Efficiency of hub motors at higher powers?

[jag]

Most of the time motors are modeled with a simple linear torque response to current. This obviously isn't true beyond a point. Magnetic material in the stator will saturate, internal resistance increase etc.

Has anyone measured actual torque response at full load operating of a 9C, clyte or similar motor? The simulators will happily extrapolate the linear model to several kW of power, but I suspect reality may be more disappointing.

[317537]

Agreed. There is a lot going on in motors other than just voltage and current and torque. We use the simple linear models to make life easier and estimate all the other fury bugs that eat at our battery power.

Everything is usually estimated at 25c temperture. Well we get some hot weather here in OZ and other places are freeezing cold.This does change the field posts somewhat.

PWM and blcd does fix a lot of problems with current speed control applications but it does have draw backs too.

[Doctorbass]

Methods seem a great dcandidate to test that for us!

Well an ebike DYNO seems to become interesting again!

Doc

[ZapPat]

From my experience this summer using a 2WD bike (dual direct drive hubs - NC's), it does not take that much power demand before two motors become more efficient than just one.

I leave my place on my 2wd bike, and my girlfriend takes a rear wheel drive bike, both using nine continent hubs. Both bikes have no chains, so pedaling power is out of the picture. Same batteries (3p12s Lipos), but the 2wd bike has two large 12FET controllers while the 1wd has a small 6FET.

First part of the trip is flat and downhill, for about 6km. The single hub bike has so far used less energy than the dual drive bike.

The second part of the ride takes us up some pretty long hills, some parts quite steep. One hill is steep enough that the single NC hub bike barely just makes it up at it's current limit of 40A, which is around 2000W. The 2wd leaves the 1wd in the distant dust anytime there is an incline, but I'm nice so I wait for my sweety to catch up on the flat parts. So by the end the ride we are at the top of the hills, and now both bikes look to have about the same charge.

This means that the dual hub motor bike is actually more efficient than the single when the load gets over a certain point, but less at light loads. Winding heat is one culprit for sure (I^2*R losses), but how much does magnetic saturation come into it, I wonder?

[torker]

Plus the dual hub setup has got to weigh at least 20 lbs. more. That is pretty impressive. Kinda like folks wouldn't believe me when I told them a 500 cubic inch Caddy that I bought would get 16 mpg. on the hiway.

[jag]

Well an ebike DYNO seems to become interesting again!

ebikes.ca has dyno setups. http://www.ebike.ca/simulator/They use dyno result in their simulator, so this is much better than just vendor data. However even their new one only goes to 50nm torque. They found that the linear model in their simulator worked well in the 5 to 50nm range they could test, but I suspect that for the majority of motors the interesting range, where non-linearities become significant, is above 50nm.

Two other alternatives:
Use a car/mc dyno. However eBike motors are way below the power range these are designed for, so it may not give very accurate results.

Do coast and acceleration testing. There are some rules of thumb and calculators on the net. But I think their models are too simplistic.

Knowing the mass of bike and rider then:
A coast test determines drag a different velocities.
An acceleration test, compensated for drag, gives at the wheel power.

Methods seem a great candidate to test that for us!

Certainly then the controller won't be the limit!

[Drunkskunk]

I think Doc's right. Time to revisit the Dyno project.
I could build one in an afternoon I bet, but the software is beyond me.

[adrian_sm]

This means that the dual hub motor bike is actually more efficient than the single when the load gets over a certain point, but less at light loads. Winding heat is one culprit for sure (I^2*R losses), but how much does magnetic saturation come into it, I wonder?

Could be just that you are going faster on the 2WD, therefore operating the motor at more efficient RPMs... Got to balance that aginat the added wind resistance.... You might still come out ahead in terms of efficiency if the 1WD is getting bogged down. Or is that what you were saying above?

[Doctorbass]

I think Doc's right. Time to revisit the Dyno project.
I could build one in an afternoon I bet, but the software is beyond me.

I am searching for a waterpump dc motor or a treadmill DC motor that could do the job for power under 5kW and used as a generator..
I already have the load.. so combined to a controller and a metal structure it could be built easily..
I am also not very good for software or ADC interface programing... mayve a NI interface and some labview could be the best..


Doc

[Toorbough ULL-Zeveigh]

to answer the original question, at full load the xlyte 4-series saturates @1.8kW.

If the electromagnets are saturated with
the flux density at near maximum, a further power increase will fall
off sharply. The only one I have solid experience in this with was the
Crystalyte 408. It saturated at ~1.8kW, and anything more put into it
was being dumped as heat. As we all know, heat is the bullshit
detector for efficiency. In my example that worked out as ~25A at 72V
= 1.8kW. But it could also be 37A at 48V = 1.8kW, or even 50A at 36V
= 1.8kW.

from post#19 http://groups.google.com/group/tidalforce/browse_thread/thread/b64253d8fd27b618#

[ZapPat]

to answer the original question, at full load the xlyte 4-series saturates @1.8kW.

If the electromagnets are saturated with
the flux density at near maximum, a further power increase will fall
off sharply. The only one I have solid experience in this with was the
Crystalyte 408. It saturated at ~1.8kW, and anything more put into it
was being dumped as heat. As we all know, heat is the bullshit
detector for efficiency. In my example that worked out as ~25A at 72V
= 1.8kW. But it could also be 37A at 48V = 1.8kW, or even 50A at 36V
= 1.8kW.

from post#19 http://groups.google.com/group/tidalforce/browse_thread/thread/b64253d8fd27b618#

That's excellent info, Toorbough!! I wish we had this info about the X5's and nine continent hubs too. 1.8kW isn't very much power at all for an X4, and I would guess that the NC hubs will saturate somewhere around this value too, maybe even a bit less since they are lighter?...

[Toorbough ULL-Zeveigh]

you'll have to ask jeff how high he's got his 5303 up to, he's an ES member as well.
sounds like he hasn't hit saturation yet using his 72V A123 pack..

[jag]

Many people quote about 2kW as the upper useful limit for a clyte 4xx, so this is roughly the same as Jeff's observatoon below. My guess is that it may be well into the flat part of saturation though, and the "knee" where the response curve goes from linear increase to flat may be before 1.8kW. Perhaps that knee can be found out with some more careful testing.

to answer the original question, at full load the xlyte 4-series saturates @1.8kW.

If the electromagnets are saturated with
the flux density at near maximum, a further power increase will fall
off sharply. The only one I have solid experience in this with was the
Crystalyte 408. It saturated at ~1.8kW, and anything more put into it
was being dumped as heat. As we all know, heat is the bullshit
detector for efficiency. In my example that worked out as ~25A at 72V
= 1.8kW. But it could also be 37A at 48V = 1.8kW, or even 50A at 36V
= 1.8kW.

from post#19 http://groups.google.com/group/tidalforce/browse_thread/thread/b64253d8fd27b618#

I'm puzzled by the numerical claim that both

25A at 72V and 37A at 48V = 1.8kW, or even 50A at 36V

are equivalent in terms of saturation. Magnetic field is caused by the current, so should occur at the same phase current, independent of voltage. I guess numbers might correspond to battery current and voltage into the controller. However, to quantify motor saturation we need amps going out of the controller into the motor phase windings (which may be higher depending on the pwm duty cycle).

I'm curious about my 9C. At low speed and high phase currents throttle response seems "flat" beyond a point, which may be a sign of saturation. However, I have no objective measurements to prove that.

[ZapPat]

I'm puzzled by the numerical claim that both

25A at 72V and 37A at 48V = 1.8kW, or even 50A at 36V

are equivalent in terms of saturation. Magnetic field is caused by the current, so should occur at the same phase current, independent of voltage.

Indeed, the numbers given by Jeff are battery side numbers by their looks?? This also makes me wonder if the tests were done with the motor stalled or what, because as jag says here, magnetic saturation has to do with motor current rather than controller input power. If this is exact, then it implies that the higher the motor speed, the more power we can draw from the motor without it magneticaly saturating.

And how would be the best/easiest way to measure the magnetic saturation point of a BLDC motor? We could feed the motor (any two phases) a square wave and observe the motor's center point voltage. Start with 0% duty cycle and raise it until the center point's positve going voltage ramp starts to change it's slope (the slope should get steeper as the laminations saturate). The slope changes as the winding's inductance lowers. My latest controller incarnation looks like it should be able to supply variable high current PWM to the motor, and I have lots of Lipos too that could easily provide mucho amps, along with a good scope... Anyone see a problem with this idea or have a better suggestion to find a motor's saturation point?

[Arlo1]

I have build a dyno. Seen here http://www.youtube.com/watch?v=XdjQ445WoJw and as a mater of fact it will work for low power ebikes. As for an ebike like mine a motorcycle dyno would likely work. It is a matter of what type of dyno it is what we want to test for acceleration is inertia dynos. But if we could vary the load it would help to test efficentcies at different levels like simulate different hills. So a brake dyno is best for this other wise It is possible to add weights to my inertia dyno as well. As for the program I have and the dyno mesureing equipement it is for inertia. But for acurate variable load tests one could use a brake mounted to a old cheap tourque wrench to hold it at a certain speed and increase load and the tourque wrench will tell you the tourqe! As for most motorcycle dynos dynojet ect. alot of them have eddy brakes so they are a light drum that spins up easy and to add load for higher power bikes they use the eddy brake at a load percentage.

[jag]

I have build a dyno. Seen here http://www.youtube.com/watch?v=XdjQ445WoJw and as a mater of fact it will work for low power ebikes.

Nice work, though as you mention a brake dyno would likely be better for the ebike. We woul dlike the measured for torque/power under sustained load.

As for an ebike like mine a motorcycle dyno would likely work. It is a matter of what type of dyno it is what we want to test for acceleration is inertia dynos. But if we could vary the load it would help to test efficentcies at different levels like simulate different hills. So a brake dyno is best for this other wise It is possible to add weights to my inertia dyno as well. As for the program I have and the dyno mesureing equipement it is for inertia. But for acurate variable load tests one could use a brake mounted to a old cheap tourque wrench to hold it at a certain speed and increase load and the tourque wrench will tell you the tourqe!
[/quote]

Not a bad idea. I hadn't thought of using an external static torque measure and a brake. An electric brake may as you say below, be smoother than a mechanical. A lever and a scale could be more accurate than a torque wrench.
[/quote]

As for most motorcycle dynos dynojet ect. alot of them have eddy brakes so they are a light drum that spins up easy and to add load for higher power bikes they use the eddy brake at a load percentage.

[/quote]
I still wonder if a regular dyno can accurately measure in the 10 or even 100w range. Given that most are designed to hundreds of hp, it would mean they have 0.1-0.01% accuracy. That would be better than my multimeter!

[Arlo1]

I think the thing to remember is we all have big torque! Me bigger then most. But lets say a 300w motor runing at 300 watts and say its 75% efficent. 300x.75=225w then lets convert that. 225/746=.3hp doesnt seem like much but lets say its at 100rpm (100rpm in a 20 inch wheel is ~10 mph) .3x5252/rpm= torque so .3x5252=1575.6/100=15.756ft/lbs or at 10rpm 157.56ft/lbs torque now most motorcycles dont even produce that torque so I think alot of them would work. I deffinatly want to try it. Mine has to have huge torque to be able to flip people off like it does. And I do plan to dyno it!