and brushing the armature, but I don't see how that would happen.

 

[MoonRise]

The 'servomotor' system is just a :

- motor (can be almost any type of motor)

- feedback position encoder (so that the drive controller 'knows' where things are)

- the drive controller

If the drive system is 'rattling' around, that sounds like either the motor bearings are shot, or the encoder or drive controller are having issues. Or all of the above.

Yeah, the tearing things apart is relatively easy.

It's the careful disassembly and repair and then reassembly that can be tricky.

 

[A_Pmech]

Kevin,

If the shaft encoder is loose on the shaft or the servo is not tightly secured to the screw, it could cause the servo to hunt for position.

Also, have you tried tuning the drive? If the acceleration is set too high and the decel set too low, the motor will over-run position.

 

[Kevin54]

Kevin,

If the shaft encoder is loose on the shaft, it could cause the servo to hunt for position.

Also, have you tried tuning the drive? If the acceleration is set too high and the decel set too low, the motor will over-run position.

The guy that worked on it did all of that. He did tell me that the motor was rebuilt. By who is anyone's guess, but he doesn't think it is wound correctly and it was wound either a 1/4 or 1/2 off when they started. I understood what he was talking about there, but this is a new problem that just started last night. It is a loud rattle, and it only does it when the "Y" changes position. I pulled the front cover off while it was running to see if the cog on the motor was tight and to make sure nothing slipped to where the belt was loose, and all of that is good. I was thinking about swapping the "Z" and "Y" servo until this rattle started. I have it running at the moment and took an automotive stethoscope to it, and it sounds like it is coming from the front of the motor (Operator's side) instead of the back machine side where the encoder is at.

 

[rockwithjason]

you can handle the tear down. time and care are the important things there. if you have the noise located then you and look at things like bearings and broken parts

 

[tdkkart]

Other than the encoder, there's nothing in there that's not in any other motor.

Hunting-- very likely the lead screw or ball screw whichever is has. Keep an eye on it and see if it happens more often in certain table positions more than others. Also try leaning against the table slightly and see if that helps.

 

[Kevin54]

Other than the encoder, there's nothing in there that's not in any other motor.

Hunting-- very likely the lead screw or ball screw whichever is has. Keep an eye on it and see if it happens more often in certain table positions more than others. Also try leaning against the table slightly and see if that helps.

It's not the leadscrew or ballscrew I'm worried about. There is an issue with the motor. The cog on the motor that drives the belt, I can take the ball bearing center off of my lathe and use it to push in on the shaft of the motor where the cog is and the noise will stop. And it's no small noise either, it is a rather large rattling sound.

I just now pulled the motor off so I am going to let it cool a little, then pull it apart. According to one site I was looking at, there is a little more to a servo that just an armature and windings. I'll see if I can find the site again, but it was talking about a brake system being in there.

This shows some of the stuff in different brands http://www.servotechusa.com/servo-motor-parts/

 

[theoldwizard1]

The 'servomotor' system is just a :

- motor (can be almost any type of motor)

- feedback position encoder (so that the drive controller 'knows' where things are)

- the drive controller

If the drive system is 'rattling' around, that sounds like either the motor bearings are shot, or the encoder or drive controller are having issues. Or all of the above.

I'm no expert on servomotors used on CNC machines, BUT the adove statements are accurate.

Depending on the vintage, sped and load requirements, you motor may be a traditional DC motor (like your pictures) or may be a stepper motor. Steppers don't necessarily hand heavy loads (tool pressure) well, unless they are geared down. Stepper motors are also called brushless for the obvious reason.

If you have a DC motor, make sure the brushes are in good shape and the commutator (the area on the shaft where the brushes touch) is not excessively worn and reasonably clean. Look for a spray can of electronic/brush cleaner.

I have seen the "jitter" you mentioned on some servo systems after the device being moved come to a stop. Make sure the encoder pickup (optical ?) id clean.

 

[theoldwizard1]

This shows some of the stuff in different brands http://www.servotechusa.com/servo-motor-parts/

Those are a LOT bigger than anything I have seen !

 

[MoonRise]

http://www.servotechusa.com/blog/bid/31931/Disassembling-your-own-DC-Servo-Motor/default.aspx

 

[Kevin54]

Well, I'm out of commission. Every damn magnet in the top of the housing is broke and shifted. I don't know if I can get in touch with the company Powertron and get just that portion rebuilt or if they are going to want to sell me a complete motor. I don't have that sort of money to toss at it right now. I'll send them an e-mail and see what they say, but I have an idea it's not going to be good at all. That is why it's always had a hard time finding it's position. The magnets are set in three different areas set evenly around the housing. Now there is like 7 magnets in there all shifted around everywhere. I probably never needed a new encoder at all and it's been the motor all along that was supposedly rebuilt.

 

[kbs2244]

I would guess it is a stepper motor.

I used to sell a special purpose printer that used one to move the print head.
It was a direct drive system, no cog belts.

In order the problems were:
Dirty encoder "wheel"
Loose shaft connector at encoder or motor.
Loose motor mounting screws
Loose bolts holding the motor halves together.
(This was very rare, but on a "rebuilt" motor ???)

Steppers do move in steps, so there is more vibration by definition.
But the problem is designed in.
I did take one of ours, consigned to NFG, apart once out of curiosity.
It turned to just be dirty inside some how.
Some WD40 fixed it, but the techs wouldn’t touch it.
So it went back into the trash.

Steppers depend on mechanical integrity for accuracy.
In your case I would try a new cog belt before I got too carried away.
A cog belt jumping a tooth can sound and feel like more than it is.

Well..
I typed up the above before you posted that last one.
For sure a "rebuilt" problem.
Holding the magnets in place is pretty basic.
Someone, somewhere, should have a warranty claim to give you some relief.
I would start with whoever sent your repair tech.
He did not see the problem.
Then work up the responsibility chain.

 

[Kevin54]

The motor is a DC Servo permanent magnet motor and not a stepper. I bought the mill used and had the screws put to me when I bought it. After I had it I spoke with the person I bought it from and he told me that he had it, but he never even had the mill programmed in all of the time that he had it. All he did was jog it to fly cut some material. When I got it home and hooked it up and tried to program it was when I found out about the positioning problem. He had given me the name of an older gentleman who works on CNC's. So I gave him a call and the first thing he had to do was tune the CRT to coincide with my Phase Converter. Once that was done, I screwed with it trying to figure out the programming as I haven't programmed by G&M Code in quite a few years. That's when I found out the "Y" axis wouldn't hold position. After a number of hours on it, he decided that it was the encoder because it had either oil or coolant in it. I spoke with BEI Industries who makes the encoders and that style is no longer made, so it was a 30 day ordeal to get one made at the tune of $500, or take a chance on getting the one I had rebuilt. And if it couldn't be rebuilt, they would have to make a new one anyways, so I just went ahead and bought a new one. I called him up, he came up and put the new one on and double checked everything, and for a week things seemed to be good then it started acting up again.

I spoke with the guy and he said he thought that he remembered that the person I bought it from had the motor rebuilt, plus it has a different tag than the other two servos. This motor has a paper tag with all of the pertinent info on it which he said is a telltale sign that it's been worked on, plus it had been repainted, and had some paint marker writing on it.

I know a lot of people have picked up this type of Bridgeport CNC relatively cheap and converted them over so they could run newer software on them. I'm gonna surf the net first and see if I can find a used one, and then get in touch with the company tomorrow about rebuilding the top half where the magnets are, or even if that can be done.

Around here, we don't have any places that rebuilt motors except one, and that's a crapshoot at best with him. I go to Columbus, Ohio quite a bit, so I may check to see if there is a place over there to rebuild it. Basically all it needs is three new magnets put into it and epoxied in if that is what they do. Looking on FleaBay, the prices are outrageous for a Servo of the type I have. Look up a 402CNSP0S72. They are going for up to $2800. I've only started to look though so I still have to go through there and see what they have.

But that's why I had problems with the mills "Y" axis finding position is because the magnets inside gets shifting position. And yesterday was actually the first time since I've had the mill that I got the programming figured out. So the longer it ran, the more the magnets shifted around. Why it quit making noise when I pushed in on the armature, that I don't know unless it just moved thing just enough to quiet things down, but when it shifted from forward to reverse is when the noise was bad and that's due to the magnets shifting back the other way. How it stayed as close to zero position and not move way off. On the cube I was cutting, center was within .003 on four sides, but there was a lot of vibration in the cut as you can really see it under a light and the circular steps were only
.100 deep at 8 steps/side.

I'll get something figured out on it, but I hate tossing more money at it as I've already spent enough. But once you are so far into it, it's either sink or swim, so I guess I'm going to have to swim.

 

[Dragoonmc]

I'm guessing by your description that the magnets are not on the rotor so I have to believe that someone in Columbus can get you new magnets. As for the motor staying in one position it isn't to much work for that and I bet the noise being so bad on direction changes was because the rotor was moving back and forth to the ends of its play

 

[Kevin54]

I'm guessing by your description that the magnets are not on the rotor so I have to believe that someone in Columbus can get you new magnets. As for the motor staying in one position it isn't to much work for that and I bet the noise being so bad on direction changes was because the rotor was moving back and forth to the ends of its play

It was making noise because the magnets were broke and some had stuck to the armature and the others were still glued to the case. So three large magnets are not about 8 different sized magnets. Part of them were stuck to the armature and shifting around every time the motor would shift direction

 

[rockwithjason]

i use a place in phoenix called keller electric (800) 888-5326. you might call them and see what your options are.

 

[454ragtop]

Two thoughts from someone who never worked on a servo motor in his life. First, I wonder if it's possible to Epoxy the broken magnets back into place, since the motor does work, just has a problem with the magnets shifting? Second, might post a question over on Practical Machinist, lot of CNC guys over there.
HTH, Jim

 

[macgyver37]

cnczone has a big group of people converting these to a newer control, they may have some of the stuff laying around and there are also vendors that do make things specifically for these, they may know of a good repair. You may also look into if there is an upgrade for it, one that has the same specs but is newer and potentially cheaper and easier to come by.

 

[zkling]

Sorry for your loss Kevin. I learned one thing with my clausing kondia. There was a reason it was cheap and that ~20 year old (at the time) CNC machine parts are just all around expensive to repair, source and replace. I ended up opting for the conversion when clausing quoted me parts ~2x what I paid for the mill. initially was running mach but then switched over to CNClinux and have been happier ever since.

Hopefully you can get it going with a sourced servo that will work as it sounds like your main controller is fine. CNCzone is the place to go for down and dirty system help. If it is just the one servo I don't think I would jump ship yet on that controller unless you really want to.

What exact mill do you have? (sorry I may over missed it in the posts). One thing I do remember from calling around is that it seemed support for the older Bridgeport machines was much greater than my clausing. Ended up trading a guy my old servos and broken controller for some tooling.

 

[Kevin54]

Two thoughts from someone who never worked on a servo motor in his life. First, I wonder if it's possible to Epoxy the broken magnets back into place, since the motor does work, just has a problem with the magnets shifting? Second, might post a question over on Practical Machinist, lot of CNC guys over there.
HTH, Jim

I don't know if you can or not seeing that the magnets are broke. One reason is when a magnet breaks in half, the poles change, so I don't know how that will react with the motor itself. IDK

Sorry for your loss Kevin. I learned one thing with my clausing kondia. There was a reason it was cheap and that ~20 year old (at the time) CNC machine parts are just all around expensive to repair, source and replace. I ended up opting for the conversion when clausing quoted me parts ~2x what I paid for the mill. initially was running mach but then switched over to CNClinux and have been happier ever since.

Hopefully you can get it going with a sourced servo that will work as it sounds like your main controller is fine. CNCzone is the place to go for down and dirty system help. If it is just the one servo I don't think I would jump ship yet on that controller unless you really want to.

What exact mill do you have? (sorry I may over missed it in the posts). One thing I do remember from calling around is that it seemed support for the older Bridgeport machines was much greater than my clausing. Ended up trading a guy my old servos and broken controller for some tooling.

I went to the CNC Zone and posted on there. The mill is a Bridgeport Series I with a Boss 9 Controller. A lot of people will buy them up and switch them over and some will take off the servos and put on steppers. For home use, and for the money, I'm fine with the Boss 9 controller, if I can just get the one servo repaired. Other than that, it is a decent mill.

 

[Kevin54]

Things may be salvageable to a point. I spoke with the person that worked on it before and told him what happened. He gave me the name of an electric motor repair place that works on servo's. So I gave them a call and spoke with a tech and explained what happened. If the magnets just came off, they can epoxy them back on. If they are shattered, they can't. After looking at it closer, the magnets that run top to bottom in the case are stacked. I have only one of the stacked magnets that is actually broke but they are not shattered.

They are concerned about that one with the same thought I had in my mind......a broken magnet changed poles. By that I mean that if you have a 3" long magnet, you have (1) North pole and (1) South pole. If you snap it in half, the area where it broke will either become a N or an S pole.

The said that it could be a problem, other than I explained the fact that the servo WAS running and holding position even with the magnets shifting. So it's not going to be too painful if they epoxy things back and it doesn't work. But if it ran with the magnets shifting around as far as position, in theory, putting them back in place, it should still work.

All three servos on the mill, the "X", "Y", & "Z" are the same servo's. So if this one runs fine after epoxying them back in place, I'll swap out the "Y" & "Z" Servo's. The reason for doing that is because the "Z" axis see's way less action than the "Y", and you don't have the forward/reverse direction in the "Z" like you would the "X" or "Y". So the "Z" servo should be in great shape, and by swapping them out, putting the "Y" into the "Z"'s location should add some life to the servo. That is because the "Z" won't have the forward and reverse action, and when it is running, is just to find position.

If it runs with the magnets epoxied back in, then all should be good. The tech at the electrical place told me that it is fairly common for magnets to come loose in a motor, but as far as putting in new magnets, it can't be done. So if epoxying in these, and it DOESN'T work, then I'll most definitely have to find either a new servo, or at the least a good used motor. If everything DOES work out, I think I'll still keep an eye out for a replacement motor if the price is right, just to have a spare.

So road trip tomorrow to the motor shop and see if they can put me back in business for a fairly decent price. I wouldn't think that it would cost much, and hopefully they can keep the price down below $100. Something like $20-$30 would be fantastic. We'll see, and keeping my fingers crossed!!!!!

 

[rockwithjason]

i am curious why they can't replace the magnets? seems like if you could find the correct replacements you could glue them in.

 

[Kevin54]

i am curious why they can't replace the magnets? seems like if you could find the correct replacements you could glue them in.

I guess it has something to do with the strength of the magnets, the radius of the magnets, thickness of the magnets, and so on. Plus there has to be a certain amount of gap between the magnets and the armature. If you ever tried to modified a magnet, it's damn near impossible.

I was making a special fixture at work one time and a portion of the fixture I used a magnet in it. I went to surface grind things, and the surface grinder barely phased the magnet at all, it was that hard. So I can understand now why it would take too much to try and find a magnet the right size for a specific motor.

I'm just hoping that they can epoxy these back in. I'll find out tomorrow. I'm fairly certain it will work if it was running with broken and dislocated magnets. So as long as they are back in place and won't move, it should run again. We'll see I guess