Motor Kv - the truth

[Erfolg]

John

 

I would read it if the link worked on my machine.

 

It is times like this, that I find computers infuriating. Links that work for others, images that I cannot see, but it is not always the case. I guess it is my firewall etc. that sometimes gets upset.

 

Erfolg

[Brian Parker]

Erfolg,
We are probably 'guilty victims' of using relatively simple explanations for complex subjects.

 Off thread, but referring to your weight on a spring. If you hang the spring onto a spring balance, give the weight a pull and then let go you have an experiment in inertia as the weight accelerates and decelerates as it comes to rest at the top and bottom of its stroke (much the same as the reciprocating parts of an engine).

 (Kinetic energy snippet. A rocket loses most of its mass in the burnt fuel and will receive in return more kinetic energy as a proportion the chemical energy than will the ejected burning gasses.)

[Roberto Gava]

The more I read the more convinced I'm that calculation software can't be accurate because it starts from data as the motor Kv that it is not clear that it is stated with the same rigour by the different brands.

There is also the problem of the propeller behaviour also difficult to calculate with enough accuracy.

Very interenting thread. Thanks Tymbo for opening it and all of you for participate and share your knowledge.

[Erfolg]

Roberto

 

The point you are making is in my opinion totally valid, in that the answer obtained cannot be considered to be absolute. For as you say there are so many variables and unfortunately not all of equal value in there range and importance.

 

I do not dismiss programs such as Motor Calc as worthless, you can at least believe you are "in the right ball park".

 

As with many modelling systems, the more, known data you have, the less assumptions are necessary to make use of standard relationships. The advantage of standard relationships is that the relationship, is either absolute or any assumptions are known at a theoretical level and the limitations of the relationship known.

 

Timbo has adopted one approach to measuring Kv. Many others apparently have also trod the same path. The obvious variable is the ESC. From what little information I have found on the web, the output voltage, approximates to a sinusoidal wave pattern (AC). It by no means follows that 10v DC equates to a 10V AC output. Annoyingly the individual who did the work (Camille Goudesuene, writer for Quite Flight, Lecturer and researcher at the University of Illinois. Mathematician by trade) did not provide values in and read from the oscilloscope.

 

After being pointed by forum members towards the "Colin Dedman's" method. I believe it has many virtues. Conceptually there are far fewer variables, it has a basis which is founded on well accepted scientific/electrical effects, in principle simple to implement as a process and finally obtaining a reading is also straightforward.

 

Above all it would be most useful to have either one standard method of obtaining and conditions for determining Kv. Or the Kv value, being annotated with basis of measurement. Just as we often see RMS (root mean square), CLA (centre line average), music power etc.

 

None of this will happen overnight, if at all.

 

Timbo has my admiration, I totally  commend and applaud Timbo for actually doing something. Hopefully this will not only spur discussion, but get some of us doing something (could that be applicable to me?) and not just talking. Although the discussion is important, as an aid in useful thought.

 

Erfolg

Edited By Erfolg on 19/05/2010 09:50:11

[Klippy]

Hi Tim, reading with interest, any chance you could do the tests using a different manufacturers ESC? It would be interesting to see if the ESC makes any difference to the maximum motor speed. I suspect that it might.

 

[Tim Mackey]

I did mention this aspect earlier, and altering the ESC timing etc will also obviously have an effect. I suspect that ( also mentioned before ) some of the variablkes are to connected with the manufacturer using maths to arrive at the fihure, and this is probably based on actual voltage at the windings rather than at the ESC input. However I still feel that the actual results obtained as here, using the mandatory ESC etc as this is how we actually use the things!

I may get around to doing some more testing....no promises - we'll see

[Alex Whittaker]

Timbo,

Immensely enjoying all your pukka technical stuff. A real antidote to dumbing down. Anything that has datalogging, sensors, laptops, and real measurements of revs, amps and volts get my approval every time. More "E" in RCME!

Reminds me of blissful days in the Electronics Labs of my youth... and before that, luvverly double Physics on a Thursday afternoon at school! Sadly no laptops then ... but you've made an old anorak very happy. In your next pics, can you please have a nice Gould double beam oscilloscope (just for eye candy) in the background? Keep up the good work.

Alex (Whittaker) 

 

[Tim Mackey]

Hmmm now you have me pondering Mr Whittaker

I may just have to sneak off and check your profile for any mis-spellings of the word ELECTRIC

And here.... just for you.... uncovered i n the darkest dustiest corner of the workshop

 

[Erfolg]

Remembering that I am not an Electrical Engineer.

 

I have quite a lot of concerns regarding what is being measured with the multimeter. I understood (from the very depths of my memory) that multimeter's are incapable of measuring AC current directly. The solution is to rectify the current, apply a root 3 correction (RMS) to provide a DC current which is used to obtain a reading.

 

From the publication Camille Goudesuene, it seems that the oscilloscope reading of the ESC provided an approximation to a sine wave. My interpretation of the traces published that the sine wave is closer to a "Triangular" form, than the classic Sine wave. At worst I saw, or think I see, gaps in the trace, equating to the frequency of supplying current..

 

My question is, if any of my concerns have any validity, would they influence the resulting Kv value in a material way?

 

I was again in deep thought whilst plastering this morning, something which I hate, so a drifting mind kills the pain. I was thinking of the 0.95 value. I thought could it be 60 degrees, root3/2, but that is 0.86 or something similar. So who knows where it actually comes from. I know it does not really matter, but does it not annoy you not to be able to spot what it is?

 

Erfolg

 

 

[Myron Beaumont]

Erfolg

Root 2 surely not 3  ie 1.414  .That's just off the top of my head mind you ! & I havn't even looked at my tables !

[Peter Beeney]

Erfolg, You are saying you wish to see one standard method of determining the Kv. Can I ask, please, whom you think may undertake to regulate this? Also I’m afraid I don’t quite understand the statement - ‘the Kv value, being annotated with basis of measurement?’ Maybe, if there were one standard method, the technical details of the procedure would be freely available anyway.
If you wish to know the Kv of a particular motor, or motor ESC combo, why not simply just measure it yourself? I wouldn’t think it’s that difficult. I would just use a tacho and a voltmeter. I’d have thought the standard type aeromodelling items would give you a pretty close approximation. Or you could use some high standard instruments, for a more accurate result, but that might be costly. I’ve not checked any prices, but I’ve have thought that a tacho and multimeter would cost around £20 - £40 for the pair? And they would last a long time and do many tests!
If I were using electric motors I think for me a very useful indicator of motor performance would be a torque curve multiplied by a rpm curve, giving you point of maximum power output, which, if you then had a set of revs obtained per different size prop tables you would be able to choose a propeller to give you the best results. (Max power). These curves and figures were available for some of the multitude of small i/c engines of yesteryear, I think this was because independent bodies undertook to establish these. I’m not sure that any of these various graphs and tables are available anywhere now, certainly if they are I don’t seem to see them.
As a point of interest, I was looking at some test results of some motors recently, completely separate to this forum and the mag, and in a least one set of results it would appear that when fitted with the smallest propeller the motor was going faster than the unloaded rpm. I couldn’t help but think this made the whole test, to say the least, confusing!
A little argument I’ve long used, is that these are consumer items, designed to be ‘consumed’. Or perhaps ‘recycled’ in today’s parlance. As with many other of our modelling ‘tins’, it would certainly seem you need to measure for yourself what is on the inside, rather than totally believe the writing on the outside.
Seriously hoping this post does not cause any thread distortion problems!     PB

[Erfolg]

Myron

 

What I should have written is peak voltage *root2, yep, got me.

 

Peter

 

Standardisation is generally seen as sensible. It is not always necessary to impose standardisation. The immediate self standardisation that I thought of is Household Music centres, where it is common to stipulate either Music Power or max, alternatively RMS, in the advertising blurb.

 

The benefit of having Kv values that we can rely on, is at the point of purchase. I am sure that most of us look at the Kv value in conjunction with other data, to determine if the motor is the one we want?

 

I also thought we we all debating, is which of the ways of measuring Kv are useful. Also trying to get to grips with the limitations of the differing approaches.

 

Another issue that has arisen, is why there seems to be anomalies between data sheet information from the manufacturer/supplier and some post sale tests. Is it the method of derivation or what and does it matter?

 

 

[Sparks]

For the benefit of those that couldn't access the link I posted earlier, here is the explanation provided for the 0.95 factor used in the 'drill press' method of measuring Kv.

 

"The reason for the 0.95 factor is that Kv (Velocity or Voltage Constant) is defined as the DC voltage that would be produced if the motor was a generator. With a brushed motor this is exactly what your average-reading voltmeter shows. A BLDC motor acts as a 3 phase alternator, so you have to rectify its output to get DC. Assuming lossless rectification of sine waves, the average voltage produced is equal to 0.955 times the peak voltage."

 

Hope that helps,

 

Richard

[Peter Beeney]

Erfolg, Yes, I certainly agree on standardisation, however I’m not convinced the model industry is always really good at this; and when I’m in cynical mode, I’m also not sure this is always entirely an accident. I’ve always thought that the music world’s advertising blurbs might occasionally also be something of a ploy, too!
Kv values at the point of purchase I might also agree with. But I would read the whole label. If it contained words like ‘Aveox’ or ‘Hacker’ then I’d consider I was heading in the right direction. If, however, I saw ‘Slippery Sam’s Re-winds’ then I think a modicum of caution might be called for.
Measuring Kv. Perhaps the most useful is the most accurate? If you use different approaches and get different results then perhaps there are limitations. Also depends on what test equipment you have.
Anomalies and test sheet data. Sum this up in my ‘consumer’ paragraph, above. Probably, in the main, I’d say, a derived calculated guess! I’ve toyed with a few motors in the past, in terms of output characteristics, and none these ever seemed to be exactly what they said they were. However, I wasn’t really serious, and consequently perhaps it didn’t matter very much.

Incidentally, the RMS value for a sine wave is the peak value multiplied by the reciprocal of root 2, 0.7071. Or at least it always was, about 50 years ago. And, as you say, it’s the same value as the DC equivalent.
Hope this is all a satisfactory reciprocation! PB

[Myron Beaumont]

PB

You are of course correct ,I was only saying that it had little to do with root 3,which off the top of my head Honestly(again) is 1.736 & I havn't looked it up on my slide rule

 

[Erfolg]

My concern regarding the RMS AC value from the voltmeter, revolves around the wave form.

 

I understood that voltmeters converted the input AC voltage to a DC voltage, on the assumption that the AC current is sinusoidal. I understood that for square, saw tooth waveforms that a multimeter will give a value which is not representative of the input value. From the one set of values produced by Camille, the wave form seems to be triangular. Please recognise, I am working from memories of +40 years, and much changes.

 

I am querying if the value produced on a multimeter is sufficiently accurate to be considered to be adequate for our purposes of measuring KV, if the wave form is a triangular form. Particularly if there appears to have  gaps in the waveform?

 

I am asking the question, not supplying the answer.

 

I am also trying to understand which method of measuring Kv, produces the most appropriate value. Why, do we not all want to understand?

 

Now that Timbo has an oscilloscope we may soon have answers to typical waveforms and how the motor receives its current.

 

Erfolg

[Peter Beeney]

Erfolg, I’m assuming you are talking about the ‘drill press’ method? As in Richard’s post, above? If so, then I don’t understand that either; but I am intrigued.
My volt meter reads True RMS, which means it should give a true indication of any form of AC voltage curves, provided the positive and negative halves of the cycle are the same. This is why, if I were to check the output of a typical 230V mains inverter, the reading would be 300+ V, because the RMS value of a square wave, which is what some inverters tend to produce, is it’s peak value. But, we found they always seemed to drive everything ok.
I think you are right to have a level of scepticism though, this is certainly one of those areas of testing that I would want some verification.
Let’s adopt the “black box‘ approach. Firstly we have to assume we need an ESC to be able to run our motor, so we have to include that in just about every test, as a motor, not a generator, that we do.
So, we have two battery wires disappearing into one side of our black box, and a motor shaft sticking out of the other. But we don’t know what’s inside, and we definitely cannot read any nameplate notations. We press the start switch, and the shaft starts to turn. We measure the battery voltage with a voltmeter and we measure the rpm of the shaft with a tachometer. Let’s assume that our test instruments are properly calibrated, accurate jobbies. If the voltmeter reads 10 volts and the tacho 10,000 rpm, then I think it we could safely make a fair assumption that for every volt supplied the shaft would turn at 1000 rpm. In other words, 1000 revs/volt. We could test at different voltages, say 1, 2, 6, 8 and 10 volts, if we obtained differing results we would need to establish why. Then we may well need your annotation to the test results.
This may well be sufficient to be called a standard test. If any other system gave different readings then you may well query them. The drill press gives the motor figures less the ESC, but it’s difficult to see, in practical terms, how you could actual run the motor, as a motor, driving a prop say, without it’s ESC.
Now the question is, how much difference does the ESC actually make.

Are we going in the right direction so far? PB

[Erfolg]

Peter

 

My concerns were centred on the RMS measurement. In that as a method I thought that there is an accepted error, as I understand or understood it. My first thought was revolving around the shape of the trace/graph and the second is the discontinuous trace of Camille's.

 

In the first instance, You assure me that it is not an issue.  I am not sure, but concerns revolves around what the instrument actually does. I understood that RMS is determined by squaring all +ve and -ve (thereby becoming positive) summating, dividing by the number of terms, where the square root is then obtained. Where as I assume that a multimeteractually  takes the peak voltage and then applies1/2^(1/2) to the value. I thought that this gives a value which is close to one standard deviation of the bell curve. Although I understood that there was some error from this approach, as it is a statistical approximation of one standard deviation.  I assume that it is acceptable, for a sinusoidal curve. I did not know that it is independent of wave shape. If it is accept that wave form does not matter great, no problems.

 

The second issue which I had concerns about, is the discontinuous nature of the oscilloscope trace. What affect does this have on the value of the reading. I took an extreme view. If there were one, instantaneous pulse per cycle, the result from the instrument would be the RMS of the one value. It seems to have  some error, as the reading is the RMS of a single peak value. at a practical level it probably does not matter. In some ways the your square wave seems similar to the instanaenous instance, in that the rise is a step change(n1) and then over t (time) the sum of n2, 23, n4................ which are essentially a single point, would there not be an error of the approximated bell curve, to the sqaure wave?

 

I remember (well almost) going back to "control theory" that black box or functions, that you needed to quantify the limits or restrictions, to be certain that the output was valid. All I am trying to do is put forward ideas, regarding why there may be a reason between the published Kv of a motor and what is being tested.

 

I am not saying, I am right. I remember vaguely that whilst doing some labworks in the distant past, that we, the students, were warned about the limitations of some of the  instruments. I thought, it now seems wrongly that simple multimeter's, were not reliable with some AC currents, I thought we had used something else and still had to undertake a correction.

 

It seems that either manufactures are supplying poor data, or they determine the value by a differing procedure, or the equipment used to determine the values are substantially different to the ones being applied by testers. But there are differences, beyond trivial. 

 

I have put one idea forward regarding a potential issue, if there is no substance to my idea, fine, but why do the values differ?

 

Regards

 

Erfolg

[Chris Bott - Moderator]

I think one point of the thread was to find how much kv ratings varied between motors and between manufacturers.  As interesting as all the above is, is there a practical distillation of the information that we can actually apply?

 

Maybe we should settle on one measurement method and all contribute some test results? The simplest might be Tim's original method where we could divide RPM by input voltage. We'd come up with a "Timbo method" kV for each motor.

 

A lot more results might throw up some answers? 

[Brian Parker]

This Month, in a rival publication, Alasdair Sutherland comments on Props and Static Stalled Props.
Funny old world.

[Ultymate]

 

 Posted by Alex Whittaker on 20/05/2010 12:02:29:

Timbo,

Immensely enjoying all your pukka technical stuff. A real antidote to dumbing down. Anything that has datalogging, sensors, laptops, and real measurements of revs, amps and volts get my approval every time. More "E" in RCME!

Reminds me of blissful days in the Electronics Labs of my youth... and before that, luvverly double Physics on a Thursday afternoon at school! Sadly no laptops then ... but you've made an old anorak very happy. In your next pics, can you please have a nice Gould double beam oscilloscope (just for eye candy) in the background? Keep up the good work.

Alex (Whittaker) 

 

 Had a few laptops back in my school days Alex, but they weren't the ones that come with a qwerty board and screen if you get my drift But maybe that's just me dumbing down this thread 

Cheers Brian 

 

 

Edited By Ultymate on 21/05/2010 12:12:27

[Tim Mackey]

Posted by Chris Bott on 21/05/2010 11:17:19:

I think one point of the thread was to find how much kv ratings varied between motors and between manufacturers.  As interesting as all the above is, is there a practical distillation of the information that we can actually apply?

 

Maybe we should settle on one measurement method and all contribute some test results? The simplest might be Tim's original method where we could divide RPM by input voltage. We'd come up with a "Timbo method" kV for each motor.

 

A lot more results might throw up some answers? 

 

Indeed Chris - but nice to see it has sparked off a lengthy discussion...even though much of it is beyond me .

I suppose with hindsight I should have not mentioned  /published the prop data as this was not the original theme of the thread.

My idea was to simply do an experiment that i had in mind for some time, and see in a real life worlkd situation exactly how a motor performed versus its claimed Kv. Using a variable PSU feeding the motor via an obligatory ESC should have, to my mind, resulted in a close result to that specified. It varied from motor to motor, and my conclusion ( which I am "happy" with ) is exactly that. It varies.

As I said, its no big deal to me as I usually experiment with props on any installation until I get the figiures and flight I want IYSWIM ...but anyone else doing the same test method with different brands please chip in the results and we can have a better sample number for analysis.

[Chris Bott - Moderator]

I agree, it has showed that the way we go about things is as good a way as any.

Buy a motor in the closest "ball park" we can. Then tweak the system hopefully just with prop selection.

 

 

[Peter Beeney]

Erfolg, Yes, I think you may well be correct in your concerns about the voltmeter and the RMS values. I have done a little research on this, which might also answer the questions that were bugging me in Sparks (Richard’s) previous post, concerning the little formula for Kv. However, because this is in a deep and dark neck of the woods, where wise men seldom go, it is indeed in connection with the way that moving coil voltmeters measure AC voltage, circa 1940/50/60s, perhaps the best thing is to gloss over this, as it’s bit of a dry explanation; and at the end of the day, probably does not take us much farther forward on the Kv issue. It may be appealing though, if you are into RMS, Average Values and Form Factors. …. Form Factors? Sounds as though that might be attractive, if only I could remember what it was about; I’m sure curves played a part in that somewhere as well. But, it is also possible, that today’s ordinary digital multimeter might also be suspect in some circumstances. I think I now sort of understand what you are asking/saying and I’m sure you’re right; but whether we would be able to establish this satisfactorily is a another matter.

Likewise regarding the oscilloscope trace. I’ve not been able to look at any of this myself, and I’m still not exactly sure what you are seeing. This, though is my idea of the black box. Whatever goes on inside we don’t know, or care. It’s the same in every case. We only have the end result. Then, given enough end results, perhaps we have a set of figures on which to make a judgement on any future purchases. Unfortunately, that may be somewhat expensive! As you say, there are a number of reasons why we could possibly be seeing figures that don’t always add up. Everyone has different ideas and different test equipment. This is only natural. As far as the manufacturers data is concerned, I’m of the opinion that this is not always 100% reliable. I’d bet this is probably the biggest reason for the discrepancies. As an example, on a slightly different note, but still well in tune, just for fun I recently tested a set of 4 AA cells, which I think I mentioned here somewhere in another thread. These are a set of inexpensive, very high capacity cells, from a reputable supplier, (should that nearly reputable?), and a very impressive performance specification to boot. They failed in every respect! And even when I went easy on them they still failed! Their own spec, that is. They simply turned out to be a very ordinary set of cells; in a comparison test other makes with only standard ideas were better. They were very uniform in failure, a fact which led me to believe that the packs either side on the assembly line would be exactly the same. I pointed this out, amongst other things, in a little report I sent to the supplier: needless to say, he didn’t send me another pack to play with!

At the end of the day, I’m sure that you will always come back to the cut and try method. Or rule of thumb. Or, as in my case, poke and hope. Or find out what someone else is using with success and copy that. As I said, I’d also like to see other figures included on the box, but that ain’t never gonna happen.

Meanwhile, I’m off to have a little pogger* on motional emf, motor speed, torque and stator current and their interaction therein. And perhaps throw in sprinkling of magnetomotive force, (MMF), for good measure…….. PB

* Pogger - An old Wessex word meaning a study of the completely incomprehensible by the totally unintelligible.

[Sparks]

 Hi folks, here are some more numbers to ponder.

Further to my previous post where I reported Kv figures for a range of E-Flite Motors using the 'drill press' method. I have taken some readings with the E-Flite 32 Motor using an ESC at different supply voltages to see how the 'no load ' Kv's compare to the 'drill press' ones.

 

I used a Black Mantis 60A ESC - tried it with both Lo and Hi timing options (the 'auto' timing defaults to Lo timing with the E-Flite 32)

 

Edited By David Ashby - RCME Administrator on 24/05/2010 15:30:30