Rave CD-ROM Brushless from GB: Part 3

[Jay C]

The previous thread is here

Tutorial
GoBrushless Research & Development -> The GB Library -> Tutorials (use version 2 for the most up to date information

Articles
GoBrushless Research & Development -> The GB Library -> Guides

Online Motor Database
GoBrushless Research & Development -> Performance Database

Online Motor Rotor Calculator
GoBrushless Research & Development -> Calculators

Use the Thread Tools pull down to subscribe to this new thread

Jay

[CharlieS]

Thanks Jay for the database update. It helps to answer my question I posted the other day regarding efficiency for GB motors.

I have a few comments:

Here's the measured data from my post, edited to correct for a mistake in the voltage:

12T, 24AWG wye, draws 8.6A x 5.9V = 60W on a GWS HD 7x3.5 turning 12,200 rpm. Measured kV = 2600 /min/V, Rm=0.1 Ohm, Io=0.6A @20800rpm (8V).

I plug the data into the calculator and I get:

Dia, Pitch, I, V, RPM, Thrust, Pitch Speed, Power In, Power Out, Efficiency % = 7 3.5 14.58 4.87 12653 14.4 41.9 92.2 68.1 73.8

This is way off, especially in the current 14.58A versus measured 8.6A! Now I've used motocalc quite a bit and seen this kind of poor agreement before, with brushless motors in particular.

I tried playing around with parameters to get good agreement and found that I had to change the prop constant to 0.6 and the Kv to 2050. The Kv is the critical parameter. There is no way to get agreement unless Kv is lowered from the measured value. Yes, I'm sure I measured it correctly: 20800 rpm @ 8V = 2600/min/V. I've done this kind of adjustment in motocalc and found that it appears to give decent agreement with experiments.

The reason why Kv needs to be lowered is that at high currents the iron is close to saturation and the back emf voltage does not change linearly with rpm. Simple motor formulas based on the inputs Io, Kv, and Rm can't take this into account.

But I'm puzzled as to why I had to lower the prop constant from 0.75 to 0.6. I live at 1200 ft elevation I wouldn't think I would need to change it by that much. Prop flexing at 12,200 rpm? Any other ideas?

Charlie

[Jay C]

Charlie, I assume you are using the simulator on the database. I too question the accuracy of the calculations in all cases. I am not the author of the equations and am not familiar enough with them to know "what could be wrong". However, using the measurements you provided I calculate a Power out of 68 watts. therefore 1 of two things: 1) your amps or voltage are off or 2) the prop isn't really 7x3.5

I'm using the Bob Boucher formula for power absorbed by the prop = K * (Diam/12)^4 * (PitCH/12) * (RPM/1000)^3

Your thoughts? also, some common constants I've seen listed are 1.11 of APC, 1.18 for thin carbon, and 1.31 for Top Flight and MAS props.

Jay

[rysium]

Originally Posted by Jay C

Your thoughts? also, some common constants I've seen listed are 1.11 of APC, 1.18 for thin carbon, and 1.31 for Top Flight and MAS props.

I have meassured GWS 7x3.5 and I got 2.5 pitch instead 3.5. Also the prop flexing at high rpm (over 10k rpm is "high rpm" for 7x3.5) makes the big difference between calculation and measurment.

At the same time, the calculator doesn't include correction for extrene values we are running through our small beast. At high current motor is not linear so all calculations will be off.

I would say for real results for thrust, power out, efficiency, ect. the best method is just measure the rpm, torque, Voltage, Current and use common formulas to calculate power in, power out and efficiemncy.
I found static thrust always off, so for comparision, it's OK to calculate it, but to make sure, that the powerplant will haul your plane as it should, you need to put it on scale and just test it.


RysiuM

[CharlieS]

Jay, yes I am using the GB database calculator but I would get the same results from motocalc with the same inputs. I don't think there's anything wrong with the GB calculator. It's just an issue of the accuracy of the standard motor model (Boucher et. al). Kv is not constant, it depends upon current. When we push our motors we tend to run high on the B-H magnetization curve of the iron, i.e near saturation. Thus the back emf which determines Kv also tends to saturate at high current. This effectively reduces the Kv. At no load the current is small, thus the iron is not saturated and Kv is higher. You just need to keep this in mind when evaluating data otherwise some things won't make any sense.


I just measured again - third time.
The amps and voltage are 8.6 A and 6.9 V = 59 W according to my whattmeter and the rpm is 12,100, and the prop is GWS HD 7x3.5. I was incorrect with my voltage previously by one volt.

I don't have an explantion for the prop constant other than to suggest at high rpm (is 12,000 high for these props?) the flexing decreases pitch or lowers the prop constant. Hmmm, I didn't try changing the pitch... I'll give that a try.

Charlie

[CharlieS]

Rysium, you posted while I was responding to Jay. Yes, I think it's more of a pitch than a prop constant problem.

Charlie

[Jay C]

Originally Posted by rysium

I have meassured GWS 7x3.5 and I got 2.5 pitch instead 3.5. Also the prop flexing at high rpm (over 10k rpm is "high rpm" for 7x3.5) makes the big difference between calculation and measurment.

At the same time, the calculator doesn't include correction for extrene values we are running through our small beast. At high current motor is not linear so all calculations will be off.

I would say for real results for thrust, power out, efficiency, ect. the best method is just measure the rpm, torque, Voltage, Current and use common formulas to calculate power in, power out and efficiemncy.
I found static thrust always off, so for comparision, it's OK to calculate it, but to make sure, that the powerplant will haul your plane as it should, you need to put it on scale and just test it.


RysiuM

Richard, please tell me how to measure power out? I could only think to calculate the power out given the prop dimensions. RPM, and prop constant. Is there a better way?

Jay

[Jay C]

Originally Posted by CharlieS

I just measured again - third time.
The amps and voltage are 8.6 A and 6.9 V = 59 W according to my whattmeter and the rpm is 12,100, and the prop is GWS HD 7x3.5. I was incorrect with my voltage previously by one volt.
Charlie

1 volt makes a big difference when you multiply it by 8.6A Now taking into account what Richard has said about the pitch actually being 2.5 then I calculate 48.6W to the prop and an Eff of 81.94% That's more believable

Now that I have my new tach (a laser tach for the cool factor and so I can run tests at night without fiddling with a flashlight), my Medusa Research Power Analyzer Plus, two high quality digital multimeters, and an oscilliscope I'm ready to start testing different winds, rotors, and stators Of course this will all take time so don't go looking for it all at once but I hope to complete testing on a 12T 24AWG wye single 22.7mm stator this weekend and maybe retest/validate my measurements for my 14T 24AWG wye single 22.7mm motor.

Jay

[empeabee]

Originally Posted by Jay C

1 volt makes a big difference when you multiply it by 8.6A Now taking into account what Richard has said about the pitch actually being 2.5 then I calculate 48.6W to the prop and an Eff of 81.94% That's more believable

Now that I have my new tach (a laser tach for the cool factor and so I can run tests at night without fiddling with a flashlight), my Medusa Research Power Analyzer Plus, two high quality digital multimeters, and an oscilliscope I'm ready to start testing different winds, rotors, and stators Of course this will all take time so don't go looking for it all at once but I hope to complete testing on a 12T 24AWG wye single 22.7mm stator this weekend and maybe retest/validate my measurements for my 14T 24AWG wye single 22.7mm motor.

Jay

Good Grief - all that hi tech kit for a $10 motor
Looks like I'll have to progress my dynomometer design. ( even less stick time )

Mike

[mmormota]

Originally Posted by Jay C

Richard, please tell me how to measure power out? I could only think to calculate the power out given the prop dimensions. RPM, and prop constant. Is there a better way?

Jay

You have to measure rpm and torque, the product of them (with a correction constant) is the output power.
Measuring torque is not too complicated, as the torque on he motor body is equivalent to the shaft torque. You need a scale and a simple shaft on bearings, and an arm to push the scale... Takao Shimizu published a photo about a simple but accurate DIY torque measuring device.

If the efficiency is your primary interest, it is very well correlated to the rpm drop at high power level.
efficiency = rpm / (Kv * U)

where rpm is the actual loaded rpm
U the battery voltage
Kv is the well known motor constant
(Kv " U) is the theoretical unloaded rpm at the given voltage

[rysium]

Originally Posted by mmormota

You have to measure rpm and torque...

150% correct answer - one thing to add:

Efficiency = Mechanical power out divided by Electrical power in

RysiuM

[CharlieS]

1 volt makes a big difference when you multiply it by 8.6A Now taking into account what Richard has said about the pitch actually being 2.5 then I calculate 48.6W to the prop and an Eff of 81.94% That's more believable

Yep, I agree the pitch must be about 2.5 as well as the Kv about 2050 rather than measured 2600 to explain my measurements. Wow an efficiency of >81% from a non-custom can CDR motor. Seems high but in my tests the motor was cool after a 20 second full power (60W) run. Maybe it is that high.

BTW my error in volts was poor memory, but I had the watts and amps correct.

Charlie

[Lnagel]

Originally Posted by power

So on my 23T motor 1 turn off, could only vary the voltage about 1/60 of 4.76 volts? Jeeesh, kind of makes me wonder why I check them at all for that matter is there going to be any physical sign that the motor is missing a turn?

The more winds you have, the lower the percentage of over-all voltage one wind represent. On the higher wind motors I wonder if missing or adding a wind or two really has that much effect. I saw one post here where a person missed winding a whole tooth and the motor still ran. Probably a lot lower efficiency though.

However, once you have the voltage, you can measure the rpm and use mmormota's formula to calculate your kv

[Kureus]

Why not just use efficiency = (Iin-Io)x(Vin-IinxRm)/VinxIn, or are these motors not modeled well with the resistor/current offset model? Using this model, and measuring the motor characteristics Io, Kv and Rm, you can set up a spreadsheet to show the perfomance at various speeds, and check against measured values for various props. I know we would like to see 80%, but the above calculations usually show peaks of 65-70% (for my motors, which may well be far from representative of a really well done motor)

[azfrench]

When building for 6 magnet poles, do you just use 6 magnets or is it ok to use 12 magnets with 2 N side by side then 2 S side by side? I thought that I saw that it was ok to put 2 N together and 2 S together but just wanted to be sure.

Rick