PDA

View Full Version : why not squirrelcaged (asynchronous) inrunners?


Moxus
05-14-2009, 03:14 PM
hi!

i was just sitting here, rewinding my hacker b40 (yet another time), and was thinking of why it uses permanent magnets in the rotor.
much higher field strength would be possible with a squirrelcage, and hence a more efficent motor at high loadings, and hence, much friendlier on the windings, and hence, i wouldnt have to rewind it after every other flight, and hence, the world would been a much better place.

is it because field strenght wouldnt be constant, and therefore wouldnt the motors have a fixed kv rating?
and since frequency isnt fixed either with our regulators, the motor would simply spool up until some load stopped it, or it self destructed?
why not use squirrelcaged motors with a frequency controlled regulator?

DKNguyen
05-15-2009, 08:07 AM
1. kV not constant, as you said.
2. Torque curves are not linear and the shape itself also varies from motor to motor depending on design requirements.
http://www.wattflyer.com/forums/attachment.php?attachmentid=100976&stc=1&d=1242371825
3. Both 1 and 2, as well as a few other things cause speed control of induction motors to not be straightforward and requires you to know the motor parameters, as well as a lot of intensive math. Here's a whiff:
http://www.ece.ualberta.ca/~knight/ee432/trim/induction_machine_transient_analysis.pdf
4. Like you said, an induction motors require variable frequency drive. This in itself is really just a software change for our current ESCs but...
5. Current ESCs are trapezoidal six-step drive ESCs. Motor phases are either positive, negative, or floating. But induction motors require a sine wave drive. These are a MAJOR pain to make and require much more complex drive circuitry and commutator sensing circuitry.
6. Also, sine wave drives require switching frequncies MUCH higher than the frequency of the sine-wave (much higher than current ESC frequencies). Higher switching frequency introduces more losses, more noise, more transient problems. Not only do you have more complex circuits and complex components, you also need larger components that can handle dissipate more heat, and it is harder to physically layout the ESC because of the noise and transient problems.
7. Sine drives must know the commutator positions at all times since the drive waveform is varying at all times. If you used a sensor, it would not be one of the typical hall sensors you see on "sensored brushless motors". These sensors only tell you where the nearest phase is. You need a sensor that tells you exactly where all phases are at all times, like a high resolution encoder. And if don't want to use a sensor, you need much more involved BEMF sensing schemes which require a lot of complex math that models the motor AND the motor parameters.
8. The ESC also requires all the circuitry to drive and adjust current through the field winding (that's basically a variable BEC right there).
9. It goes without saying that all the motor parameters that are required for the math in the steps mentioned above are neither cheap nor easy to measure.
10. Permanent magnet motors are simpler to build (making and mounting magnets is easier than field windings)
11. And finally...permanent magnet motors are smaller (magnets are simpler and smaller than field windings), stronger (no armature reaction and higher power density), lighter (smaller motor means lighter motor), and more efficient (no field current to supply) than field windings. I am not sure where you got the idea that field windings were stronger.

THe reason they use field windings is because large magnets are expensive. And during the manufacturing of large motors with large magnets, it is difficult and dangerous to manuever the magnet around the steel that the rest of the motor is made of . Have you ever gotten between the magnet and steel stator or casing of a motor the size of a room? The second, very minor reason is that you do have more flexibility in the way you can control the motor because you can the field strength whereas in a PM motor you cannot (though you do take a hit in performance compared to an equivelant permanent magnet motor).

pilotpete2
05-15-2009, 04:34 PM
I think that a true ac VFD would be quite a bit more expensive than a simple sensorless BLDC controller, more than just a software change involved.
Both a BLDC motor and ac induction motor have stator windings, but the induction motor would be more expensive due to the induction rotor being more expensive to make than a simple permanent magnet rotor.
I think that the deciding factor here is cost and power to weight considerations.
Pete

Moxus
05-16-2009, 02:50 AM
on/off/floating in current escs is voltage only. not current. very important detail.
current escs is pwm sine wave modulators, and they do indeed model a sine wave CURRENT, by... yes... turning VOLTAGE to either on, off or floating.
now, the motor doesnt give crap about voltage, for the motor its the current through it that counts.
so i dont believe thats going to be a problem, neither for squirrelcage motors.

and why i think they will be more efficient at high loadings is because permanent magnets are only so strong, and they dont deal with temperature very well.
magnetism induced from current through a wire has no theoretical limit, and practically, it can be something like 50 to 100 times stronger than a permanent magnet.
so yes, definetly, a squirrelcage will easily overdo the magnetic strenght of a permanent magnet rotor.
and as we know, stronger magnets is better efficiency for a motor.
plus, of course, you can run the squirrelcage until its red hot glowing, and it wont mind. opposed to the neodymium magnets, where some even surrenders as low as 80 degrees celcius.

DKNguyen
05-19-2009, 07:44 AM
on/off/floating in current escs is voltage only. not current. very important detail.
current escs is pwm sine wave modulators, and they do indeed model a sine wave CURRENT, by... yes... turning VOLTAGE to either on, off or floating.
now, the motor doesnt give crap about voltage, for the motor its the current through it that counts.
so i dont believe thats going to be a problem, neither for squirrelcage motors.

THat's not true. The ESCs on the market right now are not sine wave modulators. They are trapezoidal six-step drives. When they drive the motor, the motor sees a trapezoidal current and produces a trapezoidal back EMF. Just hook one up to an oscilloscope and a current shunt. You will see a trapezoid waveform with six linear segments for voltage and current, not a sine wave. It might roughly approximate a sine wave, but it is not a sine-wave.

The voltage does matter because if you apply a constant voltage to the winding inductance the current will ramp- not curve like a sine wave. The voltage might not have to look like a sine wave but it definately has to make the current look like a sine wave. For this to happen you have to vary the PWM duty cycle sinusoidally. Since the voltage is a square wave and can only be on or off, it is always at the wrong voltage to get the current to be where it should be in the sine wave. So the voltage has to adjust itself fast enough so that the current in the winding inductance never overshoots or undershoots too far from where it should be for that part of the sine wave. That means the PWM frequency has to be high enough so that the current cannot follow the voltage fast enough in the winding inductance. The result is that the PWM frequency for the voltage square wave has to be much higher than the frequency of the current sine wave which is much higher than for an equivelant trapezoidal six step BEMF ESC (which has a switching frequency of around double the frequency of the BEMF- each period of the trapezoidal waveform has a positive and negative square pulse).

You cannot simply increase the frequency- we are talking about frequencies that are at least 5-10x times higher than what ESCs currently use. That's an ESC that's at least 5-10x bigger for the same power level. The higher frequency means more heat, more noise, and more inductive spikes. That means more powerful MOSFET drivers to switch the MOSFETs faster, larger MOSFETs, more shielding and attention to RFI design, and more circuitry and careful design to protect against inductive spikes. Plus all that new circuitry and complicated software to actually control the motor speed properly. Not only is the ESC bigger, it's more complicated, and more sensitive to design flaws.

Even with that high frequency, the motor's inductance is often not enough to properly filter out the higher level harmonics of the square wave so you need to add extra filtering- one power inductor for each phase that is able to tolerate the same power level as the phase itself. Those are huge inductors. Sometimes you even need power capacitors for each phase to help with the filtering. That's about 4x more giant capacitors than what ESCs have right now.


and why i think they will be more efficient at high loadings is because permanent magnets are only so strong, and they dont deal with temperature very well.
magnetism induced from current through a wire has no theoretical limit, and practically, it can be something like 50 to 100 times stronger than a permanent magnet.
so yes, definetly, a squirrelcage will easily overdo the magnetic strenght of a permanent magnet rotor.
and as we know, stronger magnets is better efficiency for a motor.
plus, of course, you can run the squirrelcage until its red hot glowing, and it wont mind. opposed to the neodymium magnets, where some even surrenders as low as 80 degrees celcius.
Not true- electromagnets do have a theoretical limit on their field strength - they saturate. After a certain current level the magnetic strength will level off. The magnetic field from the current in the coil aligns the paramagnetic domains in the core to point in the same direction, boosting the strength of the magnetic field for the same amount of power. But those paramagnetic domains that produce the magnetic field don't come from nowhere. To increase the maximum possible strength of an electromagnet, you need more domains to align which means you need more iron. Also, as electromagnets get hotter it takes more current to maintain the same strength (you are basically fighting the Curie effect, remagnetizing the core as it demagnetizes) which means more current for the same strength which means less efficient. So electromagnets will "demagnetize" if the temperature is too high but the current is too low.

But what if you use an air core electromagnet that doesn't saturate? So all of the magnetic field is coming from your current flow alone? Well now your magnet doesn't have a ceiling on strength, but it's also much much less efficient. So you'd need a much more power just to get the same strength as if you had a core.

I mean...have you ever seen an electromagnet the same size as Neodymium magnet that was even remotely as strong? And how much current was it using? Compare that to the fact that the neodymium magnet needs no current.

ron_van_sommeren
05-19-2009, 02:15 PM
Sinus-commutation for RC
www.rcgroups.com/forums/showthread.php?t=788065&highlight=sinusleistungssteller (http://www.rcgroups.com/forums/showthread.php?t=788065&highlight=sinusleistungssteller)

Trapezoid commutation & PWM pictures/shots:
www.aerodesign.de/peter/2001/LRK350/index_eng.html
-> Why does the Torquemax rotate so slowly and so forcefully
and/or
-> SPEEDY-BL self made brushless controller

Vriendelijke groeten ;) Ron

Moxus
05-19-2009, 03:00 PM
well, i spoke to my teacher about it, and he said the escs were sine wave modulators, so now i really got to test to see whos right.
and electromagnets do not have a theoretical limit.
surrounding ferromagnetic materials (a core) will reach saturation, yes.
but the windings themselves will not. and as long as the windings isnt reaching a limit, then neither is the electromagnet as a complete unit either,
the strongest electromagnets i know of is 62 tesla of field strength, and thats about 50 times stronger than the saturation of any known ferromagnetic material. and its not a nitrogen chilled super hightech megamagnet.
its a simple coil of massive copper, drivien by a homopolar generator for massive currents, and cooled by water.
of course, this isnt something thats practical for our motor applications, but just to illustrate the point. you can get well beyond saturation with realtively simple means.
and anyway, the saturation of a iron core is still more than what the strenght of neodymium magnets is, so you can gain field strength by switching to squirrelcage.
and at the point where the iron core of the rotor reach saturation, the windings themselves will still increase field strenght if the motor lags sufficiently enough behind its magnetic rotational speed.

ron_van_sommeren
05-19-2009, 07:22 PM
Once the iron is saturated, it will behave like air, magnetic 'resistance' (=reluctance) will not go up anymore and flux lines are all over the place instead of in the iron.

Your teacher is thinking of industrial drives and inverters, he's probably not into RC. Point him to this overview of diy brushless controllers and he will see the light ;)
diy brushless ESC designs (http://www.rcgroups.com/forums/showthread.php?t=140454)

Excellent active ESC design discussion:
www.rcgroups.com/forums/showthread.php?t=200567 (http://www.rcgroups.com/forums/showthread.php?t=200567)
www.yahoogroups.com/group/osmc

Motor design, magnetism and controllers:
www.consult-g2.com
-> course

If you have money to burn, get this book:
www.eece.maine.edu/motor

Vriendelijke groeten ;) Ron

DKNguyen
05-19-2009, 07:43 PM
well, i spoke to my teacher about it, and he said the escs were sine wave modulators, so now i really got to test to see whos right.

DId you say ask about induction motor? Or brushless motors? All induction motor drives are sine-wave because they have to be. Most brushless motor drives are six-step whenever they can be because it's simpler to do that than since-wave. This is especially true on smaller drives. When they get bigger and harmonics and spikes become a problem then they have to go sine-wave. I have a professor too and if you ask him, that's what he says. He is also usually talking about industrial drives so when he says "big" he means MASSIVE (like freaking megawatts). And when he says "small" he is still referring to things that are way bigger than anything seen in our hobby.


and electromagnets do not have a theoretical limit.
surrounding ferromagnetic materials (a core) will reach saturation, yes.
but the windings themselves will not. and as long as the windings isnt reaching a limit, then neither is the electromagnet as a complete unit either,

THat's what basically what I said earlier. I did reword my first sentence incorrectly though. I meant to say there is a theoretical limit for the core's contribution to the magnetic field rather than the coil. I was first thinking the entire electromagnet was limited by the core but then I remembered air core electromagnets and forgot to go back and fix my first sentence. But that imposes a practical limit (if not a theoretical limit) on the field strength for applications that don't have good cooling or huge power sources.

THe core greatly enhances the efficiency of the electromagnet but has limits (saturation). After saturation (or for air core electromagnets that have no core), the efficiency drops a lot and all the extra increases in field strength have to come directly from increasing the coil current flow rather than a core to focus the flux lines. Then you start need much more power for any increase in field strength. So even though there is no limit on the magnetic field being directly generated by the coil current flow, it takes way more power than most applications have to get a useful strength out of an air core electromagnet (especially mechanically applications) than is practical.


the strongest electromagnets i know of is 62 tesla of field strength, and thats about 50 times stronger than the saturation of any known ferromagnetic material. and its not a nitrogen chilled super hightech megamagnet.
its a simple coil of massive copper, drivien by a homopolar generator for massive currents, and cooled by water.

THat's not saying very much about the strength of an electromagnet vs rare-earth. Of course it's going to be strong- it's massive with lots of current flowing. How efficient is it? It sounds like they are directly generating the magnetic field through current flow by forcing tons of current through the coil, rather than relying on a saturation-limited core to increase the flux density. The power required to gain useful magnetic strength would be impractical for most applications.


of course, this isnt something thats practical for our motor applications, but just to illustrate the point. you can get well beyond saturation with realtively simple means.

Yes, I agree that only the transformer core is limited, not the coil. It's a simple means but too costly to be of practical use for most things.


and anyway, the saturation of a iron core is still more than what the strenght of neodymium magnets is, so you can gain field strength by switching to squirrelcage.

Do you have any literature to back this up with? Because you can easily find transformers larger than neodymium magnets that are saturating and the magnet they produce is much weaker.


and at the point where the iron core of the rotor reach saturation, the windings themselves will still increase field strenght if the motor lags sufficiently enough behind its magnetic rotational speed.
I agree. But once you saturate the core you get greatly diminished returns in field strength for any power increase. The power cost to drive an electromagnet like that is often too high for any application that doesn't have crazy cooling and a power nuclear generator sitting nearby. Particle accelerators and ship-mounted railguns come to mind.

Moxus
05-19-2009, 07:46 PM
thanks a lot ron, thats some interesting info the chew up =D

nguyen, neodymium magnets cant be stronger by the simple fact that they arent pure iron.
pure iron is the strognest ferromagnetic material there is, and you just cant make a permanent magnet out of it.
as a matter of fact, its the iron that gives neodymium magnets their magentism.
the other elements, neodymium and boron is there just to give it stability.
fair enough, the neodymium magnets does contain a LOT of iron, and is nearly as strong as a saturated iron core would be, but its not as strong, and it will never ever be stronger.
the chemical forumla is Nd2Fe14B. as you see, a lot of iron. the boron and neodymium is there just to give it some stability, as pure iron isnt very stable as a permanent magnet. but it is stronger.

DKNguyen
05-19-2009, 08:14 PM
thanks a lot ron, thats some interesting info the chew up =D

nguyen, neodymium magnets cant be stronger by the simple fact that they arent pure iron.
pure iron is the strognest ferromagnetic material there is, and you just cant make a permanent magnet out of it.
as a matter of fact, its the iron that gives neodymium magnets their magentism.
the other elements, neodymium and boron is there just to give it stability.
fair enough, the neodymium magnets does contain a LOT of iron, and is nearly as strong as a saturated iron core would be, but its not as strong, and it will never ever be stronger.
the chemical forumla is Nd2Fe14B. as you see, a lot of iron. the boron and neodymium is there just to give it some stability, as pure iron isnt very stable as a permanent magnet. but it is stronger.
That makes sense. The transformer cores must be just starting to saturate then rather than completely saturating since you want it to maintain waveform integrity so you never want it too far into the saturation region. I can accept that a completely saturated iron core might be stronger than a rare-earth magnet. To get it to the point where it can beat a neodymium (both in overall terms and field strength), it still requires more power and cooling than is practical for most things though, and the hardware and software are still a fair deal more complicated than your typical hobby ESC, as is how to get rotor position feedback.

Last I heard, they were trying to replace the field wound electric motors (I am not sure if they were synchronous or induction) on submarines with very large permanent magnet motors which they never tried before because large magnets are expensive and it is hard to position a magnet properly during manufacturing since it gets attracted to the steel in the motor. Those definately use sine-drive whether or not it's an induction, synchronous, or brushless motor. THose things have like more than 12 phases. Could you imagine designing a sine-driven 12-phase driver?

Sinus-commutation for RC
www.rcgroups.com/forums/showthread.php?t=788065&highlight=sinusleistungssteller (http://www.rcgroups.com/forums/showthread.php?t=788065&highlight=sinusleistungssteller)

Trapezoid commutation & PWM pictures/shots:
www.aerodesign.de/peter/2001/LRK350/index_eng.html (http://www.aerodesign.de/peter/2001/LRK350/index_eng.html)
-> Why does the Torquemax rotate so slowly and so forcefully
and/or
-> SPEEDY-BL self made brushless controller

Vriendelijke groeten ;) Ron

That sine-wave driver is pretty damned neat.

Moxus
05-20-2009, 12:00 AM
well, according to my teacher, running cores up to 1,5 teslas is fine, efficiencywise. for absolute maximum, you can go beyond. something like 1,7 to 2 teslas.
the n48 grade of neodymium magnets, wich is among the strongest there is, has a remnant static magnetic field of 1.38 teslas.
and i dont think motors use these strongest grades, because they are notoriously unstable to temperature. some even have curie points below 80 degrees celcius, and thats not very practical for a motor who routinely gets hotter.
so im still pretty sure a squirrelcage is more efficient.
and om not sure the losses of going beyond 1,5 teslas is very significant either, because there isnt much need for the magnetic field to change anyway within the rotor. the rotors mechanical rpm doesnt lag much behind the rotors magnetic fields rpm.

sensing the rotors position is not something you need to do. you need to sense the position of the rotors magnetic field, and thats what our escs already do.
but as you guys say, since induction motors apparently dont work very well with non-sinusiodal power, i guess thats the reason why they arent used then.

DKNguyen
05-20-2009, 12:35 AM
I'm kind of wondering how expensive and how bigthat sine ESC that Ron posted are.

EDIT: 500-700 Euros *phew* There don't seem to be any dimensions given for physical size or weight.

ron_van_sommeren
05-20-2009, 12:46 AM
... That sine-wave driver is pretty damned neat.It's also pretty damned $$$ :D For now that is, some Japanese RC'ers are also working in sine-wave drivers. The shape of things to come?

Did you check the vids I posted there, especially the Gee Bee racer? Soo smooooooooth :)

Too bad Christian Lucas (http://www.rcgroups.com/forums/member.php?u=14492)is not a member of this forum. He is the one responsible for the first diy outrunner motor design (Elektro Modell articles (http://www.aerodesign.de/peter/2001/LRK350/index_eng.html) at bottom of page) and from that, outrunner motor popularity in general. He works for www.magnetmotor.de (http://www.magnet-motor.de/en/home/) (fuel cell research and motor development). Check out Christian's messages (http://www.rcgroups.com/forums/search.php?do=finduser&u=14492) on RCGroups, good reading.

... There don't seem to be any dimensions given for physical size or weight.Could not find it either but component size should give some indication

Vriendelijke groeten ;) Ron

http://www.sinusleistungssteller.de/gfx/SLS/SLS-Intro.jpg

Moxus
05-20-2009, 12:51 AM
lol. that is a bit expensive, i admit :P
but... it IS a nice piece of equipment =)

DKNguyen
05-20-2009, 01:01 AM
sensing the rotors position is not something you need to do. you need to sense the position of the rotors magnetic field, and thats what our escs already do.


Sensing the position of the rotor is the same as sensing the position of the magnetic field via BEMF since the rotor has the magnets (electromagnets) for outrunners (inrunners). The sensorless scheme would have to be quite a bit more complicated than what six-step ESCs are using since those only need to know when to commutate so they only sense the BEMF zero-crossing. For Sine-wave ESCs to drive induction motors, they need to know where the rotor is at all times.

I'm not sure how well it would work if you sensed the zero-crossing in an induction motor sine-drive, measured the time between them to determine the period, and then blindly vary the voltage PWM duty cycle sinusoidally according to that period hoping it would produce a current sine-wave. I think it would be really crude. It'd be breaking the the Nyquist criterion. You'd have no idea what the actual phase current waveform actually looked like. It would probably work a lot better if the load was constant and the motor parameters were known so you could model the motor in software. I have been under the impression induction motor sine-drives constantly monitor the current waveform and update their voltage output to keep the current waveform like the desired sine-wave at all times, not just at the BEMF zero-crossings.

But I think it would work for permanent magnet brushless motor. They run just fine off a crappy trazpezoidal current waveform so you could probably interpolate the sine-wave in between zero crossings- a little inaccuracy doesn't seem to bug them. I think the performance difference between a sine-drive and six-step drive is that the six-step drive is more efficient at the ESC but a sine-drive is more efficient at the motor and gets smoother operation. I have an ESC design that I haven't built yet because high current PCBs are too costly right now. It's way oversized for the current I need, even when unheatsinked. When I build it I should try increasing the PWM frequency and doing the modulate the voltage duty cycle sinusoidally using just zero-crossing detection to see how well it works. You do end up with alittle problem though...how do you measure the BEMF of the floating phase if it's no longer floating? (since all 3 phases are now driven sinusoidally 120 degrees out of phase.)

Maybe one day enough sine-drive hardware will be around that induction motors will be considered...assuming someone wants to tackle those really nasty induction motor control algorithms. I doubt there will ever be any induction outrunners though. That would be...the ugliest...most complicated motor.

kyleservicetech
05-24-2009, 01:59 AM
As for the high powered permanent magnets, they were used in the magnetic actuators for the high voltage circuit breakers where I worked before I retired. The shop management safety department had warnings posted all over the place on the handling of these magnets. You get your fingers between two of them, and if you are lucky, you will still have your fingers.

They had a magnetic attraction of over 1000 pounds!

Moxus
05-24-2009, 06:00 PM
you know what the magnetic flux density of those magnets was?
they were probably over 1 tesla neodymium magnets, or samarium cobalt magnets, if they were subjected to temperature.
anyway, they are pretty strong, yep. and if you got some size on them, they are downright dangerous. they WILL crush your fingers, if you ask them nicely to do it. that is, not taking care when handling them.

thats by the way one of the difference on quality engines, and cheap ones.
the cheaper uses neodymium, while the more expensive use samarium cobalt.
and dont go "wow, my cheap engine is better than hacker, it has a stronger magnet!".
yep true, neodymium is stronger. but it cant take heat like samarium cobalt can.

DKNguyen
05-24-2009, 06:04 PM
I've heard people use "motor" and "engine" to talk about the things that use internal combustion engines. Do people also use the word "engines" to talk about electric motors? I've never seen that before. I'd prefer Samarium Magnets in my motor than Neodymium. But...seems Neodymium is cheaper and stronger so that's what everyone uses.

Moxus
05-24-2009, 06:36 PM
the definition of motor and engine is roughly the same, and both ic engines and electric engines fit under the definition.
that is, both convert energy from one form to another.
if you want a specific term for ic, you can use "prime mover".
that dont apply for electric engines. unless you can call electricity a "fuel source".
or so i think :S cant say i ever gave a half damn about the linguistic details.

DKNguyen
05-25-2009, 01:08 AM
the definition of motor and engine is roughly the same, and both ic engines and electric engines fit under the definition.
that is, both convert energy from one form to another.
if you want a specific term for ic, you can use "prime mover".
that dont apply for electric engines. unless you can call electricity a "fuel source".
or so i think :S cant say i ever gave a half damn about the linguistic details.

It can get confusing especially when someone is talking about their electric conversion or on a mixed forum. :confused:

kyleservicetech
05-25-2009, 01:27 AM
Moxius
Not certain on the flux density, but they are samarium cobalt magnets that measured 1/2 by 1 1/2 by 3 inches in dimension, and the magnetic actuator used two of them.

This magnetic actuator's function was to move the circuit breakers contacts open or closed, with 1/2 inch travel, with 500 pounds of force. The time to open or close was on the order of 25 milliseconds, or 0.025 seconds.

Moxus
05-25-2009, 02:44 AM
thats definetly something to have in your "might become handy one day" toolbox =D