100 kRPM Brushless Motor Design 4

[sdk_imported]

Hello,

I would like to see some discussion about high-speed brushless motor & controller design(60-120kRPM, 1-5kW). Our biggest challenge is preventing rotor heating and maintaining stable speed control. We currently are using two-pole segmented magnets, segmented magnet sleeves, high resistivity materials, high speed switching, and external inductors ... but, we are still having issues. Will any of the following help me?

Slotless configuration
Non-PWM based speed control
Laminated rotor yoke
Something else I'm not thinking of..

If anyone has successfully designed and tested a motor in this speed and power class I would very interested in hearing from you. Finally, do you know any companies that make motors like this in large quantities?

Thanks.
Scott

 

[UKpete]

jb, I'm not sure that Barden are saying you can do away with TDBs. Though I have seen some wrecked mag bearings after high speed drops even with TDBs, the latter have to be rugged (and in themselves expensive) to survive. Active mag bearing systems also need controllers and possibly custom software, which isn't cheap.

I side with skogsgurra, for many high-speed applications ceramic ball bearings are going to win. There are no bearing current issues with them, and unless there are special reasons for not wanting oil (eg fuel cell compressors), it is technology that most are comfortable with.

 

[unclesyd]

Here is a outfit that does it all pertaining to rotary equipment. We used them and some others using their software on some very high speed equipment. I wasn't directly involved but I know the engineers that were involved were quite satisfied.

Though it doesn't run on electricity I have a rotary carver
that runs at 400,000 rpm on air bearings.

We had air bearings on some machines that ran 20,000 to 70,000 rpm. The touchdown was handled by the materials of construction.

http://www.rsr.com/index.html

 

[sdk_imported]

UKpete:
Our present motor design uses a Inconel sleeve with a 1mm gap between the sleeve OD and the stator. The stator is a six slot stator with ~3mm slot openings. It sounds like I should increase the air gap (with an appropriate increase in magnet thickness). Do you think a 3mm air gap would be sufficient?

The suggestion to look at back filled stator is interesting. I thought about making the stator such that there is no slot opening. Would this effectively reduce cogging torque even with the saturated regions between the teeth? Related: do you think skewing the stator would help reduce rotor heating?

The laminated magnet testing that we did was only a test case to verify that it does indeed have an impact on rotor temperature. Our present rotor design has magnet segments (high temperature Neo) that are ~10mm thick. Do you think we will see any benefit with segments this thick? What is the max thickness that you would consider beneficial?

Drive question for everyone:
I spoke with a motor controller design consultant and he questioned whether a sine drive will effectively reduce heating in the rotor because at high speeds it is difficult generate a clean sine wave. My question is this: Is there a benefit to using a sine drive even if the resultant current is not very sinusoidal? Why?

ShinEtsu has some interesting analysis presented on their website showing that the use of internal magnet rotors and a sine drive dramatically improvement on rotor heating. You may find this interesting.

http://www.shinetsu-rare-earth-magnet.jp/e/circuit/large.shtml

Thanks again.
SDK

 

[Skogsgurra]

sdk

It is interesting to learn that "motor controller consultants" think that it is difficult to generate a clean sine wave. Yes, that might be so. But when you do have a clean sine wave, then the reduced heating should not be questioned.

There are drive systems that produce sine with very little voltage distortion. The secret seems to lie in the high speed switching and a "lossless" soft switch that allows switching speeds in the couple of hundred kHz region and still having low loss in the transistors.

So the answer to your drive question for everyone is Yes, but on condition that: 1, the driving voltage is sinusoidal and 2, the back EMF is sinusoidal plus 3, the motor is synchronous. This will produce a sinusoidal motor current and a flux with no harmonics and no slip which translates into low rotor losses.

 

[sdk_imported]

Thanks Skogs. We have a 400Vdc bus and power levels that range from 1-5kW. My understanding is that we really need to use IGBT's and soft switching to keep the losses out of the controller. We have been using 39kHz IGBT's, which I thought we're about as fast as we could go with IGBT's. (Let me state here that my background is mechanical .

Do you think we should consider MOSFETS? Are IGBT's or FETS that switch at a couple hundred kHz considerably more expensive? Cost is a huge driver in this design.

I've asked two consultants about the benefit of using a sine drive and I've received two different responses. A motor consultant told me that there will be a significant reduction in losses regardless of the fidelity of the sine wave. The primary benefit comes from having three phases conducting at all times where the leading and trailing edges (normally seen when using a trap drive) are "stretched out" over a larger commutation window. The other consultant (a drive guy) suggests that fidelity of the wave form is critical at highspeeds and unless you get can generate very smooth sine waveforms there is no real benefit in using a sine drive.

I need some tie-breakers on this debate. Do you (or anyone else) have any thoughts about these competing views?

Thanks.

 

[Skogsgurra]

Yes sdk. You have no choice but using MOSFET. And they are not so bad in that power range. SiC transistors will be a lot faster and probably also more efficient, but they are not available yet - if we ever get to see them in our time. I have been involved with NFO - and still is to some extent - so perhaps I should not say too much about their solution. But, on the other hand, I know no other system that performs like theirs.

 

[jbartos]

Comment on UKpete (Electrical) Dec 7, 2003 marked ///\\jb, I'm not sure that Barden are saying you can do away with TDBs.
///Actually, Barden is using the TBDs. Visit the link for:
""The emergency touchdown bearings will support the shaft in case of complete power failure and allow the shaft to coast to a stop in a controlled manner.""
However, my remark below the posted link focused on:
""magnetic and touch-down bearings in one system
etc.""
Therefore, I do not see anything misleading or controversial in my posting.\\

 

[cbarn24050]

Hi, high voltage mosfets are quite lossy, (high conduction). High speed igbts are lossy (switching losses). The best approach is to use igbts, lower the switching speed and use a LC filter, or lower your bus voltage and use mosfets.

 

[Skogsgurra]

Yes cbarn,

And still, total efficienies above 95 percent (DC link to motor shaft) have been reached with MOSFET, soft switching to produce sinewaves for PM synchronous motors. Filters can be used in conjunction with IGBTs, but in automotive applications (fuel cells) the filter will be quite clumsy and not exactly weightless. It will also influence the dynamic properties, which are getting more and more important.

 

[cbarn24050]

Hi skogsgurra, suitable filter shouldn't weigh more than a few pounds.

 

[Skogsgurra]

cbarn,

A sine filter for a 5,5 kW inverter weighs in at 22 kilograms, which according to my conversion table, is more than 45 pounds.

 

[Skogsgurra]

Sorry, my informer now tells me that it was for a 11 kW inverter. There was none available for the 5,5 kW so we had to use the big one. The 5,5 kW seems to be about 16 kilograms.

 

[UKpete]

sdk, the increased airgap will certainly reduce the rotor heating. You could also reduce the slot gap, 3mm is rather wide for a motor of that size. It's a bit of a trial-and-error approach, ideally it's best to optimize the design before anything else whenever possible, more so with a cost sensitive application.

I don't think that skewing will help though, all it will do is skew the eddy currents.

jb you are correct, my misreading.

 

[sdk_imported]

Thanks UKpete. I have learned a hard lesson on this project. Start with functionality, then cut the cost! Thankyou and everyone else for the input. This has been an extremely helpful discussion for me.

SDK

 

[Skogsgurra]

Comment to posting Dec 3 2003:

"Suggestion: It appears that magnetically levitated bearings would eliminate lubrication problems and increase Mean Time To Repair (MTTR)"

Yes. It is very likely that a magnetic bearing will increase MTTR, but is that really what we want? Wouldn't it be a lot more interesting to increase MTBF? ;-)

 

[jbartos]

Suggestion to the previous posting: MTBF is more significant in applications where MTTR is insignificant. These are mutually related. Actually, MTTR can increase MTBF.

 

[Skogsgurra]

Come on jb!

MTTR = Mean time to repair = the time it takes to repair something after a failure.

MTBF = Mean time between failures.

How is your reasoning going?

 

[Skogsgurra]

Aw, come on jb!

MTTR = Mean time to repair = the time it takes to repair something after a failure.

MTBF = Mean time between failures.

How is your reasoning going?

 

[jbartos]

Comment on the previous posting: Yes, agreed. The longer MTTR may result in the longer MTBF, since the better quality repair is probably made.

 

[cbarn24050]

Hi skogsgura, take it from me it doesn't need to be anywhere near that weight. The filter only needs to remove the carrier frequency, 3 small chokes and some electrolitic caps is all you need.