AC motors, multipoles, torque

[acmotor]

Richo,
This one is aimed at your thinking cap.

I know we have mused over the star delta options but ....
Can you fuse some braincells on the topic of the number of poles and switching poles options with 3PIMs ?

As we know 2 pole 3000RPM 50Hz xNm
           4 pole 1500RPM 50Hz 2xNm
           6 pole 1000RPM 50Hz 3xNm
           8 pole 750RPM etc.etc. up to hub motors 24 pole
What are the options (and real differences inside) with a 2pole that can be switched to 4 pole etc.
Seems there is little? difference in the rotor and the iron of the stator, just the windings placed for two or 4 efffective poles. Is this so ?

There are 2 speed motors on the market but they have poor torque/kg. Is this just because they are DOL and that a VFD could make more use of the winding assortment ?

I can picture a 4 pole's windings being arranged to look like a psuedo 2 pole.

When you have that sorted out, can I have my 4 pole modified to a 2pole / 4pole / 8pole please ?

[Richo]

Mmmm good one I'll have to think about that.

4-pole may not go up to an 8-pole as the phyical windings may not exist.

[TimF]

This does make good sense (as noted - albeit perhaps not as succinctly - here viewtopic.php?t=505&start=3).

Also I seem to recall reading an article about a motor wound for multiple poles (8 or 9 max iirc) where each pole had it's own power thyristors in the inverter - thus the number of effective poles could be changed on-the-fly by altering the switching pattern of the inverter...just need to track down the reference now

[acmotor]

Yes TimF, I noted that you had raised the issue with that thread, however it had wandered off so the thought was to try again !
Anything to get rid of gearboxes !!

In particular, I was interested in the possibilities with off the shelf, low cost industrial motors as the high tech items are either notyetmadium, unobtainium, can'tafforium or just inyourdreamsium.

We have the options of re-connecting to windings internally or rewinding the motor completely. Some 4 pole motors present 12 wires (each pole end) to the terminal bock.
With Richo's inverter design expertise a multi pole winding VFD could be the go.

Converting a 4 pole to 2 pole seems fairly direct depending on the access to the internal windings. (rough sketch !)

[woody]

I had a quick bash at this.

For data I used the CMG largest aluminium motor frame (160L) which can come as a 2,4,6, or 8 pole motor of 18.5/15/11/7.5 kW respectively.

This is a plot of breakdown torque against RPM of 400V motors with 400V peak controller voltage. Assumes fixed VF ratio drive below sync speed (i.e. no low speed torque boost like the danfoss)

First graph assumes breakdown torque declines on the square of the frequency once saturation voltage can't be met.

Second graph assumes constant power after saturation voltage can't be met.


In either case, 2 + 4 pole windings look the most useful, 8 pole is of marginal benefit at low RPM.

I don't know what compromises you get with a multipole wound motor (e.g. higher weight, lower power, less efficiency)

I think star/delta switching looks more promising

When I get some time, I will try and add pole switching and gear changing to my spreadsheet and see what looks best.

cheers,
Woody

[woody]

I did this again with 2/4 Pole Star/Delta Switching.
2S = 2 pole Star 400V
4D = 4 pole Delta 230V

I also added in 48kW and 59kW curves to show the limit of a Danfoss 5042 / 5052 controller.



This shows with a 5042 you only get an advantage with 2 pole above 4000rpm, 3150rpm with a 5052.

Again this is with a big motor frame (a4x4 kiwi size 160L = 18.5/15kw 2/4 pole).

With a smaller frame / motor power, the advantage would kick in at a lower RPM.

cheers,
Woody

[Johny]

All the brands of motors that I've reviewed so far indicates that multiple pole-wound motors have less torque (lots less) than their equivalent one-set-of windings motor. I assume (I know - ass, u, me) that this is due to more wire thereby having to be smaller gauge. I have enough trouble trying to get a low weight motor without more wire weight when star/delta may do it. Still don't like switching between them (star/delta) but there should be a way.
Woody - could you trouble yourself to graph a 4 pole star vs delta with a straight line to indicated rated torque as well please? I gather I just place a torque limit somewhere up the graph to show controller limitation due to current limit? BTW I'm glad you are doing this - it's really useful as I'm still trying to figure if the 11kW is enough for moderate driving.

[woody]

I've switched to our favourite CMG 132MB-38 frame (11kW 4 pole) for you Johnny:

Looks like the 5052 can take full advantage of the motor in 4 or 2 pole 230V delta.
5042 only misses a few small corners off between 2100-2900 and 4900-5600 RPM.

The 5042 + 5052 lines are Max(motor curves;X kW power curve) which is why they follow the motor curves.

Edit: From Tuarn's experience I don't think this is correct: [Not all this torque is usable, I think Tuarn's Danfoss limits itself to 85% of breakdown torque to avoid "losing" the load.]

Johnny, if you send me data on drag, frontal area, weight (less motor + batts), batts, controller, diff, tyre size, I'll send you what I think your car will do with these motors.

These graphs don't show the advantage of Star windings at low speed: the delta wound motor needs about 75Amps for 85% torque, which doesn't give the controller much headroom for "Torque Boost" which it does by boosting the voltage -> current -> field -> torque. The star wound only needs sqrt(1/3) of that (44 Amps), giving more headroom for controller trickery.

cheers,
Woody

[woody]

Notice above that the 2 pole only gets better at about 3800rpm, which for my car is about 90kph.

If I limit the torque to 85% of the breakdown (as I think the danfoss does), then the gaps between the 5042 and the motor curves disappear. I.E. The 5042 is just the right size for an 11kW EV motor

I still want to do gearbox numbers, including efficiency, I think I heard efficiencies of 85% for low gears, 90% for direct drive, 80% for overdrive.

Another thing to explore is how high a frequency should you wind for: if winding 230V/400V is better than 400V/690V, where is the sweet spot?

cheers,
Woody

[Rob M]

Woody, I noticed that you used lower Kw output motors in your graphs of torque Vs Rpm for 2,4 6 and 8 pole motors. Try doing the same with similar power output motors and you will see the torque advantage. you still need the power for top speed anyway. To eliminate the gearbox, an 8 pole motor will give you double the torque at half the revs of a 4 pole, closer to what you want.

[Johny]

Hi woody - I have to do the drag figures when I get it going again - rear brakes are stuck due to lack of driving over last 2 years. (I've bought the kits). I also have to measure front area which I haven't done yet. It all should be similar to the Cortina anyway - the Vogue is a similar sized and shaped car. I have 2 days at end if Victoian school holidays slated for work on the car. I'll get figures then. Star/Delta switching (and pole switching if that is worth it) will be dangerous for the drive so it is going to have to have some solid advantages to go that way.

[Johny]

woody. What line does the 4 delta join (yellow line)? The motor must surely saturate earlier if left in 240V mode (delta) so there has to be a disadvantage in NOT switching to star at lower revs due to not being able to boost v as much (when in delta). I can't see the disadvantage in the graph.

[acmotor]

Danfoss does not limit torque to 85% of breakdown. Since the controller is 48kW max and the motor nominal 11kW it just pushes out according to the settings you make of max torque and current. You can be conservative and select 300% or push out past breakdown. This occurs out over 400% at low revs.
The consequence is extra current for in fact a reducing torque as the standard torque curve shows. The motor is still pulling hard but the efficiency is obviously shot.

It is all quite smooth except for the battery amp meter being higher than expected for the torque.... and the motor warms up.
It feels a bit like a slipping clutch in an ICE. In the same way you feel a slipping clutch under power (you know, worn plate or oil on it) the VFD has lost the ACIM. Back off the accelerator a bit and it gathers it up again. You can hear the magnetic field rotation go over the motor speed.
The VFD seems not to know it has lost the motor, unless an encoder is fitted, when you are pushing the boundaries like this.

Richo will come up with a natty circuit to always hold the motor / VFD at max torque when you apply max accelerator - independent of speed.

[Richo]

The circuit doesn't change just the firmware.

[acmotor]

You can't get out of it that easily Richo.
We want full working details right now.
I'll need a 100kW intelligent controller for the Rodeo.

[Richo]

Can't talk building - need controller for BMW

[woody]

Woody, I noticed that you used lower Kw output motors in your graphs of torque Vs Rpm for 2,4 6 and 8 pole motors. Try doing the same with similar power output motors and you will see the torque advantage. you still need the power for top speed anyway. To eliminate the gearbox, an 8 pole motor will give you double the torque at half the revs of a 4 pole, closer to what you want.

Hi Rob,
I used the specs for same frame (CMG 160L-42) with different windings as I figured that was more realistic. The manufacturer offers lower power in the same frame with more windings -- I guess because more windings take more room...
cheers,
Woody

[woody]

woody. What line does the 4 delta join (yellow line)?

The 4D line is buried under the 5052 line all the way up to 3900 rpm -- looks like the way to go.

The motor must surely saturate earlier if left in 240V mode (delta) so there has to be a disadvantage in NOT switching to star at lower revs due to not being able to boost v as much (when in delta). I can't see the disadvantage in the graph.

Yes, I left torque boost off the graph as I'm not real sure what shape it is -- ignoring the inefficiencies of over magnetising, you get starting torque of 418 in star vs 240 in delta, but the advantage drops off at 10kph. So it may be worthwhile, or in our case we may just be spinning our skinny tyres

[woody]

Danfoss does not limit torque to 85% of breakdown. Since the controller is 48kW max and the motor nominal 11kW it just pushes out according to the settings you make of max torque and current. You can be conservative and select 300% or push out past breakdown. This occurs out over 400% at low revs.
The consequence is extra current for in fact a reducing torque as the standard torque curve shows. The motor is still pulling hard but the efficiency is obviously shot.

I was thinking about actual breakdown torque, not nominal.

I think danfoss gives more voltage than VF -> more reactive current -> more field -> more peak/breakdown torque than nominal.

This means you can get more torque than breakdown, which would be impossible without this voltage boost (or some other wizardry)

It is all quite smooth except for the battery amp meter being higher than expected for the torque.... and the motor warms up.
It feels a bit like a slipping clutch in an ICE. In the same way you feel a slipping clutch under power (you know, worn plate or oil on it) the VFD has lost the ACIM. Back off the accelerator a bit and it gathers it up again. You can hear the magnetic field rotation go over the motor speed.
The VFD seems not to know it has lost the motor, unless an encoder is fitted, when you are pushing the boundaries like this.

Hmmm, that's not as scary as I was expecting. I guess you're just losing some of the torque, you don't lose all of it. But it's not something you'd want to happen at the wrong time.

The encoder fixes all this? Do you get the same/more torque with the encoder?

Richo will come up with a natty circuit to always hold the motor / VFD at max torque when you apply max accelerator - independent of speed.

Yep, it looks to me like a "Richo Industries" 100kW controller is about right for a 22kW ACIM, since a 5042 is about right for an 11kW motor.

cheers,
Woody

[acmotor]

The encoder allows the VFD to go promply back to sync with the motor as power is backed off.
The internal sensorless control algorithm is a little slow at recovery when the motor has gone past pullout.
The encoder also allows the VFD to apply more torque boost at low speed as it actually knows what the rotor is doing when it is belting it around the ears.
The encoder also allows you to stop / reset / power on /off when in motion and have instant recovery to motor speed. The other option is the VFD's 'flying start' but this takes a few seconds to 'find' the rotor.
Having said all that, the system also works without the encoder. Just not quite as polished.

[Johny]

My experience with my bench-top setup for the Lenze VFD pretty much tracks what acmotor says of Danfoss. Flying start without encoder is a bit frightening without a load - but it does find it within about 0.5 seconds. That is my concern with sequencing a star/delta switch at full throttle. Lenze says the same thing - that flying start doesn't even have to be selected if an encoder is used.

[woody]

The encoder allows the VFD to go promply back to sync with the motor as power is backed off.
The internal sensorless control algorithm is a little slow at recovery when the motor has gone past pullout.
The encoder also allows the VFD to apply more torque boost at low speed as it actually knows what the rotor is doing when it is belting it around the ears.
The encoder also allows you to stop / reset / power on /off when in motion and have instant recovery to motor speed. The other option is the VFD's 'flying start' but this takes a few seconds to 'find' the rotor.
Having said all that, the system also works without the encoder. Just not quite as polished.

Awesome info. You can't argue with the real world .

The low speed torque boost is the hardest to predict since it involves magnetic saturation curves of the motor, which aren't easy to express let alone write on the sepc sheet

But I know the boosted torque is somewhere between the nominal breakdown torque and what the max active current of the controller would generate. (For Red Suzi's ABB I get nominal breakdown of 3 * 72.5 = 217.5Nm, Delta (400V) current limit of 397Nm, Tuarns reported 337Nm)

I think if Tuarn tries Red Suzi in 690V Star config, my formula says 687Nm, so maybe he should see 585Nm of torque, up to about 10kph, which should take about 3/10ths of a second on level ground.

Star will run out of puff really early though, Suzi will get to 40kph in ~6 seconds, 50 in 14 secs, but only get to about 57 kph.
cheers,
Woody

[Johny]

Hey woody. I hadn't even thought of using star in an unmodified 415/720 V motor. That means in star it would use way less motor current at low speeds and not require the controller to be as 'grunty'. Lower capacity VFDs are significantly cheaper and would only need to be twice rated motor power. The extra cost would be in the star/delta contactor - AC so not too expensive - and the drive enable delay. Lenze can maybe do the enable delay as a user function block but it's a DC low-power delay so not complex anyway.
I.e. Disable gate drive (coast) at the same time as energise change-over contactor (contactor will be way slower then drive disable so is safe), delay, then re-enable gate drive (start). Maybe I don't have to delve into the power section of my Lenze after all (I still will - I can't help it).

[woody]

Hey woody. I hadn't even thought of using star in an unmodified 415/720 V motor. That means in star it would use way less motor current at low speeds and not require the controller to be as 'grunty'. Lower capacity VFDs are significantly cheaper and would only need to be twice rated motor power. The extra cost would be in the star/delta contactor - AC so not too expensive - and the drive enable delay. Lenze can maybe do the enable delay as a user function block but it's a DC low-power delay so not complex anyway.
I.e. Disable gate drive (coast) at the same time as energise change-over contactor (contactor will be way slower then drive disable so is safe), delay, then re-enable gate drive (start). Maybe I don't have to delve into the power section of my Lenze after all (I still will - I can't help it).

I plugged your controller (16kw/30A * 180%) and CMG 132MB into my cortina (165/70R13 + 4.4:1 Diff) spreadsheet and figured:
1. Rewound 230V Delta is no good for lenze, it's current limited the whole way: 0-60 in 13s, top speed 72, 400m in 23 seconds.
2. Standard 400V Delta is OK 0-60 in 8s, 0-100 in 25, 400m in 22. Top speed 114. Better than the Cortina 1200 ICE
3. Star 690V up to 16kph almost doubles your torque, still better all the way to 30kph, saves you a second. 0-60 in 7s, 0-100 in 24, 400m in 21. Top speed 114.

So:
0-16: Star 690 is 75% more torque
16-30: Star 690 is better, flat power curve
30-35: Star 690 + Delta 400 the same, flat power curve
36-107: Delta 400 is better, flat power curve - insufficient voltage for star
107+: Delta 400 is better, V/F limited

Also 14kW of drag at 100kph, I don't think you'll be cruising at much faster than that.

cheers,
Woody

[woody]

I finally got around to doing numbers on AC + gearbox.

I found with my planned setup:

Yes, quicker by ~ 1 second to accellerate IF you ignore the time it takes to change gears, and the extra cost of upgrading your gearbox/diff to cope.

It may be worthwhile for a 2 pole high RPM motor.

I think rewinding for approx 415V star / 230 delta is a better deal.

I used 85/90/95% efficiency for over/direct/under drive gears. + 93% for the diff.

cheers,
Woody