Understanding outrunner motor specs.

[David Hall 9]

Hi, new to electric flight, I've bought a few motors and an ESC to start me off with two airframes.

I understand that the way motors are specified might vary from manufacturer to manufacturer. One example is the length of the motor body being used, or sometimes the length of just the magnets. My 2826 motor could be a 2812 motor.

My last (very cheap) motor to arrive is described as 2212 2200KV, and has a body length of around 26mm and a diameter of 28mm, confusing!

I thought that if the physical measurement can vary so dramatically, that the KV spec might also be guesswork.

This is where the confusion really sets in, I assumed that the speed rating might be related to the number of poles on a motor, but this 2200KV motor and my 1000KV motor each have 12 poles.

What does set the off-load speed rating?

 

 

Edited By David Hall 9 on 26/01/2016 12:46:47

[Martin Whybrow]

Motor dimensions do seem to be a mystery, I've struggled to find a certain size motor, only to realise that there were loads about the correct size, but the model number didn't make it obvious.

Off load speed is determined by a lot more than the number of poles, it's also dependent on the strength of the magnets, the air gap between the magnets and rotor, the number of turns on each winding and the shape of the pole pieces; all of these alter the back EMF of the motor which is what determines the off load speed (ignoring bearing friction and windage losses).

[Biggles' Elder Brother - Moderator]

Hi David,

well the first thing to say is that this whole business is a bit of a "movable feast"! Conventionally the numbers do not refer to external dimensions (although they might do with some manufacturers!) What they most commonly refer to is the stator diameter and the stator stack length. So a 2812 motor has a stator 28mm long and the stacks on that stator are 12 mm long. The physical motor will be somewhat bigger than this over its external dimensions.

kV is, as you say, is the no-load revs/volt applied. You can think of it as a sort of "gearing ratio" for the motor. So for a given motor-size/battery combination a low kV motor will generally have more torque and swing a bigger prop (slower) whilst a higher kV motor will have less torque but swing a smaller prop faster. The amount of power you get as a result could be very similar in both cases - but you are getting that power in a different way.

kV is strongly related to the number of turns on the poles. Generally high kV motors have fewer turns and vice-versa.

You'll have noticed that a lot of my comments above are hedged about with terms like "generally", "usually" etc. I'm afraid that is somewhat in the nature of the beast! Motor specs are not a fully standardised international system, they should be but the reality is they're not! Different manufacturers use different systems! I find the numbers (e.g. 2812 etc) are only really useful as a guide and I tend to rely a lot more on data such as:

1. maximum continuous current, maximum burst current (determines the maximum power),

2. input voltage range (determines which battery I can use), and

3. kV (determines the approximate prop size).

Finally I use the external dimensions if given (not the stator dimensions) just to check it will physically fit in the model.

BEB

[Denis Watkins]

BE assured David

The lower kv is desireble

the more useful torque is produced in the lower kv motors

as opposed to screaming high revs

[Geoff S]

Motor design has been a black art even when few of them had permanent magnets. My father knew more about motors than I did and he cursed that the manufacturers of ac/dc motors kept their data secret. I know I deliberately took little notice of the parts of my electrical/electronic engineering courses that involved permanent or electro magnets because I could pass the exams knowing the bits that interested me most ie semiconductor circuits.

Thus I know little about what internal hardware structure (winds, poles etc) is needed to provide the motor characteristics I need and like BEB I use the black box principle and simply study the characteristics. So I look first at the maximum and continuous current permitted to keep smoke to a minimum as well as the voltage (though that is less important - current is the killer) and then the kv, which, though not always accurate, is usually somewhere near.

How the motor number relates to actual physical dimensions seems to a moveable feast. eg my EMax GT3526/04 has dimensions of 44mm dia and 52mm long. I guess the 3526 must refer to the rotor size but I don't know. What is useful is the weight (265 gms) because it gives an idea of the power rating roughly 3 watts/gram so I know this motor is likely to be capable of roughly 800 watts. Actually it's bit less than that but it's in the right area and it powers my 3kg Tiger Moth easily.

Geoff

Edited By Geoff Sleath on 26/01/2016 14:30:25

[Frank Skilbeck]

Another thing worth doing is downloading a motor drive train programme, drivecalc is one I use

[Biggles' Elder Brother - Moderator]

Posted by Denis Watkins on 26/01/2016 13:55:55:

BE assured David

The lower kv is desireble

the more useful torque is produced in the lower kv motors

as opposed to screaming high revs

I don't wish to strongly disagree - I too prefer lower kV generally as it means I can usually put a more "in proportion" prop on the model.

What I would slightly take issue with here though is that one of the complications of electric power is that generalisations that become "rules" don't really work! There isn't a "one-size-fits-all" solution.

So, while that is generally a good "rule" for say scale models of propellor driven aircraft it most definitely isn't a good rulle for an EDF! There you want fan speed - the fan works better the faster it goes. This is because it is usually of very limited diameter, the tip speed will still be below optimium even at fairly high revs. So you need to really push the revs to say 30,000rpm or so. For that you need a very high kV motor. If you use the proposed rule "low kV is always good" you will end up with a very poorly performing EDF!

Another issue is that if the kV is too low then you may end up not having the ground clearance for a big enough prop to get the full power out of the system. The result will be that you fail to harness the potential of the motor and and up carrying around useless extra weight!

I think the only rule that really works with electric power is "there are no rules", you have to evaluate each application on a case-by-case basis!

BEB

Edited By Biggles' Elder Brother - Moderator on 26/01/2016 15:28:59

[DH 82A]

One "rule" you might consider is to get a wattmeter . Also I use " Web- O - Calc " ( Google it) as a very simple easy to use calculator for electric flight . A rev counter can be handy too, it will give you an idea of the airspeed .

[Dave Hopkin]

Also take the "maximum wattage" rating of a motor in context, most motors will handle a range of Lipo's - it common to see something like "2 to 5 Cell" given with the motor data with a wattage figure of (say) 500w - you will only achieve that power on maximum voltage (ie 5S in the example) with the correct prop - bigger/coarser prop means more load, which means more current drawn as the motor tries to reach its KV rating, which in turn equates to more power generated - go over the top and you will draw more current than the motor can handle and smoke will result.

[Swissflyer]

Hello David,

Yes, Drivecalc & WebOCalc make it all rather easy as others have already commented. WebOCalc will get you in right ballpark from the flight perspective but first you need to have a grip on motor kv etc.

DriveCalc exposes what the real rpm/V of motors is and lets you see if you will be operating with good efficiency towards the top of the efficiency curve.

In the jpg below you will see that I am thinking about a 4S setup (14 volts is a conservative number) turning an APC 11x7 E prop at about 10’000rpm.

The blue line here shows the motor at 10’210 rpm.

The green curve shows the motor efficiency 77.4% at that prop speed.

You should also notice that the pitch speed is 109 km/hr (about 70mph) and thrust at 2.4kg – powerful.

Now for the magic, you will probably remember, from David Burton’s excellent series, that a propeller running on the ground is pretty much stalled and will only start “flying” correctly with some airspeed under its belt.

As the propeller “unstalls” it produces less drag so it can speed up.

In the days of Speed 400 can motors we used to call it “unloading”, typically a Speed 400 would unload by about 20% in the air, they were low on torque.

A good, modern brushless motor, with a well chosen prop, will unload by about 10% (these days you can check that in real time)

So now look back at the blue line and imagine that our motor unloads in flight from 10’200 rpm to 11’200 rpm.

Now look at the efficiency curve, you will see that the motor has moved to the top of the efficiency curve at about 80%.

You will know when you have a motor set up this way as you seem to get more power appearing by magic as you power up going into up into a big open loop it is a glorious feeling and you will be amazed by how many mAh are left in your battery at the end of the flight.

A couple of health warnings!!!

Do remember the old pilot’s adage about burning fuel to carry fuel.

An efficient set up will let you fly longer with a lighter LiPo, however, the system I am describing is drawing nearly 600 Watts (one bar of an electric fire), thankfully, most of it is being used to fly the model and you only have to get rid of 120-130 Watts (a good sized light bulb) from your engine compartment.

I have seen models (and read too many reports) of models that are incorrectly set up and so produce 300-400W in the motor body and engine compartment, and then…

Unfortunately a Watt meter will only show you that the total energy drain is 600Watts and you may only get the bad news in the sight of blue smoke and a crash in the visitors car park (the RC Groups report was terrifying).

So my second request is please avoid motors that are not in the DriveCalc database until you know how to measure the efficiency and set up power systems safely yourself.

The good news is that the predictions of DriveCalc will correspond very closely to your Wattmeter and you can fly feeling safe & really enjoy the magic of high efficiency electric flight.

Enjoy it

Mark

[eflightray]

A lot of good technical advice, but don't forget the basics, -- the right prop makes the plane fly how you want.

Once you have a pretty good idea of what prop you want, then you go about choosing a motor and importantly a battery combination that will turn it at a suitable rpm.

Motor diameter and length are relatively unimportant unless the model has a very restrictive nose.

Ray.

[David Hall 9]

Thanks all for such detailed replies. There's a lot of technical info here for me to work on. The performance graph is really interesting.

But, just to return to my low-tech post, given an outrunner motor with very dubious/no specification marking, there is no definitive physical property that can be used to deduce the KV rating? (without running it).

 

Edited By David Hall 9 on 27/01/2016 09:38:35

[PatMc]

I don't trust the stated KV of motors. I use a tach & input voltage reading to calculate the actual KV of a motor running at WOT with no load (i.e. prop). It just needs a couple of marks on opposite sides of the motor for the tach to "see".

BTW there is no practical max volts or max watts for any motor. The max voltage is determined by the ESC & max watts by the minimum useable prop.
OTOH it's often advantageous to use a heavier motor than necessary in order to achieve the correct cg for a model & have the motor running cooler at a current level lower than would be possible with a lighter motor producing the same power.

If practical, it's usually better to use a higher voltage battery with a low KV motor than a low voltage battery with high KV to run a given size prop at the same rpm.

Edited By PatMc on 27/01/2016 11:39:16

[Biggles' Elder Brother - Moderator]

Posted by David Hall 9 on 27/01/2016 09:38:02:

But, just to return to my low-tech post, given an outrunner motor with very dubious/no specification marking, there is no definitive physical property that can be used to deduce the KV rating? (without running it).

 

The short answer is: "No".

As I said earlier the number of coil winding strongly influences kV and in the simple theory is the determining factor. But in practice lots of other things come into play, most of which are very difficult (read almost impossible) to accurately quantify. These include internal electromagnetic losses, mechanical loses due to friction and drag etc. etc.

So, yes you can work out a theoretical kV value for that motor design if you know the number of windings and poles etc. But it might not bear much resemblance to the value you'll actually get when you connect the motor up and test it.

On the subject of the various calculation software systems available. Personally I have very little faith in them. I suspect they are mainly based on a combination of some fairly simple theoretical models interpolated with empirical data. They may have their uses but I take the output they give with a big pinch of salt.

One instance of this was when I blogged the electric conversion of my Black Horse Chipmunk. I did my usual back of fag packet calculations and predicted a given max power for the set up. Someone ran the same data through one of software systems and suggested I was predicting the power about 30% high, if I recall correctly. I pressed ahead with the conversion anyway - quitely confident (hopeful!) in my calculations.

When it came to testing the Wattmeter told me I was actually about 10% low - ie I had 10% more power than I predicted. That's OK not a bad estimate at all. What was not so OK is that result suggested that the software was a whopping 40% incorrect! Food for thought!

BEB

Edited By Biggles' Elder Brother - Moderator on 27/01/2016 12:52:07

[Steve Hargreaves - Moderator]

The two most important figures on an electric motor spec are the max current & the kv figure. The given "dimensions" of the motor are pretty irrelevant other than for physically mounting it. As you have found out they do tend to differ. Some manufacturers quote the outer casing size whilst others quote the stator size.....daft but true!!

The max current will tell you how much power the motor can handle & the kv will guide you to a suitable propeller size.....the final arbiter is the wattmeter.

A lot really does come down to experience however & as a wise man once said "Good judgement is the result of experience & experience is the result of bad judgement"...

[Swissflyer]

Posted by David Hall 9 on 27/01/2016 09:38:02:

Thanks all for such detailed replies. There's a lot of technical info here for me to work on. The performance graph is really interesting.

Hello David,

Good to see that this thread is useful to you, thanks for the comment on the efficiency curve; yes those curves have helped me enormously. Ready for another step?

Since this thread is about understanding electric motors, I think BEB’s point on “Why is my fag packet so good?” needs to be addressed.

On the one hand an electric motor, especially a brushless one, looks like a simple piece of kit but why do so many calculators fall down and fag packets work?

Please allow my to tell you a little about me and radio controlled models as part of the answer comes from there. Basically I grew up in the 50’s & 60’s with a wartime flight engineer and pilot for a father. He gave me his Mills 1.3 and later bought me a Mills 0.75, I scaled up (doubled) a Keil Kraft Playboy plan and spent my life savings to buy a McGregor single channel RC with a rubber sequential escapement.

That became a “galloping ghost” and soon I learnt how to strip components off computer boards to make my own RC gear. Grown ups in those days could afford 10 channel reed outfits!

At university, our professors instructed us in the 3 constant model for electric motors, that is where I met our good friends :

Rm (Motor Winding Resistance)

Io (Idle current assumed independent of voltage applied to the motor)

Kv (The motor internally generated back emf when rotating)

All was well under I returned to modelling at the end of the 90’s and jumped into the electric end of the spectrum. In those days the batteries were very heavy and the Speed 400/600 motors were not so efficient.

I quickly found those old equations from university posted on RC Groups by Joachim Bergmeyer. All seemed well; they said increase the operating voltage, set ground rpm about 20% to the right of the peak of the efficiency curve (see my first post) and things will improve.

They didn’t and although Io is supposed to be constant and independent of applied voltage it wasn’t and the promised efficiency gains didn’t come!

So was it back to fag packets? Or did the 3 constant model need improving?

Various experts dug into this and realized that the 3 constant model takes no account of the eddy current losses between the laminations (and other secondary effects). Worse, eddy current losses increase with the square of the current.

In concrete terms, increase your operating current from 10 Amps to 20 Amps and your eddy current losses multiply by 4.

So called 4 constant (and maybe more) mathematical models were developed that incorporated the eddy current effects and gave an excellent fit between theory and practice.

Peak efficiency in a 4 constant model happens at a lower current (less eddy current losses) than in the 3 constant model. However, if you choose propellers to maximize the in-flight efficiency, you will be helping to minimize the eddy current loss effect.

If you look at the efficiency curves you will see that a motor close to peak efficiency in flight will only draw about half the maximum recommended current so will be creating 4x less magnetic eddy losses (remember to square the electric currents).

However, and here is where the fag packets win, most of the calculators (freeware or paid for) that are out there use the 3 constant model so beware if you trust them too much.

continued in next post

[Swissflyer]

Christian Persson, the main author of DriveCalc knows about these deficiencies and designed DriveCalc to “choose between several motor models for the computation, depending on what measurement data are available, to achieve the best possible accuracy”.

I do confirm (from practical experience) that DriveCalc adapts its accuracy according to the quantity & quality of data you put in and its accuracy, with good data, suggests that it goes one or more steps beyond the 3 constant model.

Better still, Christian Persson built tools into DrivCalc for estimating the motor (& prop) calculation reliability. If you look at the extended version of the Hyperion HS3026 data (below) you will notice Motor Calculation & Prop Data Reliability at the bottom right.

The motor calculation reliability is derived from the “Measured Data” on the left.

You can see that DriveCalc indicates -13.57% and +17.09% swings in the measured data and says expect only low to medium reliability from this data.

Whoever measured the data for the Turnigy 540S did a much better job and is told to expect medium reliability, I guess that DriveCalc had enough good data to choose a better algorithm.

I used exactly this motor to power my Ripmax Spitfire some years ago, with the weight gain from moving from NiCads to LiPos and the power gain of the brushless motor it flew like a dream.

If you look carefully, DriveCalc says you can be quite confident on the motor results but not with the prop. DriveCalc relies on us (the modelling community) to supply that prop data and no one has said much about a Graupner 9x6 folder on a 45mm hub.

I switched to some well known 10x6 props APC & GWS HD. Then DriveCalc told me to expect 22.6 Amps current draw at 11.25 Volts, my Wattmeter measured 22 Amps at 11.25Volts so I was comfortable with that.

Christian Persson says if you find an error of more than 3% relative to the DriveCalc prediction, look carefully for the reasons. I have followed that advice over the years and found a 12 turn motor missing a winding i.e. 12:12:11 turns, LiPo battery packs with a rogue cell hidden in the middle, faulty ESCs and motors with Kv labelling that is just wrong etc.

In summary, yes, there are traps and many variables with electric motors.

However, well measured motors in the DriveCalc database let you check what you are seeing against a calculation reliability scale & home in on anything that looks odd.

If it looks odd, it probably is and be sure that you understand the cause.

Happy flying
Mark

[David Hall 9]

Enjoy it

Mark

Mark, thanks for the info and quick run-down on the importance of a well researched setup.

I keep coming back to the graph above. I think that I see how your choice places the efficiency (green) past it's peak, to improve (to nearer the peak) as the motor unloads. The power input curve (red) looking linear here, but losses associated with higher loading dramatically reducing the efficiency (my guess would be that this is due to running the coil poles to saturation). I can't quite see what the blue trace shows us.....

David.

[Swissflyer]

Hi David,

Great, you are spot on with your understanding of the green efficiency curve & the red power curve (I guess that you noticed the scales are on the right hand side of the graph).

On the left hand side of the graph you can see a blue scale called n (rpm) and that is what the blue line is. It says that with no prop, the motor will turn close to 12’500 rpm, the bigger the prop you put on the motor, the more the rpm will fall.

In the case that Drive Calculator is showing you, an 11 x7 APC E is “slowing” the motor down to 10’120rpm. Put on a bigger prop and you will go further down the blue curve to the right (lower rpm).

Do let me know if all this is clear to you?

If you would like to, we could also look at how to avoid the “max power” trap that some unscrupulous vendors use to relieve the unwary of their hard earned cash?

Mark

[David Hall 9]

Thanks for the clarification, Mark.

As I thought more about this and looked at the green trace in the graph, I realised that I don't really understand it. My confusion (taking the last graph as above) is that I have understood that the efficiency curve is plotted as RPM vs current input, so for a fixed input voltage, will be directly proportional to efficiency.

Perhaps the green trace is not plotted against rpm? (as it can't be 78% efficient at 42A input and 20A input for the same prop rpm).

[Toni Reynaud]

The Blue trace is RPM. The Green is efficiency - Power out/Power in. If you look against the three curves, the efficiency is the same for 190W @ 17.5A,11,500 rpm (approx) as for 450W @ 42A, 10,000 rpm. To get the higher RPM figure a smaller prop must be fitted, which will tend to use less current, and still give the same efficiency. It's another illustration of how it's all a balance and trade-off.

[David Hall 9]

Um....so the green trace and the blue trace have prop size as an axis?

So that at the left of the graph, a small prop is used, and increased across to the right. The green point chosen to operate at corresponds to a particular prop size. If we moved left or right along the green trace, the prop size would be different?

Also, I assumed that the power was increased to the train to get a change in current, but is it that power is fully on for the cells chosen and the current is the result drawn for a particular prop?

Sorry to be dumb....

Edited By David Hall 9 on 09/02/2016 16:32:28

Edited By David Hall 9 on 09/02/2016 16:32:43

[Swissflyer]

Hi David,

Thanks to Toni for his support and compliments to you on your tenacity.

Of course there are no dumb questions, only dumb answers.

So as to your question: “Is it that power is fully on for the cells chosen and the current is the result drawn for a particular prop?” The answer is yes, and that is a very useful conclusion as you can experiment with different props and voltages (2S, 3S, 4S LiPo packs etc) to find a good solution for your application.

A very practical way for you to move forward would be to download and try DriveCalc (it is free of charge). Then you could look up the Hyperion motor we have been discussing, set the voltage to a “constant voltage” 14 Volts and play.

**LINK**

Hopefully you will then discover that the curves are always the same, however, different props will load the same motor in different ways.

If you have questions, please ask.

Mark