My father and I have a company that services electric motors. We can repair every ac, dc, induction motor.. But these days factories are replacing there dc brush motors with servo motors. So if I want to stay and work for the factories I need to start repairing servo motors. The winding itself isn't different then normal induction motor, but the rotor has permanent magnets, encoder at the end,,, etc.My biggest problem is that once I dissmental the motor and put it back together it doesn't work anymore. I am thinking that is has something to do with the position of magnets in rotor and position of the encoder on the shaft. But I need to learn more.
So any help or info would be very appreciated.

My email is boris.matijasevic @ siol.net

Regards
Boris

 

The feedback device must be aligned properly on a BLDC servo motor. You may find hall effect sensors, UVW channels on an encoder, an encoder index mark, a sinusoidal encoder, a resolver, or a "smart" encoder. Or some combination of the above. It can be complicated. Some manufactures provide phasing info, some may not. Good luck.

 

There are a number of possible problems. One is that there are usually sensors that have to be aligned correctly. Sometimes when you take a servo motor apart and put it back together again, these sensors end up getting moved. Servo motors still have to be commutated. This is just done electronically in the drive instead of with a commutator. The sensors in the motor provide this commutation postion to the servo drive. How to do this properly varies according to the servo motor design. You would have to get this information from the manufacturer (or some other source).

Another problem (and there is a good chance that this is the problem you have) is that the magnets used in servo motors are often different from those used in normal permanent magnet DC motors. Often, if you break the flux path by disassembling the motor, the magnets lose their flux. This happens instantly on removing the armature. Whether or not this happens depends upon the magnet material in that model of servo. In order to disassemble the motor without losing flux, you would need to insert an alternative flux path which is inserted as part of the process of removing the armature.

Once the flux is lost, all that can be done is to recharge the magnets with a magnet charger. Even if you have a magnet charger with the right number of poles, you would have to know how the original manufacturer did things (outside in? inside out? with the sensors present or removed? etc.).

 

That's not even the whole story. Many motors have exacting requirements for commutation angles and there is a specific procedure to align them after repair. This may or may not be published and is very likely specific to each brand or configuration. I have seen some for Siemens motors that required special equipment and some that looked doable on the floor. I would expect that once you get the idea of what they are setting up, you might be able to tune them in a pinch without the specs. We do have local shops that manage to fix them with good results most of the time.

Regards
cww

 

Boris,

You are quite correct, that servo motors are differend from standard motors. BLDC motors are mainly synchronous 3 phase motors. They have no slip rings for curent commutation. Thus the 3 phase currents are commutated in the servo amplifier. It is a requirement that the current commutation be synchronized with the position of the motor armature. The trend is to use a resolver on the motor shaft and it must be synchrinized to the motor armature position In the manufacturing of these motors, the sychronization is done on special fixture. Once this is done, the position transducer (resolver or digital device) should not be changed. This position signal is used to control the current commutation, and to create a synthetic velocity loop by differentiating the position signal.

In actual use, some service people do not know this and can not figure out why the motor does not work. In our machine tool company, a bldc motor that has had the position transducer moved, goes back to the factory and gets reset. The manufacturer has the test equipment to readily reset the position transducer. It can be done in the field IF one knows what they are doing, but it is cheaper and more expedient to send the motor back to have the transducer realigned.

In your case, I recommend that you contact a motor manufacturer and explain that you need to know how to reset the resolver, etc., so you can do it in your business. It is not a major problem to reset the resolver. I am sure you can do it with some instruction and instrumentation. In our case we just did not want to bother with it, with the many drives being installed daily.

Regards

George Younkin, P.E., MSEE
Life FELLOW IEEE
Bulls Eye Research, Inc.
Fond du Lac Wisconsin, 54935
[email protected]

 

Hi Boris,
As mentioned previously in this thread, in a BLDC, application timing of the encoder with respect to the motor is critical.

In these applications the encoder has a set of three Commutation tracks (U,V,W) which correspond to the three phases of the motor (R,S,T). Alignment of the encoder commutation signals to the motor could be thought of as being comparable to timing the distributor of an auto engine. Just as the distributor tells the spark plugs when to fire, the encoder tells the amplifier/drive when to fire the windings in the BLDC motor.

Just like in a car, if the alignment is off, the motor will not run correctly, will run inefficiently, or not at all.

If two phases are reversed, it may even run backwards.

While it is always advisable to have the manufacturer realign the encoder to the motor, if you have the right information, it is not impossible to do in the field.

While I have not done this myself and am therefore not an expert on it, as I understand it the process goes something like this:


1) One phase of the motor is energized which locks the motor into position.

2)The encoder is rotated to a given position which is usually the start of one of the commutation signals (I.E. leading edge of U). Often times this corresponds with the encoders index pulse.

3)The encoder is assembled to the motor and the the shafts are locked. (usually via the encoder set screws) The encoder flex mount is not yet secured.

4)Motor winding is de-energized. Encoder is powered.

5)The motor/encoder is back driven by another motor and two waveforms are displayed on an oscilloscope. One waveform is the motor phase, and one is the encoder commutation channel.

6) While the motor is rotating, the assembly is fine tuned by rotating the encoder body to align the encoder signal to the Motor waveform.

7) Once alignment is achieved, the encoder flex mount is secured, locking in the phase relationship between motor and encoder.

I am describing the way this would be done using the encoders I am most familiar with. Please note that not all incremental encoders would have a "flex mount". Modular encoders do not for example. and not all are secured to the motor shaft with set screws.

Keep in mind that you will need to have the information on which motor winding corresponds to which encoder commutation signal, and to what degree they need to be aligned.

If you plan on doing this in the field, another thought is that you may want to mark the encoder and motor shafts with respect to their bodies before disassembly. This may ease replacement.

Jim Miller
Application Engineer
Quantum Devices Inc.
[email protected]
www.quantumdev.com
(608) 924-3000

 

I am just adding a blog post and video that speaks to this topic.

Blog post:
http://www.quantumdevices.wordpress.com

The video can be seen here:

 

One of the problems you will have with repairing servo motors is that the manufacturers of most servo motors want to do the service themselves. They will often not sell parts, not even connectors. Obtaining components such as tachs, hall sensors (these for older motors) and encoders (for older and newer motors) can be a problem.

Also know that, with some brushless motors, mere removal of the permanent magnet rotor will result in demagnetization of rotor. Without the remagnetization fixture and power supply, you can not remagnetize the rotor and very little torque will be achieved from the motor even when properly reassembled.

Sorry to be so negative, but......

 

Larry,

I agree with most of your points, but I just had one comment/question. Most motion controllers' "defacto standard" connection to a drive is via analog torque command (with the servo drive doing commutation and current loop closure). Short of using a serial connection (SERCOS, EtherCAT, MACRO, etc) how would you interface to the drive? Are you suggesting a pulse train (pulse and direction) to an amplifier is better than analog torque mode? I would imagine this really depends on the drive and motion controller you pick as to which may be better.

> Stay away from any controller that outputs a current (torque)
> command to the servo amplifier, using the encoder to close

KEJR

 

I am fairly sure that he means an analog torque command as opposed to an analog velocity command.  The analog velocity command moves the velocity loop down to the drive.

Bill Sturm