SPM VS IPM Motor - And Field weaken, is it worth it

Emoto

10 kW
Joined
Jun 13, 2011
Messages
551
Location
Australia
So This a follow on discussion as seen here https://endless-sphere.com/forums/viewtopic.php?f=31&t=75494&start=275
im personally interested as im considering 2 inrunner motors one spm kv of 68 and a qs ipm kv 41 ,my main question is is fw worth the hassle .

As LFP has quoted ipm's with large stator iron can achieve only 2/6% hit on efficiency, and if this is the case then the low 41 kv ipm could work for me and sounds awesome. There is also a bonus reluctance torque advantage that ipm's can produce but is this a free lunch?? as you have to tune this in and im sure it would need extra amps ? but again if this is inclusive of the 2/6% then im excited.
maddin talks about reluctance torque https://endless-sphere.com/forums/viewtopic.php?f=30&t=93627&start=125.
 
Hi Emoto,

FW is not worth the hassle its mandantory to run a IPM machine at max efficiency! The Torque of a sync machine is always the sum of reluctance and magnetic torque.
T = 3/2 × Iq × (Kt + (Ld - Lq)×Id)
So if the difference of the Inductances Ld and Lq is significant then running the machine without negative d-current is a waste of Current. When utilising the reluctance torque u can reach higher torque with less current. There is a method called MTPA maximum torque per amper which calculates the needed Id for different load conditions providing maximum torque for the current.

About the 2-6% gain in eff.: on what basis was this increase calculated? Automotive IPMs normally provide more then 30% of their torque from reluctance. U can even build a motor the just uses reluctance part of the torque and get rid of the magnets and i dont mean those stepper like reluctance motors.

Regards

Alex
 
Is it true that SPM motors cannot make any reluctance torque?
Almost every paper or study says the potential is zero, but wouldn't a rotor with less than 100% magnet coverage allow reluctance torque?

Like the Surron motor which has steel between the magnets facing the stator (see yellow lines):

 
Hi Maddin,

the surron motor i would see as an rotor with inset magnets, thats why u are right, it should have significant reluctance torque.

Please make a short measurement. Please measure the phase inductance while turning the rotor. Mark down the lowest and the highest inductance. If u then take half of the inductance difference as (Ld-Lq) and take 70% of the nominal current squared, u can multiply those values to get a rough estimate for the maximum reluctance torque.

I would assume for this style of motor u get around 20% of the torque from reluctance.

Just to give u an other notion of reluctance torque and its relevance. Assume a rc motor like the 6374s ore simular. Thiese have an optimal timing value which u need to set in the bldc controller. This is basicly a kind of what here ist normally known as FW. Or in a more theoretic wording it compensates for the current phase lag and sets an Id <0
Making use of the little reluctance torque that is also there for surface mounted magnets.

Regards

Alex
 
Hi tecnologic !!

So If if i understand , if you dont utilize id / reluctance torque in an ipm that will sort of fight against iq / magnetic torque. making it less efficient.
So if we compare it with the 68kv spm at the same speed [kph ] the 41kv ipm would be past its not fluxed state @ 3000 rpm[4500 rpm ] would it have better efficiency.
So the MTPA is a algorithm That the controller has to be programed with.

This is the rotar of the motor i am considering looks like it has a bad magnet configuration as it doesnt have a good path/ space for reluctance between magnets, what do you think , would this affect the difference of the Inductance's .

Edited as i got my iq and id mixed up
 
I wonder if the controllers sold by QS are already tuned to run those motors for maximum efficiency? Also given the benefits why aren't IPMs much more common.
 
Hi Emoto,

its not that both torques fight each other, its more for what purpose the motor designer designed the motor.

The V-grove arrangement of the magnets is actually a very good one to reduce magnet mass to iron ratio with the side effect of a slightly higher max torque. But this is only the case when the motor is driven the way its ment to. Just by the looks of this rotor i would assume that using a current command of 2/3 iq and 1/3 id ist more efficient that id = 0 and the hole current command on iq.

Regarding MTPA there is a nice easy to understand (my opinon :)) app note:
Sensorless-FOC With Flux-Weakening and MTPA Motor Drives - Texas Instruments
http://www.ti.com/lit/an/spracf3/spracf3.pdf
 
larsb said:
Why do you think it's a bad geometry for the reluctance torque?
image.jpeg

Nice reference, there u have it the magnetic path through the q axis is designed to be explicitly bad to create a reluctance difference between d and q axis. Providing Reluctance torque aditional to the magnetic torque.
 
Grantmac said:
I wonder if the controllers sold by QS are already tuned to run those motors for maximum efficiency? Also given the benefits why aren't IPMs much more common.

There are no common controllers that can set IPM torque optimisation (rpm dependent field weakening) AFAIK. I think the APT96600 can but i haven't got it working yet due to no manual 8). To create the best performance each motor has to be tried on a dyno which mostly puts it outside of the ES market unless you're OEM or motor/controller kit seller.

A google search doesn't give much either but I guess that setting only one field weakening value like the common controllers do can give a lot also.
 
Even on a dyno u will not get the max Performance out of the motor because this is also totally temperature dependant. But there is a possibility to track the drives parameters at runtime and calculate the correct mtpa map always a new. Thats what ti is discribing in the paper.

But most oems in automotive are just running there motors to an extemsive dyno session with several temperatures and then the build a map out of this data just like it is done for a combustion engnie.

Setting a constant id aka FW all over the operatimg range ist defenitly better than leaving id at 0. For an IPM u could gain around 30% of max torque with the same current demand just by this.
 
tecnologic more home work :) Its funny that a from small question to myself Should i get one of these low 41kv ipm's spawns a lot of research
Here is a vid for those interested, [youtube]bZwLFpXhFbI[/youtube]
in the link'' texas instruments'' video they talk about pinching the flux affect , and thats why i assumed it would happen to that small gap between magnets . @ 17 MIN.

While i read your other posts. You missed my question , i edited it a bit.
So if we compare it with a 68kv spm [non fluxed ] at 4500rpm and both @ same speed/ load kph... would the 41kv ipm [ now past its not fluxed state @ 3000 rpm ] @ 4500 rpm, would The ipm have better efficiency than the spm.
Edit after reading a bit more id say yes :pancake:
 
tecnologic said:
larsb said:
Why do you think it's a bad geometry for the reluctance torque?
image.jpeg

Nice reference, there u have it the magnetic path through the q axis is designed to be explicitly bad to create a reluctance difference between d and q axis. Providing Reluctance torque aditional to the magnetic torque.

Maby, but isn't it a higher q inductance that you want compared to d?
This (pic) seems to be the way more "reluctance based" rotors are shaped.

1-Figure1-1.png
 
What inside the motor has a strong temperature dependency?

Rotor magnets lose some strength, phase resistance goes up but why would mapping against temperature be important/critical? For sensorless FOC position detection?
 
Emoto said:
tecnologic more home work :) Its funny that a from small question to myself Should i get one of these low 41kv ipm's spawns a lot of research
Here is a vid for those interested, [youtube]bZwLFpXhFbI[/youtube]
in the link'' texas instruments'' video they talk about pinching the flux affect , and thats why i assumed it would happen to that small gap between magnets . @ 17 MIN.

While i read your other posts. You missed my question , i edited it a bit.
So if we compare it with a 68kv spm [non fluxed ] at 4500rpm and both @ same speed/ load kph... would the 41kv ipm [ now past its not fluxed state @ 3000 rpm ] @ 4500 rpm, would The ipm have better efficiency than the spm.
Edit after reading a bit more id say yes :pancake:
Now u got it. I would agree just from feeling but only definite numbers will give u truth. Do u have datasheets of both motors?
 
larsb said:
What inside the motor has a strong temperature dependency?

Rotor magnets lose some strength, phase resistance goes up but why would mapping against temperature be important/critical? For sensorless FOC position detection?

For FOC sensorless algorithims the parameter changes are not that big of a deal. I calculated some time ago around 5degrees position estimation error with 25% parameter error. This maybe mor for less sophisticated algos but just to give u a number.

But my point was the temperature dependancy of the maximum torque per amp curve. If u have a look at the vid from ti @ 17min, there they have a simplified formular for the optimal id current. This is dependant on the diff of Ld and Lq and the flux both are temp depandant. But what then kicks u in the ass is that with increasing temp the flux is first increasing and at higher temp decreasing and the diff of Ld and Lq may act differently. And the optimal id current is totally dependant on the devision of the two. There are papers at ieee regarding the design of an IPM with a nice temperature curve.

But just for clarification we are now talking about the last 5% of the maximum torque. Most automotive companys will just run parameter observers and upd date the map according to them and for most hobbiests its just calculating a good curve and use it.
 
toolman2 said:
Maby, but isn't it a higher q inductance that you want compared to d?
This (pic) seems to be the way more "reluctance based" rotors are shaped.

1-Figure1-1.png
This is what u get. Inductance in q is higher because the field will not be saturated from the magnetic field. All the field lines go through the d axis and saturate the stator iron in this axis leading to a lower d inductance. And by that create more diff in Ld and Lq for the same magnets as with an spm config.
 
I assure you this is mostly incomprehensible to me, but it sounds as though once one person maps a particular motor/controller combination then it should apply to any serial products correct? Would a change in voltage significantly alter the curve?
 
Grantmac said:
I assure you this is mostly incomprehensible to me, but it sounds as though once one person maps a particular motor/controller combination then it should apply to any serial products correct? Would a change in voltage significantly alter the curve?

Voltage is totally out of the equation. Or in other words: Voltage is irrelevant for thiese calculations.
 
So the shape of the curve is dependent only on RPM, load and temperature?
That sounds a lot like tuning an EFI if I'm honest.

Are there constants for temperature which apply across all motors? Like does it shift the curve in an predictable way?
 
I don’t think it will be affected a lot by different voltage, anyway all motors have a rated voltage so the map would be done close to that voltage.

If we are lucky then Vasili will map the qs3000w motor in his controller development before i have built my dyno 8)
I’ll measure the inductances on this motor when i get back from christmas.
 
Inductances would be nice and the Kv or the Torque constant would be needed.

The parameter changes with temp are kind of predictable. Resistance is easy as cooper has a temp coefficient. But inductance is dependant on air gap and back iron and flux as well
 
tecnologic said:
Hi Emoto,
About the 2-6% gain in eff.: on what basis was this increase calculated? Automotive IPMs normally provide more then 30% of their torque from reluctance. U can even build a motor the just uses reluctance part of the torque and get rid of the magnets and i dont mean those stepper like reluctance motors.

Regards

Alex
Liveforphysics= wrote It's not uncommon to eat 2-6% motor efficiency to extend useful RPM powerband by +100% in a high iron% IPM rotor motor, where an SPM motor might be eating 10% efficiency to get +50-60% over base speed.

At some point rotor magnet strength being high is almost a penalty to motor performance, and you end up just wanting the highest temperature permanent magnet material, and then using minimal amounts of it in a mostly iron rotor to leverage reluctance torque.
iveforphysics didnt really elaborate on it.
 
Ah ok there is a missunderstanding. What we were talking about getting efficiency in terms of using all available Torque. Liveforphsics is refering an operating point past the torque optimum and there the ipm torque is decreasing faster with with increasing FW.
 
Back
Top