VFD with 'partial' support for PM motors

[lmarcantonio]

At the moment we are running with standard induction motors, however we choose the VFD (Hitachi WJ-C1) since the vendor/FAE said "it will works with PM motors too" and we wanted to keep the option for that. However (in the small writing of the spec) they actually say

"The permanent magnet motor control function is only suitable for variable torque applications such as fan and pump. When using sensorless vector control for permanent magnet motor (PM), please contact your dealer."

Of course, we are hoisting and the vendor fully knew that. Now, before trying and having things falling around, do you ever had such a statement? Motor vendor said that installing a PG should make it work (the C1 has the input so it's doable) but I'm wondering about why and the consequences (the VFD vendor is useless, he says "it works anyway").
My theories:

• in constant torque an overload would cause a slow down (i.e. loss of synchronous speed, due to no slip allowance) with catastrophic load consequences; a PG would probably allow the VFD to at least get the field in sync. Wouldn't sensorless get it however? maybe at low speed even PM motors have tracking issues?
• maybe the control algorithm doesn't support field weakening? shouldn't matter since it's an issue for quadratic loads too and it only matters at high speeds; if I drive a motor at its rated frequency it shouldn't be an issue

Anyone with insights?

 

[cswilson]

If you are driving a PM motor with a VFD with no feedback of any kind, you are basically treating it as a big stepper motor in microstepping mode. As you suspect, overload would cause a catastrophic failure - not just loss of torque, but rapidly alternating torque tending to bring the motor to an immediate stop and vibrate violently about this point.

The stepper motor people call this a "stall" condition. But it is one thing for a 100-watt motor to stall in a desktop system, completely another for a large hoist motor. Also, the sudden loss of back EMF from the stall can lead quickly to an overcurrent condition if there is not a good current loop to quickly reduce the current.

If the VFD is using interal voltage and current sensors for "sensorless vector" control (what academic papers call "shaft sensorless" control, things are better. However, at low speeds (and so low back EMF), it is very difficult to get good enough readings (low signal-to-noise ratio) for proper closed-loop operation. So I believe many of these VFDs really run open-loop at low speeds.

With a shaft position sensor (PG in your vendor's terminology), it should be possible to get robust control at all speeds. You have everything you need for full servo control.

I do not know how good this particular vendor's algorithms are in any of these modes. But hopefully, you will be more ready to find out.

Curt Wilson
Omron Delta Tau

 

[jraef]

“Partial” PM motor support could mean that they support IPM (Internal Permanent Magnet) but not SPM (Surface Permanent Magnet) motors, or some combination. For example the PowerFlex 525 drives are like that. You can use them in Open Loop on SPM and IPM motors, but if you want closed loop, it only works on IPM. This issue is not uncommon for the small “component class” Asian based drives, because their microprocessors don’t have enough number crunching HP to do it all.

It could also mean that they have tested their algorithm with certain specific PM motors but not others. That’s somewhat common too because not all PM motors are created equal.

You really need to get them to nail that down for you, because if it is that they can’t do Closed Loop control on an SPM motor, and you have an SPM motor, it ain’t gonna work!

And… don’t trust what a salesperson tells you!

 

[lmarcantonio]

Eh, too bad that this salesperson is also the main support for Hitachi Italy; luckily he has some technically competent people at the branch. Anyway, I solved the mistery: that VFD has a pulse input for the shaft encoder but doesn't use it in PM mode. It's only used for stabilizing V/f speed (and other unrelated function).
Actually Hitachi doesn't seem to make a VFD with that support, they said "here's a servo drive, it only costs 15 times more"
Time to go shopping, thanks for the assistance