DC Electric Motors 101

[MJM460]

Thanks Allen.  So the Io value adequately allows for friction, wind age and other losses that contribute to the overall efficiency?

Thanks also for the section on brushless motors.  Looking forward to the next chapter.

I don’t think you are in any danger of having no one reading.  Clearly there is interest in these things.

MJM460

[Allen Smithee]

Backing up a bit - I promised to cover cogging and turn count.

A while back I mentioned poles, teeth and slots as details that we'd try to skip over, but it looks like we can't. I've been deliberately using the vague term "coils" to avoid confusion, but we'll have to go back to basics again. This is another area where text on the interweb isn't a good medium. As Jo will tell you (we used to work together) my natural medium is scribbling on a whiteboard with coloured pens - I think engineers think best in pictures! But creating pictures for this takes more time than I have, so we're stuck with text.

Remember that a DC electric motor is based on the force that a current-carrying conductor (aka "a piece of wire") experiences when it's placed in a magnetic field. If you place a bit of iron or steel in a magnetic field it is attracted to the magnet, but if you put a copper wire carrying a current into a magnetic field it experiences a force which is at right angles to both the magnetic field and the current. This is usually visualised using what's called "Flemming's left hand rule" which is why you see people in Physics and Electrical engineering exams waving a hand with the fingers making a Horus, Tau, Spock or "down wiv da kids" shape (depending on your age and cultural background). While we're at this it is worth pausing for a moment to focus on the point that the force is produced by the CURRENT. Voltage pushes the current along, but it's the CURRENT that makes the force. Remembering that may assist in understanding a lot of this stuff. This also means that if you take the wire around the back and then along the path again the same current gets used again to make more force. So four "turns" of the same piece of wire carrying the same current gives four times the amount of force. We'll come back to this.

So to make a motor you need to align your magnetic field so that it is across the shaft axis (think spoke of a wheel) and then you need the current to be flowing parallel to the shaft so that the direction which is at right angles to both is a tangent to the rim of the spoked wheel. In the very simplest DC motor a wire is wrapped around a centre core "front to back" (parallel to the shaft) so that when the shaft rotates the current in the wires on one side is all flowing front to back and on the other side it is flowing from back to front. This means that the force on each side of the rotating part is in opposite directions, and so we get the classical force couple that produces rotation. And we then use our commutator (physical switches in a brushed motor or electronic wizardry in brushless motor) to swap the direction of the current as soon as the rotor goes past  the magnet so that the force on the shaft is always in the same direction as we covered above.

That's the basic stuff. When you apply this to a "real" motor and make something that's reasonably efficient you want to get as much current into the small space as possible and you want to get the best magnetic field you can. The first you achieve by using lots of turns of wire, and the second is achieved by winding them around a core of magnetically "soft" iron. I'm not going to explain that any further - if the details of the ferro-magnetic stuff interests you then you'll find the answer in books like Hughes (but it's not exactly bed-time reading).

So we want to use a whole load of wires to add up all the forces we get from the current, and we want to bunch them all together so that they pass the magnet at the same time - this allows us to time it to get the maximum effect when we switch the current on and off with the commutator (similar to timing a camshaft). and we want a soft iron core to optimise all the magic magnetism. Give this a few seconds of thought and the obvious solution is to make the core of the rotor from a lump of iron and mill a slot in each side so that all the wires can be stuffed in the same radial place. This works fine - the core actually has to be made from a stack of insulated laminations due to a thing called "eddy currents" that you can look up for yourself (they won't be in the exam) so the slots aren't really milled (the laminations are usually either punched or photo-etched), but we have just made the motor more efficient.

But not THAT much more efficient - we can do better. We can put a magnet on each SIDE, and indeed make the magnets the shape of a half-cylinder so that there is a large field over most of the circumference of the rotor (half "north-inwards" and half "north outwards") and the wires experience forces for almost the whole of the rotation. That's another big improvement. But then we look at the rotor and realise that there's actually a lot of iron and not much wire with just two slots. So the next step is to have multiple slots in the iron, and to thread the wire "windings" around the slots in creative patterns so that a multi-switch commutator can actually switch them on and off in a sequence to give almost constant force on the wires. This is one of the biggies - take the armature our of a typical DC motor and you'll see it might have anything from four to a dozen or more contact pads with lots of wires soldered to each. And if you look at the armature itself you see the iron core has lots of slots in it to get the maximum amount of wire into the space, and here FINALLY we get to the terminology that was the question for this piece!

We call the permanent magnets the POLES. When a current carrying conductor is next to a pole it experiences a force, and when any conductor moves in the field produced by a pole it induces a voltage in the conductor (and a current if there is a circuit). If you think that one through and remember that iron is a conductor you're on the path to understanding the eddy current thing and why the core must be laminated. If the motor has two magnets we call it a 2-pole motor - one will be a "north" and the other will be a "south". The number of poles in a motor must therefore always be an even number.

As already detailed, the iron core has cavities we stuff the wires in which we call SLOTS. In higher-performance motors it's quite common for the slots to have undercuts so they have a T cross-section to help get the maximum amount of wire into the smallest space. While it's not a fundamental requirement it is normal to have an odd number of slots (to make the motor self-starting as mentioned previously) and it is usual but not universal to have a number divisible by three because it makes the commutator configuration simpler. From the motor design perspective the critical features are the slots and the poles. But people see a feature and like to name it, so the part of the pole between two adjacent slots is called a TOOTH. The tooth is not really an important aspect of the design except that it leads to the phenomenon called "cogging" (which was the question that triggered this!).

"Cogging" is the thing you see on some electric motors that almost feels like turning the motor over a compression stoke, or more usually several compression strokes, so it feels like a radial engine. It occurs in motors whose windings are laid out around the teeth of a soft iron core (ie most of them). We covered how the force on the wires due to current is at right angles to the motor axis and tangential to the rim. But the core itself is a chunk of iron that experiences simple attraction to the pole magnets. This is the effect we call "cogging" - the teeth of the core grabbing the poles as they rotate. If you have a small number of poles and/or a small number of teeth the effect is very marked, whereas with a large number of both the effect is less noticeable. It's only important if you want to do slow-speed running and need it to be smooth, because excessive cogging can make make slow-speed running very lumpy which could be a problem for (say) a gramophone turntable drive or an astronomical telescope tracker. It is often said that strong cogging is an indicator of a low-quality motor and there is some truth in it, but it is not universally true.

Now all of the above implied that the poles ae static and the wires are moving - ie that we were looking at a brushed motor. It is all equally applicable to a brushless one where the poles rotate and the wires are static, but the brushless one has another trick up its sleeve. If you look at the picture of a brushless outrunner I posted above you will see that the rotor has more than two poles - it actually has 14 (those silver-looking bars are rectangular nickel-plated neodymium). These alternate north/south around the circumference. The stator has 15 slots, so five iterations of the three switches in the commutation sequence. Bit the real advantage is that as each pole is narrow it allows the wire to be wound AROUND each tooth, because the current flowing on the other side of the tooth (and thus in the opposite direction) will be in the field of the NEXT pole which is of opposite polarity. This allows a motor which is much more progressive - it has a much lower Kv or (if you prefer) much "torquier" making the motor a more practicable power unit which is less likely to need a gearbox. There is a general rule of thumb that the higher the pole count the lower the Kv. It's not universal but it holds pretty well.

It is possible to do the multi-pole thing with a brushed motor but the commutator becomes extremely complex and intricate so it generally isn't done except in very large motors.

I didn't get to turn-count - I'll do that next time.

AS

[Allen Smithee]

Thanks Allen.  So the Io value adequately allows for friction, wind age and other losses that contribute to the overall efficiency?

We generally take the I₀ value as mopping-up the mechanical and magnetic losses into a single easy-to-measure number. It's not literally true, but it's almost always extremely close - enough that DIN and ISO standards will use it as if it was.

PDR

[Admiral_dk]

I spottet a single error so far :

you want to get as much current into the small space as possible and you want to get the best magnetic field you can. The first you achieve by using lots of turns of wire

It's more a bit missing, than wrong .... The magnetic force would be Current multiplied with the number off turns on the coil and the magnetic properties of the Core. Just putting more turns on, usually means more Resistance in the wire + more Inductance => less current .... a compromise .... and I'm sure you can explain it much better than me (you have done this so far).

[Allen Smithee]

I'm trying to keep it simple. I purposely haven't mentioned magnetic lines of force or flux densities or indeed anything to do with magnetic circuits at all because I don't think people need to understand that. But I have alluded to "turns" above where I mentioned that you can feed the same current through again by wrapping the wire around the core. I was going to expand on this in the next bit which is about turn count...

AS

[dieselpilot]

Sensorless startup really isn't a problem today unless you have very high inertia to overcome. Some controllers are better than others. But you still wouldn't use it in a case where you want the fastest response guaranteed. I've cranked engines with sensorless BLDC.

For trivia, it's possible to build a half turn (one conductor per slot) BLDC motor. This isn't common, and is mainly practical in larger motors. A carefully wound motor can pack more copper into the slots to reduce losses. Some hobbyists rewind motors to take advantage of this by rewinding with larger wire than the factory machine wind and laying it neatly in the slots. Large motors can incorporate many other tricks to improve efficiency which are not practical in small sizes.

Disregard number of turns unless you're interested in (re)winding motors. For the application, it's the electrical characteristics which determine performance. That is to say Kv and Rm are most important. Kv and Rm are related to the number of turns for any one motor design. Turns don't really anything when comparing different motor types.

If you are going to get into the concept of outrunner and slot/pole counts we can't ignore winding factor. Most of this is irrelavant unless you're designing motors, or have a highly optimized application where knowing this might help with selection. Fundamanetally, outrunner layout increases airgap area for a given diameter, which is what increases torque.

Cogging only really influences starting current and very low speed running(like a servo drive). Motors can be designed to eliminate cogging completely (skewed armature or magnets) or don't exhibit any by design(slotless, ironless).

No load current Io is a generalization. If it really matters, measuring it at the intended operating RPM can get a more accurate figure.

Generally, I've found the highest quality motors operate closer to calculated figures. Cheaper motors have losses that can't easily be accounted for. But using these formula, you can easily get within 10% or real world performance.

When it comes to motor selection, a concept the should be understood is that motor size is related to it efficiency. A 90% efficient motor of the same size as a 80% efficient motor will be capable of making twice the power. This is due to heat dissipation. The physical size determines how much heat can be shed. Efficiency determines how much heat is required to be shed. If the motor can shed 20W of heat, the 80% motor can handle 100W, while the 90% motor can handle 200W. So in turn if you need 80W output the 90% eff. motor can be smaller. Heat is the limiting factor.

[dieselpilot]

The problem I have is with understanding the motor power curve. From what little I understand, and so far as I can see your explanations confirm this, the torque should drop as a straight line from 0 to max speed (cos the back-EMF is proportional to speed), and that this should result in max power at half of max speed, at least in theory. So why am I getting max power at around 80% of free running speed?

This can probably be explained with verifying measurements. I see the battery was quite large, but did you measure motor voltage under load? If the voltage was sagging as current increased, this would explain most of it.

[Allen Smithee]

The problem I have is with understanding the motor power curve. From what little I understand, and so far as I can see your explanations confirm this, the torque should drop as a straight line from 0 to max speed (cos the back-EMF is proportional to speed),

No, because this sort of defined power curve and torque curve is a feature of IC engines, not electric motors. The forces are the result of the interaction between the magnetic fields and the current in the windings (not the voltage). The current varies with load and so the torque depnds on the torque demans, so comparing it to the power curve of an IC engine is a chalk & cheese thing.

If your load is only drawing three amps and is not turning fast enough then the only way you're going to turn that load faster is to increase the voltage. Try doubling it.

AS

[Admiral_dk]

OK - so I know just enough to get into trouble ..... again .... 

[Ginger Nut]

Great stuff look forward to reading it fully spent yesterday reading 1st 3 posts

Thanks
Ray

Sent from my SM-T580 using Tapatalk

[gunna]

Thank you so much, Allen. I spent my entire working life in electronics but not much to do with motors. This has been brilliant!
Ian.

[Charles Lamont]

No, because this sort of defined power curve and torque curve is a feature of IC engines, not electric motors. The forces are the result of the interaction between the magnetic fields and the current in the windings (not the voltage). The current varies with load and so the torque depnds on the torque demans, so comparing it to the power curve of an IC engine is a chalk & cheese thing.

If your load is only drawing three amps and is not turning fast enough then the only way you're going to turn that load faster is to increase the voltage. Try doubling it.

For a given motor, with a constant supply voltage, and a variable load, if you increase the load the motor slows down so that the back-emf reduces, so the current increases, so the torque increases to match the load. At different loads you get different speeds. There is a load at which the motor will stall, and a speed at which the motor delivers no torque. A graph can be plotted of torque against speed. Surely?

(I don't have a problem with my application, I am trying to understand the theory).

Charles

[Allen Smithee]

Ah, OK - I didn't realise you meant with a constant voltage. Always keep in mind that a DC motor has a fixed characteristic "Torque per amp" (Kt), but the relationship to speed depends on the voltage. When it comes to  relating torque (or power) to RPM it can only be defined for a particular motor & voltage source system (not for the motor in isolation).

AS

[Don1966]

Ah, OK - I didn't realise you meant with a constant voltage. Always keep in mind that a DC motor has a fixed characteristic "Torque per amp" (Kt), but the relationship to speed depends on the voltage. When it comes to  relating torque (or power) to RPM it can only be defined for a particular motor & voltage source system (not for the motor in isolation).

AS

Excellent presentation Allen, I have dealt with motors and generators for over 55 years and you tutor is well presented for the novice to understand. Thank you for keeping it simple without all the engineering terminology to confuse readers. I have to say even I am picking up a few pointers I had forgotten. A big “ thank you Allen “ for taking the time to help others less educated on this subject.

Regards Don

[Allen Smithee]

You are all far too kind. All this flattery isn't good for my ego, you know...

I think the outstanding item was about "turns count". This is an artificial term mainly used where there is a standard motor "size". The good example here would be  the motors used in RC cars and buggies - these are almost all a "can motor" (ie one based on a deep-drawn steel casing) based on the Mabuchi RS-540 family of industrial motors. This is simply because Tamiya put these motors in the kits* for the very first "proper" RC cars back in the mid 1970s (the XR311 and Cheetah off-roaders), and when other manufacturers joined the market they used the same motors so that they could allow their customers to take advantage of the large ecosystem of aftermarket performance add-ons (motors, bearings, brushes, pinions etc) which had sprung up.

These original cars ran on a 6-cell nicad, but they soon offered 7 cell upgrade packs. Also the same motor format was used in things from basic RC cars to off-road buggies, racing trucks and even motorbikes (I still have one of the original graupner RC bikes - with no gyros or anything!). To some extent the differences in requirements could be accommodated with gearing, but the RS540 was an industrial motor that was available in a range of different Kv versions which offered more flexibility. These alternate versions were described to the Car people as "different winds", and soon the demands of RC car competition led people to tinker by replacing Mabuchi's stock windings with their own, perhaps using different wire sizes or a different number of coils to get the best performance for their needs**.

Now as I said previously the route to maximum performance involves getting as much current through each slot as possible. This is achieved by wrapping "turns" of insulated copper wire around each tooth of the iron core. Essentially the number of turns per tooth determines the Kv (more turns means lower Kv) while the size of the wire determines the amount of current they can handle (this isn't completely true, but it's a good rule of thumb). I hope it's fairly obvious that in a given size slot you can fit a large number of turns of thin wire or a small number of turns of thick wire. Those are your design choices. A "nine turn" RS540 will have a higher Kv than a 13-turn one, and it will have a higher current capacity.

This is also where you can see that "Kt=1/Kv" relationship. The smaller the turns count the fewer the number of times the current passes through the slot so the smaller the torque per amp of current.

There is another wrinkle here. The coper wires are usually round, while the slots in the iron are usually square (or at least rectangular) and you know what they say about fitting round things into square holes. So when looking for the optimum it might be that each turn is actually made of several strands of a thinner wire in parallel, because the thinner wire can better fill the cross-sectional area of the slot. The downside of this is that all the strands are insulated, and as you go smaller in wire size a greater proportion of the overall cross-section might be consumed by the insulation. Optimising the wind layout is a specialist subject that I'm not going to talk about in any detail - this was just a taster of the sort of issues which must be considered in the design.

So in summary "Turns count" is just another way of talking about Kv for a specific size of motor.

I think that covers all the specific questions, but if I've missed anything or if there are more questions just shout!

AS

* To be strictly accurate they originally included the smaller Mabuchi RS380 with the RS540 being offered as a "performance upgrade" - it was convenient because it had the same mounting configuration and shaft size so it could be easily swapped

** One of these was a young lad from Stevenage who at the age of eight was embarrassing the adults by regularly wiping the floor with them to become a multiple champion even though he was using inferior equipment. His name was Lewis Hamilton.