Bafang SWXK-5

[NRG]

The panasonic benefits because it drives through the gears just like all crank systems. These multiply the available torque hence better hill climbing, all hub motors are stuck in one gear...just like driving a car using only top gear...this becomes inefficient the slower you go.

I can't answer your last question as a) these motors come in different RPM versions and b) any figure changes depending on what size wheel the motor is mounted in. A smaller wheel is like a lower gear so will give more apparent torque.

The best controller is the one matched to your motor, typically for the 250W motors no more than 15~18amps and if going the BPM motor route up to 25amps or so...

Take a look at http://www.ebikes.ca/simulator/

Choose the eZee motor in 26" or 700c wheel, eZee 14Ah battery and 20amp controller. It gives a good indication of what to expect at various speeds although not accurate for the Bafang...

 

[jbond]

The data from that link seems to suggest that the controller is not intelligent, P (Watts) = 1.54 * N (rpm) all the way down to zero. I feel so stupid but I have to ask, how do I add an image from my PC to this post?
Can anyone also help me find a chart for POWER versus VOLTAGE for the same motor?
The goal of the project is to improve the power output of the motor at below 100 rpm, not breaking the 250W legal limit.

I have trouble reading those graphs because I can't quite tell what is caused by the motor and what by the controller. As far as I can see it only covers 180rpm to 225rpm. So I think it's the very top of the range where all the figures peak, except for current. I think if you increase the load so the rpm falls below this, the current will continue to rise unless it's limited by the controller which in practice it always is. So on those graphs, that's what you get at various loads and rpm where the controller is set to a 10A limit and it's not being hit. At the left we have the no-load speed, 0Nm torque, 223rpm, 0.71A. At the right we have 15Nm, 180rpm, 10A, 280W Pout.

Roughly (mod waste), the motors are constant torque for a given constant average current. So in the region where current limiting is happening, you have constant current, hence constant torque, hence Power out is directly proportional to rpm.

What I'm not sure about is what happens if you have a conservative current limit but push up the voltage. So comparing 36v-10A limit with 48v-10A limit. At the top end above current limiting the no-load rpm goes up. But running both setups at say 10mph, the 48v is having to be chopped more to stay below 10A. Does that mean the power out actually stays the same?

 

[trex]

jbond, I could not make head or tail out the the website data either. The guy climbed up to the top of the mountain, putting in a fair amount of leg power. Then he produced a near straight curve (yellow N).
I think the effective limit of the SB at low revs is determined by the maximum torque that the SB can deliver, 15.01 Nm @ 180 rpm with 36V 10A kit.
What would the torque be at 100 rpm with 48V at input? the current may be limited to 5.2A to stay legal. I reckon the SB unit may still achieve 75-78% efficiency at this RPM.

 

[banbury frank]

Hi trex

if you use the infinion controller and cycle analyst and limit the speed to 15 MPH I would not worry about the watts no one will bother you it is all about puling power you need 48 volts The cycle analyst can tell the real speed by counting the rpm as it fires the motor and you can input the rim size to get a true MPH reading

 

[trex]

Frank, I try to design an addon for the Bafang kit that should help it deliver much more torque at low speed, I try to get about 22Nm at 100rpm, 40Nm at 50 rpm. Outside this range, the motor response won't be affected. Would you be interested by such a device?

 

[NRG]

Interesting, I don't see how you could make such a device...I think the Bafang efficacy will be closer to 65~70% maybe less at 100RPM...

 

[trex]

Take a look at ebikes.ca Hub Motor Simulator

I tried the emulator, it's useful but it can only calculate for a few existing controllers and can't factor in any add-on.
Linear Technology has a single chip solution for a DC / DC converter. It's nearly 95% efficient so a credible alternative to a mechanical solution. My device (only on paper at present) should use the Hall signal to work out the rpm and boost the input voltage to the SB unit accordingly.

 

[NRG]

So how would that work Trex? Up the voltage at low RPM and then cut it back further up the speed range?

 

[trex]

The idea is to make a programmable stepup DC to DC converter. Outside the 50-100 rpm range, the gain is set to 1 to protect the controller and the motor. Inside the range, the booster is active and the gain of this stage is inversely proportional to the RPM. The solution is simple and should work as long as the gears will hold.

 

[NRG]

Interesting idea but wont the controllers phase current limit be the overriding limitation?

As soon as its reached it will use PWM to limit current no matter how much voltage you feed in....

 

[trex]

NRG:
wont the controllers phase current limit be the overriding limitation?

Yes, this is true but the controller outputs up to 36V at 10A supplying the DC/DC boost, at 100 RPM, the latter should be able to output 48V 5.2A with ease. Giving 90% overall efficiency, the controller would only need to use 77% of the max width.
There is a driver after the DC/DC converter before the motor. The driver is synch'ed to the controller's phase output.
Testing will show if it works.

 

[flecc]

I'm sure you can safely bet it's been tried by the motor manufacturers over the last decade of these motors, and especially since the five year resurgence of the Panasonic type units which has been taking their market away.

In 2005 to 2007, SB reigned supreme in the export market with hardly any competition, but since then they've lost the entire top end market and suffered huge inroads on their middle market, much of this due to the drive-through-the-gears systems. This position is getting worse with the new introductions from Daum and Bosch and powerful new rival hub motors.

Suzhou will have been looking hard for an answer.

 

[trex]

RNG:

As soon as its reached it will use PWM to limit current no matter how much voltage you feed in....

Yes, this is true. The battery voltage cannot change, at maximum output, the controller delivers pulses corresponding to around 36V at 10A. The DC to DC converter outputs much higher voltage at shorter pulse width.
Assume that the SB has 10 poles, has constant impedance, admits 10A as nominal current at 36V, works at 75% efficiency at 100 rpm (16.6Hz). The booster has 90% conversion efficiency.
At 100 rpm, 154W output, 15Nm torque, the calculated pulse width of a dumb 36V controller will be 25.6ms without booster.
If the booster is programmed for a gain of 2, it needs to deliver a pulse width (same impedance, same efficiency): 25.6ms * 2 (gain) * 36V (battery voltage) * 10A (SB at 36V) / (48V (booster output) * 15A (SB at 48V) *0.9 ) = 28.4ms

 

[NRG]

Ah! OK I see what you are trying to do, will be interesting to see if it works in practice. Good luck and keep us posted

 

[trex]

flecc:
I'm sure you can safely bet it's been tried by the motor manufacturers over the last decade of these motors

Agreed but it's not their job to add management function to their motors. Otherwise, they would have added digital control such as I2C bus.
NRG: thank you very much for your input. I sincerely don't know if it will work. Maybe modern controller have already got torque boost in low revs.

 

[NRG]

...
NRG: thank you very much for your input. I sincerely don't know if it will work. Maybe modern controller have already got torque boost in low revs.

There is a programable setting on the Xiecheng 116 controllers called Block Time which can be set from 1 to 10s. It overrides the current limiter for that set period of time to give a temporary boost to the motor say when moving away from standstill. I think it only works when opening the throttle and is not used when the motor speed slows down for instance...