as you Know, The PM Motor either should rotate same as stator field speed or not rotate anymore, however there are a little vibration in application with PM motors, whats the reason of vibration?

vibration can result from Unbalance. but in PM motor why it happend? as you know Pm Motor has no slip.

Hi David

1. The assembly could be wrong.
2. The magnets always try to align them as per the magnetic fields even when there is no current given to the motor.
3. Are you talking of the bower vibrating ? If yes that may be mainly because of shaft selection.

You can choose to give more details of your problem also .

I talked about vibration below 1% of nominal speed.
all type PM motors that i used have such vibration.
im sure that these motors are intact and there are no problem about shaft and balancing.
i know when the load torque change, a little vibration occurs in the motor speed.
also we Know PM motor should be running either with the preset speed(Command speed) or stop. but always there are a little vibration in motor speed.

At Max RPM . Vibrations will depend on assembly. They become a bigger problem when you install in higher static pressure places because of the back pressure on the fan blades. They become disastrous if resonance is caused . You will have options of skipping certain RPM in the PM controller which can help in reduction of the vibration or should i say skipping the vibration.

Does the vibration frequency changes with motor speed? You have to identify the relationship between the vibration frequency and motor speed. Many reasons could cause motor vibrate.

David,

There can be multitude of reasons of having vibration in PMSMs.
these can be broken into two categories: 1. Vibrations due to the mechanical elements (coupling, rotor alignment, mechanical resonance, and other factors which some were mentioned in the discussion). 2. Vibrations due to the electromagnetic elements.
Even if assuming that the PM is perfectly (mechanically) designed and coupled, vibrations due to the electromagnetic components will be present! the vibrations due to this component can be attributed to three factors: a) cogging ripple torque - the attractions of the PMs to the stator elements b) B-EMF - current ripple torque - the interaction between the B-EMF and the stator currents c) current-current ripple torque - interaction of current harmonics due to position dependent reluctance (only in salient machines).

Nir has the best answer. Torque ripple can be caused by current loop issues such as offset, imbalance, and lack of resolution in the amplifier or by the magnetic design. (Sine wave motors are much better than trap or 6 step in this area). Velocity ripple is usually caused by inaccuracy of the position sensor (dp/dt calculation can amplifier the error). Softening gains and adding inertia can hide some of these. Its amazing how the distribution of error on the feedback device affects the dp/dt calculation!!

Nir has some good answers. Low speed/frequency magnetic cogging can cause vibration noise. This vibration would be a function of the rpm, the number of PM poles on the rotor and the number of stator wiring slots. The rotor will accel/decel repeatedly during rotation as the permanent magnets experience the changing flux paths which change the magnetic attraction to the steel stator. I have observed this cogging using Velocity waveform recording software on a PC. Strong, rare earth magnets on 4 pole motors with low inertia rotors, usually show the worst velocity ripple.
Skewing the magnets helps, increasing mass of rotor or adding a flywheel helps, adding more poles to stator helps, but limits top speeds. For lowest velocity ripple you can use a PM motor stator with slotless wiring design, so the magnetic flux path and air gap never changes during rotor rotation. There are some cost/performance trade-offs, but for velocity sensitive applications it may provide a solution.

Very interesting discussion guys! I am learning a lot.
Wayne, you are absolutely right. Here is another prime example why rare earth magnets with a very high Lambda_m coefficient are not always the best way to go (especially with low-medium power applications). Initially, one would think that you get more bang (flux linkage) for your buck, but the truth is that this is hardly the case.

Another interesting design that i saw regarding the issue of start up of a PMSM with significant cogging is to add 'squirrel cage like' bars on the outer part of the rotor (I would imagine it would be mechanically easier to do so in an IPMSM than a SPMSM). So essentially you start the machine as a induction motor and than switch to a normal PMSM control method once you have sufficient speed.
I found it to be a really neat and simple idea. what do you guys think?