Seeking an explanation as to how an ESC functions

[Mike Gunn]

Hi everyone

For no better reason than expanding my understanding of life, the universe and everything I would like to understand how an ESC really works. I have got a fairly sound electrical knowledge and have tried to find a well explained run through, but all I have found so far is partial understanding and I think a lot of people who attempt to explain do not have the whole story, so I thought I'd post myself and see where it all leads?

What I think I know so far :

3 phase motors are locked to the supply frequency

Prior to using small brushless (and sensorless) 3 phase motors I have used 3 phase with speed control to power my lathe by connecting an inverter drive and powering from the mains. This achieved speed control by varying the frequency of the 3 phase supplied to the motor from 0hz to 100hz. This in turn varied the speed of the rotating magnetic field which the motor's rotor was locked to and gave around 0 - 2x the normal motor speed.Voltage applied to the motor was always the full supply voltage hence low speed with little cooling was to be avoided.

Our little motors are fed with a simulated 3 phase supply annd hence if they are locked to supply frequency also, how can a motor be given a Kv rating as it's final max speed would not be dependant on voltage. I know this is where my understanding is failing as clearly these motors just get quicker and quicker as you step up the voltage, but I cant understand how this is happening?

I have read that back emf is used by the ESC to sense rotor position in order to switch the next coil(s) on but that still feels to me like altering the rotating fiels speed, so somewhere I'm missing a big part of the puzzle?

Can anyone shed some light please perhaps by explaining exactly what happens inside the ESC when we advance the throttle?

cheers....Mike

[Chris Bott - Moderator]

Mike I'll have a go at explaining how I beleive this works.

The ESC senses the position of the rotor (by back EMF on the unpowered winding I think) and times the switching of pulses so that they are the correct amount ahead of the rotor to pull/push it round. This is actually the 3 way low frequency switching of the ESC.

At the same time, it is doing a much higher frequency PWM switching, this varies the onff ratio of the high frequency switching which varies the effective voltage applied. This is controlled by the throttle position, it affects the strenth of the pull/push and therefore varies the speed.

[Gavin Mack]

From what I understand they are not 3 phase motors, they are in fact the same as brushed motors with 3 sets of windings using an esc to replace the commutaor and brushes

[Mike Gunn]

Hi Guys

Those 2 replies gave me a good deal to chew on, and thank you both for replying. Given that 100 plus people read my enquiry in a few hours I'm assuming it's an interesting topic for everyone, so perhaps you will forgive me for spelling out how I now understand things with the hope others might find it helpful, and secondly you can confirm I have understood things correctly?

Gavin.......I had never thought of it in this way, and to be honest it's a very clever way of looking at these motors. Given that a DC motor uses brushes to continually prevent the rotor from "catching up to" the magnet that it is being drawn towards, then I suppose that is exactly what the ESC is doing. I guess that from this you could argue that a DC motor is in effect actually a 3 phase Ac motor, but equipped with a device (brushes&commutator) that artificially generate a pseudo 3 phase supply from DC....lol....much head scratching there.

Chris

I have (I hope) understood what you have both said and will now try to say it as I understand it?

PWM (pulse width modulation) effectively is the modern version of what we all used years ago on our trainsets to control speed which was a variable resistor. PWM is much more efficient and switches the full voltage on and off thousands of times a second as opposed to reducing voltage by heating up a resistance wire. If we switch the full voltage on and off rapidly with a timing of 50% on andd 50% off then the real effect on the motor is just the same as halving the applied voltage....ie the motor will do half the work or achieve half the rpm. This is the high speed switching you refer to .

The ESC also has another clever bit which is in fact what Gavin has referred to as it also monitors the motor wires for voltage. It relies on the fact that the motor will actually generate voltage from the coils that are not currently energised because magnets are whizzing past them. By analysing the frequency of these voltages it can deduce the position of the rotor and switch the current to the next coil in line thus preveningt the rotor from ever catching the magnet (just like brushes do) This is the (comparitively) slow switching that I assume is carried out by the bank of FET' (field effect transistors)

So.......big breath ..... The PWM system controls the power of the electro-magnets and hence the speed/force at which the magnet rushes towards the coil., The low speed switching FET's just ensure that (like brushes) the rotor is always being attracted to the next magnet irrespective of how fast it is rotating and is a bit like the carrot on a string in front of the donkey. Therefore applying more voltage increases the magnetic attraction, which in turn speeds up the rotor, which in turn increases the speed of the rotating magneic field . The motor speed increases until it reaches an equilibrium and then stabilises unless more voltage is applied.

This sits comfy in my brain....please tell me I have got it.....lol

Cheers.....Mike

[John Privett]

Mike, I think the relative lack of replies is because either, many of the 100+ aren't quite sure themselves, and/or those who do know (or believe they know) realise that Gavin's post and the second half of Chris's describe what is happening.

To comment on your summary I would add that the FETs are doing all the switching - both the "high-speed" PWM speed controlling and the lower speed "commutator/brush equivalent" switching.

Perhaps a bit of history may be helpful (or maybe not!!) Back in the days of brushed motors the ESC just had to do the PWM bit to control the motor speed. At low throttle it would output pulses with mark:space ratio of 0:100 - ie nothing at all! At half throttle 50:50 and at full throttle 100:0.

When brushless motors first appeared they included a sensor and the ESC detected the position of the motor by monitoring the output from that sensor. It then switched the power between the three motor windings depending on the position - exactly as the commutator/brushes do on a brushed motor. Then somebody realised that the sensor was redundant as the unpowered winding could perform the same function.

[EDIT] Having just re-read Chris's post,  I think the first half of it is actually just saying what we are all saying, but in a slightly different way.  So Gavin and Chris (in my opinion!) are indeed both correct.

Edited By John Privett on 19/02/2013 01:31:55

[Peter Beeney]

Mike, Do you think this might be an opportunity to run through the complete working principles of a basic electric motor? Jumping in at the beginning, starting with the ‘motor principle‘ and then eventually arriving at a DC brushless motor. I think it’s sometimes difficult to extract some parts and just try to explain these, without knowing what is going on elsewhere. Also any comprehensive description will cover all levels of knowledge, many folk are perhaps fully aware but others are not so conversant with it all.

There have been other threads along these lines before, one is 'What is the problem', it’s a bit long, but there is some motor discussion at the far end. Another 'Engine Starting Current'. As a result of the ‘what is the problem’ thread, I concocted a three phase ESC output, but only on paper, which I eventually got to drive a motor ok. But again only on paper! I was going to post it, but never did. If I can still find it, I could spruce it up a bit and and then publish, complete with what I think is an explanation of what is occurring. However, I don’t have any actual working circuit details, only what I think must be the principles involved.

I’d be happy to have a try, maybe a bit at the time, perhaps. Once you’ve established what happens in one revolution, all the rest become very similar.

Another option might be starting another thread, specifically relating to ‘Setting it all in motion’, say, or something similar. This could then relate solely to this idea, encompassing the intricacies of motors and ESCs, and maybe a few other such contraptions, as wide as you like, without going off track too far. And it might slot into the ‘Anything and Everything‘ heading.

Just a thought, but maybe it’s not really needed, I’m never quite sure…

PB

Edited By Biggles' Elder Brother - Moderator on 19/02/2013 21:05:36

Edited By Biggles' Elder Brother - Moderator on 19/02/2013 21:09:55

[Steve Hargreaves - Moderator]

My understanding concurs with Chris & Gavins explanation.....it is voltage that controls the speed NOT frequency just like in a brushed motor. As John observes the ESC is taking the place of the brushes & commutator in a brushed DC motor.....a sort of "electronic commutation" if you will.

Amazing what that little component is actually doing...& we complain if they cost more than twenty quid!!!

[Chris Bott - Moderator]

Mike I think you have it

I'll add though that the MOSFETs do both the commutating and the PWM

[Chris Bott - Moderator]

This may (or may not) help.

I think, at any one time, one top and one bottom MOSFET in the bridge is applying PWM to a winding, or two.

Edited By Chris Bott - Moderator on 19/02/2013 21:30:55

[John Olsen 1]

Well, an ordinary DC motor with three armature windings and a commutator is effectively a three phase motor with a mechanical inverter synchronised to the armature position. If you added three sliprings and connected them to each commutator segment you could pick up a somewhat spiky and squarish three phase AC. The frequency would of course depend on the rotational speed which will vary with applied Voltage and load.

Similarly the ESC circuit is generating a rather spiky three phase waveform, and attempting to synchronise it to the rotor position by sensing the back EMF generated by the windings. That this is not always successful is evidenced by the need to change the timing sometimes, to suit the type of load etc. However it is a very useful feature since if we just apply a fixed frequency to a permanent magnet motor, it is liable to drop out of phase lock quite readily when a load is applied. The sensing effectively determines that the rotor is lagging behind and so alters the frequency to suit.

There are some similarities to the inverters used for three phase motors in industry and in my workshop, but a big difference is that most such motors do not have permanent magnets in the rotors...although the Gentle Annie washing machine made by Fisher and Paykel here in NZ did. Without the permanent magnets you can't use the sensing trick, but the squirrel cage motor does not need it.

When the motor is approaching maximum speed, the back emf generated by the magnets passing the coil will oppose the Voltage applied by the ESC, so the current will fall off...and so the torque will fall off. So unless the load is very small, the rotor will not reach synchronous speed. So it won't be able to do that with a propellor connected. So with a load applied the ESC and motor will find a speed where the current being taken is sufficent to match the load, in pretty much the same way as a commutated DC motor will. If the load increases it will slow down a little.

One of those previous threads had a link to an application note with a lot of useful stuff about how the ESC works.

Another John

[Shaunie]

I've dropped in on this topic a couple of times without posting anything because it seems to be getting on quite nicely without me upsetting the applecart .

A little thing I might add is that the MOSFETs doing all the switching must be kept in one of two states:

OFF, where the voltage across them is high but the current is insignificant and therefore the power dissipated is effectively zero and;

ON, where the current through them is high but because their on resistance (RDSon) is in the low milliohms the voltage across them is also low and therefore the dissipation is also low (but not zero though).

The halfway state where the current is significant and the voltage is relatively high is where the power dissipated will heat the device to destruction in milliseconds.

Translated: they must be fully on or fully off otherwise the magic smoke escapes almost instantly.

The result is if we wish to speed control the motor by effectively reducing its voltage this can only be done by PWM (pulse width modulation) techniques i.e. if we wish to run the motor at half voltage we have to run it at full voltage for 50% of the time by switching it rapidly on and off. So as already mentioned not only do we have to "electronically commutate" the motor (exactly the right term btw) but during each phase we need to rapidly switch the MOSFETs on and off to achieve the power level we require. As we cannot switch these devices on and off instantaneously they spend some time in that nasty no mans land of high dissipation each time we change their state. As a result at nearly full power they switch less often i.e. once per cycle than at lower power where they may switch several times per cycle. This gives rise to the possibility that at mid throttle the ESC can get warmer that when running flat out.

Modern ESCs are a masterpiece of technology treat them right, they deserve it.

If I can think of more I may bother you all with it

Shaunie.

[Mike Gunn]

Guys

That was a really great set of answers

Thanks to everyone that has contributed

Cheers,,,,Mike