Do BLDC motors have saliency similar to IPMSMs?

Question

I am experimenting with an outrunner BLDC motor, which is a Gaui GUEC GM410. It is a 12N14P motor and it has sinusoidal bEMF.

If my assumption is corrent, it can be considered to be a SMPMSM with no iron teeth between the permanent magnets, thus there is no dependence in reluctance (and consequently inductance) with regard to the rotor angle, unlike in an IPMSM.

However if I connect a small amplitude (1-2V range) 200kHz sinusoidal test signal between two phases and measure the voltage on the third while slowly rotating the rotor by hand, the ratio of the output and input amplitude is clearly changing a bit, from about 0.45 to 0.55, so around the expected 0.5.

My question is what causes the saliency in this type of motor? Since the test signal is low, I don't believe it has anything to do with saturation.

My final purpose of knowing the source of this saliency would be to fully understand the model here, where the expression of the A phase inductance is $$L_{aa}=L_s+L_mcos(2\theta_e)$$ with Lm being negative as default in the parameters section.

In this article on the second page in the corresponding equation it is negative and is specified as $$L_{ms}= \frac{1}{2}(L_d – L_q)$$ so the sign eventually depends on whether Ld or Lq is bigger. Currently I cannot measure them, but that would not help to identify the reason for their difference anyway.

Answer 1

The PM has to be magnetised in a such way that BEMF is sinusoidal. Therefore the flux density is not even on the whole surface area. This makes the difference when the magnet is passing over the slot gap. A minimal saliency is almost inevitable.

Answer 2

I'm not sure but for my understanding, BLDC motor has trapezoidal b-EMF. If your motor has sinusoidal b-EMF waveform then it's probably a PMSM with distributed stator windings. Then the saliency is up to measurment.

For your question, I think one possibility is that your motor actually has an IPM rotor. Then you can measure the Ld and Lq. Normally Lq is larger than Ld because the permeability of magnet is very close to air so on d-axis it's facing equivalent longer air gap.

Another possibility from my guessing is due to your test signal is of so high frequency that the iron losses can not be neglected anymore. Because for eddy current losses ∝ f^2, for hysteresis losses ∝ f.

Hope it's helpful, and I'm also curious about the reason if you find out.