Low Kv Motor Dyno Data/Sheets/Charts?

[neuraxon77]

I'm looking for real dyno tests from zero to screaming on sub 200 Kv (preferably sub 100 Kv) motors like the 137 Kv Astro, 130-180 Kv Turnigys, 75 Kv Colossus, 150 Kv Scorpion, Infinite etc or any custom designs suitable for a single stage reduction drive after rewinding if necessary.

A quick search here only found data on motors like the currently prototype 65 Kv, 727W motordude one which doesn't tell me what I want to know, and the Cyclone which does.

I'm particularly interested in the torque characteristics and how low the RPM is where peak torque is produced. It's been difficult finding any useful data; does it simply not exist other than RC plane thrust measurements?

Found this on Youtube while in search of sub 100 Kv... does everyone do this to get there?
[youtube]4DAUdC67HJo[/youtube]

 

[Jeremy Harris]

I will probably be able to help out with some of those plots as soon as I get my dyno finished. I'm still waiting on the delivery of some magnets, that seem to be taking forever to get here (nearly two weeks to get here from London, which is only around 80 miles away).

Jeremy

 

[apullin]

Did that guy rewind a motor for lower Kv? That's pretty neat. I was just thinking about this recently, as I was looking at my S4025-16. Are these motors already a distributed LRK winding, or could they be rewound in dLRK or LRK for a lower Kv?

 

[Jeremy Harris]

Did that guy rewind a motor for lower Kv? That's pretty neat. I was just thinking about this recently, as I was looking at my S4025-16. Are these motors already a distributed LRK winding, or could they be rewound in dLRK or LRK for a lower Kv?

Most of them are wound dLRK, connected as delta. You can easily lower the Kv by around 1.73 by just reconfiguring the existing windings to wye. Rewinding for a lower Kv is pretty easy, particularly on the bigger outrunners. I've done a few now and all have worked OK. The lowest Kv motor I've would so far was a 5330 wound for a Kv of 88 when connected in wye.

Jeremy

 

[neuraxon77]

I'm going to update this thread anytime I find some low Kv dyno data.

mrbill posted on the Mars Electric LLC Axial Flux Motor using a Cycle Analyst measuring energy in, and a PowerTap hub measuring energy out after 3?-stage reduction*. The Kv he listed is 90-97 depending on voltage.

For other mrbill motor data to compare with he has another nice thread*.

Great job mrbill.

* Fine print guaranteed to confuse/enlighten you. :?

 

[Biff]

Attached is some data on a motor that I designed, built and tested.

As you can see, the kelly controller and motor combined had an efficiency of over 80% most of the operating range, and peaked at just over 86%. The efficiency at low torque is really poor, as this was not designed to operate at low torque. I still haven't had the opportunity to re-wind this motor and conduct more tests, I think once I do that the continuous torque would be much higher (80 maybe 90Nm), but this should give you something else to compare with. The temperature of the stator didn't go over around 50degrees C while conducting these tests, the coils approached 80degrees.

-ryan

 

[neuraxon77]

Thanks for the data Ryan.

Looks like a great project, keep us informed if you do the rewind! What are your plans for the motor?

 

[Biff]

Thanks for the data Ryan.

Looks like a great project, keep us informed if you do the rewind! What are your plans for the motor?

There is no plan for the motor. If there were enough interest I would build some, but small runs are more expensive than anyone wants to pay. To get the stator / rotor laser cut and bonded was around $1400. If there were more than 10 sold, I could probably make them for around $700 each. The casing used to test the rotor/stator was spesifcally designed for protyping, so that we could put different rotors and stators in it without having to make many modifications.

-ryan

 

[neuraxon77]

To get the stator / rotor laser cut and bonded was around $1400.

Damn that's expensive for the prototype's laser cutting and bonding, as you say only really makes sense then to make them in numbers if you can find the demand. You never know who might find a use for it if you wanted to market it.

 

[neuraxon77]

I found another I'm adding to the list.

Golden Motor have data for a heavy 11 Kg BLDC for 3-7.5 KW that appears to be around 72-87 Kv. The efficiency shown on the performance curve doesn't match their performance sheet (the rpm indicates different motors).

Nor does it show the low rpm torque characteristics I want to know, however the efficiency curve does start at 750rpm and peaks at 4000rpm for 90 lbs (122 Nm), so I'm wondering if i can take from this that 750 rpm is where peak torque of around 90lbs for this motor is produced from?
I read Ryans motor chart as about 150 rpm for 50 Nm (37 lb/ft).

and
Performance Sheet

 

[Jeremy Harris]

"Peak" torque will be at maximum current and will be fairly independent of motor rpm. This is a fundamental characteristic of PMDC motors, both brushed and brushless. In essence, torque is a function of the effective radius of the rotor, the number of turns and the current flowing through the stator windings (this determines stator flux strength) the flux strength of the magnets and the width of the stator/rotor. The torque is created by the force created by the combination of the rotor magnet and stator magnetic field size and strength and the length of the lever arm with which this force is applied to the shaft. The effect of rpm on torque is modest over the working range of the motor, although it will tend to reduce slightly as the losses in the motor increase. For example, as current increases winding resistance increases, which tends to then act to reduce the available voltage to drive current through the windings, causing it to reduce slightly. Any reduction in winding current will reduce the strength of the stator magnetic field, reducing torque slightly. Also, as the electrical rpm and stator switching speeds increase eddy current losses in the stator core tend to increase, and these eddy currents can create magnetic fields that oppose those that are desired, so reducing flux density in the stator/rotor gap, which again reduces torque slightly. Both of these effects aren't really seen in the normal operating range though and it's reasonable to assume that a PM motor is, to all intents and purposes, a constant torque per unit current machine.

The GM plot is a fairly standard looking one, with max torque of around 150lbs ins at about 145A. It would be pretty reasonable to assume that you'd get 150 lbs ins from this motor at pretty much any rpm over its normal operating range if you stuffed 145A though it.

Jeremy

 

[neuraxon77]

Hi Jeremy,

Thanks for your knowledge here, lots of useful info there, however I think I should have been clearer.

I'm a relative newbie at motor design and reading these charts, so please correct me if anything that follows is incorrect.

What I'm really interested in is how low an RPM different designs produce their peak torque and the torque characteristics or speed-torque curves up until that point, basically what I meant by low rpm torque characteristics is the starting torque characteristics rather than the working range torque you might be thinking about. I read an excellent article about smooth, quiet and high torque PMDC motor design that for the life of me I can't find now, which is what I'm interested in learning about as well as gathering information for my single-stage drivetrain design, power and Kv requirements. I may end up building my own motors and dyno to learn what I want to.

I certainly agree that the working range can be roughly considered "a constant torque per unit current machine", at least until you reach the limits of the permanent magnets before needing to look into electromagnets for higher flux. However most data/charts leave out the starting speed-torque data that I want, and I'm trying to extrapolate this from what data there is, and I'm not sure the constant torque per unit current machine applies here.

To give an example of low rpm peak torque from the data I've found, the new Crystalyte HT24 (high torque, 24mm stator) hub motor produces it's max torque of 47.74 Nm @ 46.1 RPM using 35 V according to their figures, while the HS24 (high speed) produces about the same 47 Nm @ 21.9 RPM because of the fewer turns and lower resistance, however this is at the expense of efficiency which approximately halves resulting in a nice frying pan to cook on at the summit of any long and steep hills with a view.

The Cyclone chart illustrates the ramp up to the low rpm peak torque I'm talking about, where the blue dot torque curve shows peak torque at 600 RPM of 100 somethings (Nm or lb/ft - too lazy to work that out)... that tapers off over the working range because of the losses that occur that you mentioned. On compact, low power, low Kv motors this ramp appears very important to get up and over as it can be a large percentage of the motors operating range.

So what I'm looking for is actual data for non-hub motors that include the speed-torque curve data up until peak torque RPM in different designs including; inrunner, outrunner, radial, axial, skewed/straight laminations, # poles, teeth, windings, etc.

The following is a nice illustration of speed-torque curves for very small motors of different motor types in one company's range that only tells part of the story. Ultimately I'm looking for speed-torque curve data over these motor types as well as reluctance motors.

Thanks for pointing out about the GM motor at 145A to get the higher torque, my brain was stuck on the peak of the efficiency curve!

Thanks for your input.

 

[Biff]

Isn't that the most annoying thing ever. There are so many ways to plot motor data that it makes a comparison between motors less obvious.

I believe that Golden Motor chart, uses a battery as a power source, and puts the motor controller at full throttle (so there is no PWM on the FETS). Then varries the torque on the motor. This will change the RPM slightly and the battery voltage will sag a bit with the load aswell.
The advantage of this type of graph is that it is easy to make, so you can probably find some other motors that have the same type of chart, I believe that Mars Electric makes the same graphs.
Although it gives you a good idea of the operation of the motor at one RPM it doesn't give you a good idea of how it would operate on a normal cycle.

When I tested my motor I wanted to collect data so I could make a chart like the one in this pdf
http://www.rasertech.com/download/18/

The tests I did on my motor were all at constant torque, with varying RPM all the way up to full speed the controller could provide with the battery voltage. At low RPM I got 120Nm of torque, but the temperature increased quickly, and my load and the controller wern't actually capable of sustaining that power either. The winding pattern I had on my motor was rather difficult so I only got a 20%fill factor after I did the calculation. The pattern has 100 turns per tooth, then lots of teeth in parallel. With thicker wire, fewer turns per tooth, putting them in series and having the motor in WYE rather than Delta, I should be able to get 2x as much copper in there fairly easily, so I could push 75Nm of torque without any sort of air movememt, so if it were on a bike, with air passing by I might get 100Nm continuous (or about 10kW of power at 1000 RPM) which I think would be pretty good for a hub motor on a scooter or something. The motor would weigh about 10kg, about 2/3 of the weight of the Enertrac hub motor.

-ryan

 

[Jeremy Harris]

Hi Jeremy,

Thanks for your knowledge here, lots of useful info there, however I think I should have been clearer.

I'm a relative newbie at motor design and reading these charts, so please correct me if anything that follows is incorrect.

What I'm really interested in is how low an RPM different designs produce their peak torque and the torque characteristics or speed-torque curves up until that point, basically what I meant by low rpm torque characteristics is the starting torque characteristics rather than the working range torque you might be thinking about. I read an excellent article about smooth, quiet and high torque PMDC motor design that for the life of me I can't find now, which is what I'm interested in learning about as well as gathering information for my single-stage drivetrain design, power and Kv requirements. I may end up building my own motors and dyno to learn what I want to.

I certainly agree that the working range can be roughly considered "a constant torque per unit current machine", at least until you reach the limits of the permanent magnets before needing to look into electromagnets for higher flux. However most data/charts leave out the starting speed-torque data that I want, and I'm trying to extrapolate this from what data there is, and I'm not sure the constant torque per unit current machine applies here.

To give an example of low rpm peak torque from the data I've found, the new Crystalyte HT24 (high torque, 24mm stator) hub motor produces it's max torque of 47.74 Nm @ 46.1 RPM using 35 V according to their figures, while the HS24 (high speed) produces about the same 47 Nm @ 21.9 RPM because of the fewer turns and lower resistance, however this is at the expense of efficiency which approximately halves resulting in a nice frying pan to cook on at the summit of any long and steep hills with a view.

The Cyclone chart illustrates the ramp up to the low rpm peak torque I'm talking about, where the blue dot torque curve shows peak torque at 600 RPM of 100 somethings (Nm or lb/ft - too lazy to work that out)... that tapers off over the working range because of the losses that occur that you mentioned. On compact, low power, low Kv motors this ramp appears very important to get up and over as it can be a large percentage of the motors operating range.

So what I'm looking for is actual data for non-hub motors that include the speed-torque curve data up until peak torque RPM in different designs including; inrunner, outrunner, radial, axial, skewed/straight laminations, # poles, teeth, windings, etc.

The following is a nice illustration of speed-torque curves for very small motors of different motor types in one company's range that only tells part of the story. Ultimately I'm looking for speed-torque curve data over these motor types as well as reluctance motors.

Thanks for pointing out about the GM motor at 145A to get the higher torque, my brain was stuck on the peak of the efficiency curve!

Thanks for your input.

But the key point is that there isn't really a distinct "peak torque rpm" for a PM motor, they have a pretty flat torque/rpm line. It's a different story for other types of electric motors, like sep ex, induction or variable reluctance, but the motors mentioned so far seem to be PM types and, AFAIK, non-PM motors are pretty rare on ebikes, primarily because of their relatively poor efficiency compared to PM types..

Motor plots are traditionally produced by varying the torque and recording the other data, because, as Biff rightly says, this is the easy way to do it. All that's needed is to stick the motor on the dyno and increase the brake load while logging data.

If you want to know the torque for any PM motor (brushed or brushless) over the normal range of operating speeds (normally from standstill to a max limiting rpm) then it's only a matter of knowing the current and the Kt (the latter can be derived from Kv). There's no "curve" to speak of, it'll be a flat line, maybe with a very slight slope downwards with rpm for very high speed motors, due to eddy current effects.

For example, lets say you have a 75Kv motor and that it has a maximum current capability of 50A. It will have a torque constant (and it's called a torque constant for a good reason) of (60 / (2 x Pi)) / 75 = 0.1273 Nm per amp. 50A through this motor will generate a torque of 0.1273 x 50 = 6.37 Nm. Note that this calculation doesn't need rpm, because torque is independent of RPM for a PM motor over the majority of it's working rpm range.

If you look at plots like those from Justin's simulator, then they are for the system and are dominated by the controller, rather than the motor, when it comes to the apparent torque "curve". The bottom line is that PM motors do behave very much as constant torque machines over their usable range, which is why everyone uses the tried and tested design formulae to characterise them and design them.

Jeremy

 

[Biff]

If you want to know the torque for any PM motor (brushed or brushless) over the normal range of operating speeds (normally from standstill to a max limiting rpm) then it's only a matter of knowing the current and the Kt (the latter can be derived from Kv). There's no "curve" to speak of, it'll be a flat line, maybe with a very slight slope downwards with rpm for very high speed motors, due to eddy current effects.

The other factor which affects torque at high speed is the inductance and is more difficult to calculate. As frequency gets higher the impedance due to the inductance increases, which makes pushing current through the motor in phase with the voltage more difficult, so you need higher voltage and a sophisticated controller.

-ryan

 

[Jeremy Harris]

The other factor which affects torque at high speed is the inductance and is more difficult to calculate. As frequency gets higher the impedance due to the inductance increases, which makes pushing current through the motor in phase with the voltage more difficult, so you need higher voltage and a sophisticated controller.

-ryan

Very true, but I suppose that, taking a purists view, this is a controller issue rather than a motor one. The challenge for the controller is trying to shove current into a motor at high commutation frequencies, as the winding inductance makes things more difficult.

Jeremy

 

[Biff]

The other factor which affects torque at high speed is the inductance and is more difficult to calculate. As frequency gets higher the impedance due to the inductance increases, which makes pushing current through the motor in phase with the voltage more difficult, so you need higher voltage and a sophisticated controller.

-ryan

Very true, but I suppose that, taking a purists view, this is a controller issue rather than a motor one. The challenge for the controller is trying to shove current into a motor at high commutation frequencies, as the winding inductance makes things more difficult.

Jeremy

I agree with that, but looking at the torque curves for high torque motors, suggests there is definatly something limiting torque at high speed. Here is the torque curve and efficiency map for the Remy HVH250 motor.
View attachment hvh250 torque curve.png
View attachment hvh250_efficieny_map.png

either their controller can't deliver the current at higher speeds (either due to voltage limitations or other factors)
or they intentionally didn't want to to go over around 75kW power delivery. I don't think I have seen a torque curve graph that has about the same torque available at low RPM all the way up to Max RPM, except for the one I posted above.

I think it is graphs like those, and the one he provided which make neuraxon77 understand that there is significantly more torque available at low power. But I think we agree, that it is probably more the controller than the motor which makes the torque curve higher at low speeds.

-ryan

 

[Jeremy Harris]

That torque curve does look very suspiciously like some form of limiter that comes into action at the rpm where the sudden roll-off occurs, so I think it must be an artefact of the controller. I can't see how the motor could suddenly create such a steep roll-off. The clincher for me that this might be the case is the very flat torque curve from a 0 rpm to the roll off point, which looks to be exactly what I'd expect and what PM motor theory predicts.

Nice motor though, but probably not exactly ebike sized!

Jeremy

 

[neuraxon77]

Thanks guys, I think I have a much better understanding now.

Take a look at the following small brushed PMDC motor speed-torque efficiency chart and please tell me if I'm reading it correctly. It's a little backwards.

To begin with I was a little confused as to where it lists the motor torque, then I realized the lines don't line up with the current scale down the side. Also, the speed ramps up as the dotted line from the bottom right to top left. I see a number of speed-torque charts do this. Anyone know what the grey area they highlighted on the left is called?

So for the DP30-60V12, peak current of 27 A at 12 V (red line) results in 162.5 Ncm (230 oz-in) and tells me it should produce that torque from 0 with an almost linear efficiency increase with RPM to 1550 RPM. I then looked at another diagram for a BLDC on the same site that showed flat torque from 0 (that I believe is more an infographic than actual performance data). So I looked for other BLDC speed-torque data like the HVH250 torque curve Ryan links in and came to the same conclusion.

I didn't want to believe this, because I don't believe things happen instantly. But I've decided I must be reading the Cyclone torque curve wrong, or it's showing dyno operator or controller and power supply influence on that torque curve. I note that Cyclone chop off the start of other charts for their performance data: 360W and 500W.
However Crystalyte's listing the peak torque RPM as they do still confuses me but may be an indication of the performance of the controller for these new sensorless motors?

It really reaffirms to me how important the way the information is gathered, presented and then interpreted is.

There are some interesting torque-speed curves for AC induction motors from stall I've been reading about here. And another showing a miniscule dip at start for the Tesla Roadster which could be drivetrain for all I know how it was tested.

I also learned from here at the bottom of the page there is rated torque and peak torque which relate to continuous and intermittent torque zones for applications. But I get now that there is no peak torque RPM when properly driven. Thanks Jeremy and Ryan.

My brain hurts now :x

 

[roller]

I'll re-up this thread in the hopes it catches on with some motor data. I'm trying to find the same.

I attach a curve for a 55kv 3kW (legit) 5kg motor I'm making available from Transmagnetics, see this thread: http://endless-sphere.com/forums/viewtopic.php?f=31&t=44058&start=15#p641495

neurax - you can see the amps increasing as it approaches stall torque. If I could stall it on my dyno (can't, it's electrical load, not friction), the amps go asymptotic (vertical, on this graph) at maybe 13-14ft-lbs. This is saturation of the stator core. More amps in = no more torque out. You won't find many curves that dwell in this area of motor performance (if your system is running high duty cycle near zero efficiency, it should go back to the drawing board) ; and as stated in the thread, it is largely controller dependent (current limit, etc) . . though the motor (assuming unlimited current) will have its inherent characteristic.