DIY Electric Car Forums banner

Switched Reluctance Motor

156K views 142 replies 19 participants last post by  cts_casemod 
#1 ·
I "rediscovered" this motor in my searches and it seems to be making a comeback because electronic controls are now ubiquitous and the design has many advantages and few downsides. As a point of reference this was touched upon here:
http://www.diyelectriccar.com/forums/showthread.php?t=75088&highlight=reluctance

Here are some links I posted elsewhere but I'll consolidate them here:

http://machinedesign.com/article/the-switch-to-switched-reluctance-1211
http://www.youtube.com/watch?feature=player_embedded&v=jt4Fa4H43Iw (simple VR motor)
http://www.youtube.com/watch?v=nQ8G5wnH5sc&feature=player_embedded (variable reluctance test motor 3p stator 4p rotor)
http://www.youtube.com/watch?v=b3hmkehrcUg&feature=player_embedded (SRM test)
http://www.youtube.com/watch?v=G2qS2TxU9KY&feature=player_embedded (first run)

I made a drawing of a proposed design for a switched reluctance motor:



Here is a video of a simple SR motor and controller I just made:


From the description:

This is an experimental switched reluctance motor which uses only external electromagnets in the stator, and no windings or permanent magnets in the rotor. Thus it is a very simple, rugged, and low cost design. The expense and complexity may be in the controller. The principle is essentially using electromagnets to align a piece of steel and then switching the excitation to adjacent magnets to achieve motion. I think I need to work on the design of the rotor and the pole pieces so that the force aligns the rotor at an exact point of rotation, which means wider stator pole pieces or more narrow rotor tips.

I may make some changes and see if I can get significant improvement, especially for self-starting. I think I will need to use a full three-phase H-bridge so the pole pairs can be driven both positive an negative (N and S). I also found that it helps to drive two sets of poles together to get higher torque and smoother transitions.
 
See less See more
1 1
#32 · (Edited)
I guess that could be done. It could allow for stacking of multiple rotors, too which would be nice... This could be achieved in a similar manner with a radial flux motor by just increasing the number of laminations in the stator and rotor, although it wouldn't exactly be "modular" once manufacturing was complete...

I'm going to stick with a radial flux setup for the moment though; I've got a new and improved rotor in my head which I'll draw up in CAD tonight.

Honestly, I'm not qualified to speak with authority on this subject. My background is mechanical, not electrical (or anything regarding magnetism), I'm just applying what I know from a mechanical standpoint and making assumptions based on my limited understanding of magnetic motor theory. So I guess take what I say with a pinch of salt... lol. I'm ready and willing to be corrected on any of this.
 
#33 ·
Like most of you my background is mechanical and not electrical also. But motors have always been a on going interest.

@iti uk - I have been wrestling with the idea of commenting or not commenting on your drawing of your motor. And decided to make a comment, to maybe keep you from wasting some time and effort. This is not meant as criticism but just from the point of helping.

The stator would be great for a BLDC but not so great for a SRM. The SRM makes its torque/rotation, from moving the rotor from an unaligned to an aligned position. By that I mean, the rotor pole needs to be close to a stator pole but not under it. The SRM is a "doubly salient pole motor". For this to happen there needs to be "gaps" between the poles. The stator gap widths need to be close to the width of the rotor pole faces in size. The drawing you posted doesn't have the size gaps necessary to do this. Again this is meant to help not as criticism.

Here is a PDF that gives a lot of good information about pole sizing - http://eprints.gla.ac.uk/2851/1/optimalmiller.pdf


Here are good patent PDF's that explain a lot too -
http://www.freepatentsonline.com/5111095.pdf this one is good for the use of five phase short flux

http://www.freepatentsonline.com/4883999.pdf

http://www.freepatentsonline.com/5015903.pdf


To all - the SRM is used in all different sizes, from the new Dyson vacuum sweeper all the way to 100 or more kW steel mill drives.
 
#34 ·
Criticism is always welcome so please don't hold back if you've got something to say. I'd rather know that I'm wrong sooner than later. So am I to understand that the poles on the stator should have gaps of equal to/greater than a pole width? I'll include that in my next update, which I'm about to start messing with now.

Also, those links are good reading. I'll pour over them properly later when I get home.
 
#36 · (Edited)
Here's a quick idea of my rotor lamination idea (shown with only one endplate, the other being hidden to show the iron) - each pole pair is a seperate lamination stack, and the stacks are positioned using dowels in the stack/endplate holes. Assuming that the rotor doesn't get hot, the void could be filled with any choice of lightweight material for windage reduction purposes.
The endplates are aluminium.

edit: added an extra screencap with both endplates, for illustration (minus fasteners or dowels, for now).

Out of interest, the pictured assembly weighs 0.596kg, excluding the (arbitrarily chosen) 20mm diameter steel shaft (0.843kg including the shaft). The overall diameter of the rotor is 100mm.
The previous non-seperate pole-pair rotor weighed 0.678kg (0.925kg inc. shaft), so that means a saving of 82g (overall saving of 9%, inc. shaft)

edit 2: added a shot of the rotor sitting in the stator, with one endplate hidden.
 

Attachments

#37 ·
Now your cooking! Glad you didn't take what I had to say as an offense. Some people do, and especially on line where you can't have a one on one talk. Don't want you or anyone else to get the idea that I'm an expert or anything close to that. Just a old guy with too much time on his hands that loves doing things and figuring stuff out.:)

What I'm using as a test bed for the stator is a GM car alternator. The stator has 36 slots just like the 5HP motor I'll be using for the full size one.

Starting with an alternator or AC electric motor gives (in my opinion) a head start. The stator laminations are already made, even though they need modified. And the end bells are ready made. All thats needed to be fabricated is a rotor and shaft. To me thats a win-win.

I'm going with a three phase my self. That way I can use a MC3035 BLDC controller to do the commutation logic. http://www.onsemi.com/pub_link/Collateral/MC33035-D.PDF

http://www.onsemi.com/pub_link/Collateral/AN1046-D.PDF
 
#38 ·
@ PStechPaul, I want to apologize to you for hijacking your thread with my ideas and links. Didn't really mean to, it just sort of happened. I will understand if you ask a moderator to move the stuff to another post. Just thought it would be good to have it all in one post. Again I apologize.
 
#39 ·
@PStechPaul, my apologies also. I've come into this thread treating it as an idea assembling zone...

@shortbus, no worries. I have had to learn to take criticism at work, as it is better to be seen to admit fault and then make a change than to deny and take it personally.

Sounds like you're well on your way to making a working motor! My thinking is that I can get all the laminations water-jetted fairly easily, so all that remains is a shaft, a housing and the endbells of the motor. I'm going to shop around for shafts before I specify a shaft bore diameter in my rotor design, and I'm trying to cook up something interesting for the housing which might allow machining from billet if I can find someone with a CNC mill. It's a lot of work, but I've got time and only vague vehicle plans in mind. :D

In a final model (not this prototype), I'd like to take advantage of the fact that most of the heat in the motor will be produced in the stator by including some sort of liquid cooling in the housing. It's a complete ground-up project so I'm trying to include as much beneficial stuff as I can while still remaining plausibly manufacturable...

Who knows where it'll lead.

As for the 5-phase nature of my motor, I haven't even considered controllers yet. That may be a project for the future, or I'll find a 5-phase controller somewhere... :eek:
 
#40 ·
Don't worry about "hijacking" the thread. Your replies are very much on-topic and I am reading them with great interest. I've been sidetracked on other projects but I want to get back to doing some work on my ideas for a SRM design. I made a controller using a PIC18F2431 and two L298 drivers, so I have six independent drives which can provide either polarity to each of three pole pairs, or separate unidirectional drive to all six. I want to see how various switching algorithms work. :cool:

I have a fan motor which has six pole pieces but they are shaded poles so I might need to remove the shorting bars, and the gap in each pole might not be ideal. I tried to rework the rotor and I was able to cut off the shorting caps to expose the aluminum winding ends, but it was not practical to remove the laminations and reuse them. I was able to drive the shaft out and I may make a rotor from a stack of steel washers, but I'd have to do a lot of cutting and milling, and the magnetic properties would be far from ideal. I might instead look at a short path design where the rotor would be mostly non-magnetic material and have sections of steel on the outer surface.

Thanks for all the ideas. I might be able to try some things and post results soon! :)
 
#42 ·
I was thinking of cooling the stator . I like the idea of lower voltage motors with increasing voltage as speed is increased . So the Remy motor uses hairpin("U") square wire silver soldered together to make the winding . So now you can use ceramic insulators and if the copper is a tube, cooling gas or oil can be run through the conductor .Also the coils have less winding to winding voltage differential , standard wires can have large wire to wire voltage differential . that's if the first and last wires are next to each other .
 
#43 · (Edited)
PStechPaul,

I liked the prototype you posted at the beginning of this thread. One thing occurred to me as I watched it, though; have you tried different diameter rotors? That rotor looks very large for the number and size of the stator poles. If it were me, I'd be tempted to halve the diameter of the rotor (and rearrange the stator poles to suit, of course) to see what happens...?

Is your controller scalable to 5 phases?

@aeroscott, thru-winding liquid cooling sounds interesting, although I'd be concerned about the drop in cross-section of the copper. Rather than replacing copper, my initial thought is to replace several of the laminations (1 in 10, maybe?) with a hollow "laminate-a-like" through which to pass coolant. I might mock this up this evening, I'll see how work goes...
 
#44 ·
your concern for loss of cross section is a good point. This makes me think about the tube being thick wall with very small cooling hole .this may mean H2 or He as the need for low viscosity and high velocity . I just remembered where I saw tube conductor , supper conducting generator. It had very low number of turns .
The wound coils(as seen in some pm and most sr motors) have some room around them(space between coils) . I wounder if spaces could be made between layers of winding for coolant gas or oil . These type coils are more like transformers .
 
#45 · (Edited)
Another difficulty regarding hollow copper wire is manufacturing it to our requirements. Further, once manufactured, it might be tricky to actually wind the wire effectively without crushing or kinking the tube at the corners of the laminate.

edit: Although I love the idea of cooling the coils directly either by having tubular wire or replacing laminates with laminate-oid cooling channels, after an afternoon of thinking, I'm coming to the conclusion that a (relatively more) simple solution is to use a housing which includes a water jacket or passageways to cool the outside circumference of the stator. Doing it this way doesn't compromise the performance of the electromagnets and is most likely far more simple to implement.

I really should get started on mocking up a housing...
 
#48 ·
Since there are no magnets involved the heat issue isn't as big a deal, like in PMDC and BLDC. The heat that the high temp magnet wire can withstand is hard on rare earth magnets. Thats one reason most commercial DC motors use ceramic magnets, they can take the heat better without loosing their magnetic properties.
 
#50 ·
Here is the schematic for my motor controller:


It has six independent outputs which can be driven to ground or V+ by the L298 bridge drivers. So it should be scalable to 5 phase if you are using a unipolar configuration.

I realize that the narrow pole faces are less than ideal and thus do not really lock the rotor to their position. The fan motor I plan to use has six poles with wide pole faces so I think by matching the rotor to the stator it should achieve maximum reluctance when aligned. Another idea was to use a smaller radius on the tips of the rotor so that the gap would be smallest when exactly aligned.
 
#52 ·
This is really just proof of concept. Rather than using half-bridge or hi-lo drivers and large MOSFETs or IGBTs, I decided to use the L298s because I had them already and they are simple to hook up in the prototype. I'm using 24 VDC relay coils which draw only about 100 mA so my motor will be very limited power. Once I have a better understanding of the concepts I may proceed to make a more comprehensive design (probably with a bigger PIC18 part or possibly a dsPIC). I already have the C18 compiler so I can code the PIC18 parts in C. The dsPIC compiler is a bit pricey so I'd rather not commit too much money to the project unless I can really justify it.

I'll probably make it so that it can drive a SR motor as well as a three phase ACIM. :)
 
#56 ·
I've been away for a while, but in between working I've had some thoughts regarding these SRMs. Specifically, I've been thinking about axial-flux setups (despite what I've said previously lol). I've got a question to run by people here; Is the torque of the motor proportional to the number of rotor-stator interfaces (i.e. a single stator, single rotor ARM has one "interface", a stator-rotor-stator sandwich has two interfaces, etc). If this is the case, I wonder what a split-rotor design might be like (See attached diagram).

My thinking is that the extra passive stator in the middle of the rotor sandwich will provide an extra two interfaces while adding little in the way of mass to the motor. I imagine that this would weaken the field (extra air-gaps, etc), but I wonder if overall an increase in torque might be felt...

The diagram is a bit abstract, but imagine that you are looking top-down on an axial flux motor assembly. The red sections are the stator(s). All poles on the stators (including the poles on the "middle" stator in the split rotor design) are aligned.

In the split rotor design, imagine a horseshoe magnet arrangement for the "outside" stators, with the EM coil wound on the horseshoe magnet. The "middle" stator is passive (has no coil).

Has this been tried before? Thoughts?

I'm still (slowly) going ahead with my plan of creating a prototype, but I'm currently having difficulty visualising ways of securing horseshoe EMs in a stator. Google's not been too helpful and I'm not totally happy with any of my ideas so far.
 

Attachments

#58 ·
I think it is a good idea to have at least two interfaces as you picture, which is similar to a disc brake with a caliper that grips both sides of the surface. But the problem I see is that the components would need to be very accurately aligned and kept from touching. The magnetic field strength increases as the gap decreases, but if the disc is not perfectly aligned between the stator poles, it will pull harder to the side with less gap, and may bend the rotor enough to contact, at which point it becomes a brake.

I have also considered the possibility of using some means to hold the alignment and maintain the proper gap in the axial flux design. One possibility would be to use a Teflon spacer or coating on the surfaces, although I think it would still create too much friction and would be subject to contamination and degradation. Another idea is to use a ball bearing of some sort on the periphery of the rotor to keep it aligned. And yet another idea would be to have grooves in the rotor and stator for ball bearings which would also provide a magnetic path with essentially zero gap. That might be good for very low speeds and high torque.


The radial alignment is more robust because the rotor can be held on center with good bearings and the surface can be machined accurately to hold a very small gap. The problem may be at high speeds where the centrifugal force tends to expand the rotor (although expansion from temperature may be more problematic). However, the heat will be generated mostly in the stator and expansion would increase the gap.

I need to get back to work on this project. I'm trying to finish up another couple of jobs and then maybe I'll have time again.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top