Can I get more power from my 36V 250W Bafang (10Ah battery) ?

[daniel.weck]

Hi Pedelecs.co.uk !

Now that I benefit from the Cycle Analyst readout on all my cycling trips, I have a better idea of how my Bafang setup performs.

I noticed that the current draw rarely climbs over 10A, or just under 400W peak. This matches the 1C rating of my 36V Li Ping LiFePo4 battery (10Ah capacity), and it only occurs for short bursts, i.e. when climbing hills (by the way, the voltage doesn't sag with this kind of discharge). This also seems to be the factory rating of the motor itself, so I am using it within the designed tolerances (note: this is a 255rpm motor, used on my 16" Brompton wheel).

Given that the Bafang is known to be reliable when "overclocked" to 48V, I wonder if I should be able to "push" more power into the motor with my current 36V setup ?

I think the power cutout currently programmed into the controller is around 12/15 Amps, but I need to figure how much exactly. The Ampere limit on the Cycle Analyst is way greater, I am not actually using that protection feature yet.

According to Li Ping's specifications, the battery is able to deliver more power:

Rated Discharging Amperage: 10 Amps
Max Continuous Discharging Amperage: 20 Amps
Maximum Discharging Current: 40 Amps
Discharging Cut-off Protection: 25-30 Amps

36V 10AH V2.5 LiFePO4 Battery Pack

So, can I technically get more power out of the Bafang if I increase the controller protection threshold ? (e-Crazyman / Keywin Ge / Infineon / Shenzhen Sucteam / EB806-XC-C-3 controller + USB cable with ParameterDesigner.exe software). More importantly, how safe do you think it would be ?

Many thanks,
Cheers, Daniel

 

[Andy_82]

i wait for what Flecc's got to say in your matter. he is the most tech-e chap here. that's very interesting what you ask

 

[daniel.weck]

This video shows a 15A cut-off from the battery itself (BMS tripping), but I don't think they are using a Ping v2.5 BMS though, so probably not representative:

YouTube - Lithium LiFePO4 Battery Load Testing on electric bike

 

[flecc]

I don't know anything about that new motor design Dan, so cannot give a specific answer. Given the way previous designs permitted higher outputs on higher voltages, the new motor probably has the inherent mechanical strength since it's drive cogs and rack look more robust but electrically I can't even guess at it's temperature limits.

Given what you've measured, an increase of 30 to 50% seems possible.
.

 

[fishingpaul]

The 36v 10 amp lifepo4 shown first in your clip the battery cut out at just over 15a not 33a, i would have thought that if the bms cuts the power, it would need switching power off and on to reset,i would guess that your controller is limiting power if the motor is capable of more,after replacing my ezee torq controller with a 36v 350 watt controller(i thought with my 250 watt motor 350w should be plenty they must mean 350 watt max output} i have found it only supplies a maximum 15 amps continous,and the power is well down on the original controller,i am still waiting on a bigger controller and will let you know the difference it makes.

 

[daniel.weck]

The 36v 10 amp lifepo4 shown first in your clip the battery cut out at just over 15a not 33a

Oh, that's correct, typo rectified, thank you !

However I don't think they are using a Ping v2.5 BMS, so their test is probably not representative here. Li Ping is usually quite conservative with its ratings, he never exaggerates them...so I guess we can trust his technical data (25-30 A cut-off)

 

[Herb]

I think the best way to find out the controller maximum is to stall the motor by locking the front brake and read the current on the cycle analyst as the hand throttle is opened – A short blip should cause no harm. This how I found the cut off point on my Brompton with Crystalyte controller (11A using a clip on ammeter)

If the cycle analyst is not illuminated the battery has cut out by its own protection circuit.

Interestingly I have the opposite problem on my other Bike - a 26” Bafang 36V 250W front hub motor.

On hills I need to set the 3 position switch to the middle setting to keep the current under 20A.

Herb

 

[daniel.weck]

Thanks for your input Herb ! About "locking" the front brake to observe peak Amps: it's just like in the YouTube video I linked to...I was a bit hesitant to do the same, but after all it's only for a short burst.

I bet that my 6-FETs controller has very conservative settings, which is fine by me, but I'd like to make use of the 15A safe limit if possible. I'm already monitoring the temperature of the controller + Cycle Analyst shunt (both located inside a waterproof frame bag), and so far so good: it never climbs over 30 degrees, more like 20 degrees Celsius average. I could also add a temperature sensor over the FETs built into the battery BMS, to see how the Ping copes with higher discharge (although according to the specs, we're still at relatively low Amps with 1.5C rate)

/Daniel

 

[Herb]

I should have viewed the video before posting my reply.

Just a thought, have you tried your other controller, From memory I think you bought one for each motor.

I am also interested in the software - to lower the amps on the controller I bought from Frank, No label on mine but presumably the same e-Crazyman / Keywin Ge.

Herb

 

[daniel.weck]

Just a thought have you tried your other controller, From memory I think you bought one for each wheel.

The Bafang motor has a different phase power cable than the Tongxin (both 3 wires, but different connectors), so it would be hard to mix the controllers. I trust that Frank would have correctly configured the 2 kits, he is quite thorough and is well-aware of the risks associated with a Tongxin running at high Amperes.

I am also interested in the software - to lower the amps on the controller I bought from Frank, No label on mine but presumably the same e-Crazyman / Keywin Ge.

You need a USB-Serial data cable, and the correct version of the Windows software (ParameterDesigner.exe is easy to find for Infineon controllers, but not necessarily the right version for you). Frank is able to supply these but it voids the warranty (due to all the returns he had because of people screwing things up).

 

[emissions-free]

Daniel,

It doesn't sound like your controller is limiting the power TBH. You'll need to go to 48V to get a bit more power and speed With the CA you can reign in the current if you require. I'm sure the Bafang will be just fine at this voltage and as long as your controller is up to the job (you'll likely need a 20-25A controller) it will make a very significant difference to the power, speed and acceleration.

The BMS doesn't limit the power in the same way a controller can, it will just trip, and need to be power on/off to reset. I can't see the BMS tripping on a 36V 250W motor TBH, maybe 500W 36V.

Cheers
Paul

 

[emissions-free]

BTW daniel I have a really neat little motor here, 180W at 24V, that makes about 350W nominal at 36V Has all the revs you'll need in a 16" rim and enough power to do about 20mph. Will soon be fitting into a sub 100mm fork and also have 28 holes

 

[daniel.weck]

Thanks emission-free. I will keep my Bafang-Brompton as it is (400W peak is fine), mostly because it is my main workhorse so it needs to be reliable. I agree with you about using a higher voltage, but I'm not that keen on upgrading the battery and controller ($$$). I will make experiments on my unused folding bike with the Tongxin and a torque sensing solution I should receive shortly. This should keep me busy for while
Regards, Daniel

 

[emissions-free]

No worries Dan,

I've sent you a PM

BTW after speaking with Ananda it seems they have had some issues with the torque sesing cranks and additionally they require special controllers as the signal outputted is completely different from the typical speed sensor. I can get the torque sensor and they're not overly expensive (but much more more expensive than a speed sensor, which costs buttons).

 

[daniel.weck]

BTW after speaking with Ananda it seems they have had some issues with the torque sesing cranks and additionally they require special controllers as the signal outputted is completely different from the typical speed sensor. I can get the torque sensor and they're not overly expensive (but much more more expensive than a speed sensor, which costs buttons).

Yeah, the cadence sensor is such a simple device (for new e-bikers reading this in the future: it's an array of rare magnets spaced-out around the circumference of a small plastic disc mounted to the crank, with a fixed detector attached to the frame ... pretty much like a regular bicycle computer, only with several magnets at the chainring level instead of one attached to the rim spokes).

The Chinese manufacturer/supplier who I intend to buy a kit from says that they've had problems with the 5-speed display/handlebar control, so they are now recommending a 3-speed version. Yes, the kit includes not only a special 44T crank/chainring, but also a small controller. I am waiting for further technical specifications before I make my final decision about the purchase.

Reminder of the thread with many photos and details:

http://www.pedelecs.co.uk/forum/electric-bicycles/1172-home-made-pedelec-pedal-torque-sensor.html?highlight=torque#post62512

Cheers, Dan

 

[emissions-free]

Daniel,

keep us informed on the progress with the torque sensing crank. The other solution, wich uses a special bracket mounted on the rear dropout looks good and I can fully understand how it would be implemented too

In my previous life I used to fit strain gauges and have a reasonable knowledge of the science of the application. We used to use strain gauges from the no.1 strain gauge manufacturer, Measurements Group. You can also buy decicated ICs which supply the necessary excitation voltage and measure the output signal. With the right mechanical design of bracket, it's not so difficult to measure the strain generated. Both the strain gauge and electroics could be very small and very low powered, but a little expensive. However MG gauges are very high precision and that wouldn't really be required for measuring a bicycles torque, so much cheaper could be used. You're not looking for precision, just relative values so high precision is not required.

Cheers
Paul

 

[daniel.weck]

Interesting. The hard bit must be in calibrating the strain gauge input/output signals so that the riding experience feels smooth and "natural". This is where the current crank/hall-sensor solutions fail: from what I've heard, the torque-sensing hardware is fundamentally not that bad, but the integration with the rider/controller/motor is poor (jerky acceleration, unreliable behavior in slow-moving traffic).

If only there was an easy way to program the electronics in the controller, something like this little board (where the SD-card bootstrapper contains execution instructions):

http://www.pedelecs.co.uk/forum/electric-bicycles/4821-watts-up-vs-turnigy-power-meter-analyser-4.html#post64611

I've seen a handful of electronics "hackers" on Pedelecs.co.uk and on the equivalent German forum, and of course on Endless-Sphere ... unfortunately I don't have enough skill to contribute in that area. The subjects of chipset programming and torque-sensing devices have been debated already, but I have yet to see a workable solution (to be honest, it's probably because it's neither the "hottest" topic nor the most useful feature).

Cheers, Dan

 

[emissions-free]

I can't see why it should be so difficult to make the power feed in more gently, or to have some kind of lookup table, similar to how a cars ECU will control fueling and ignition depending upon load, rpm, air flow etc whereby the controller or PAS is not a direct link but is more of a fly by wire system, where the input signal does not direct control output, but rather is modified depending upon conditition/state. Different motor characterics could be accounted for etc.

For sure the technology is available to do this type of thing, it's just a matter of implementing it better and the motivation to improve. This is after all something that requires a lot of additional work but doesn't give any concrete improvements. It will not go faster, go for longer per charge or make the motor smaller/lighter. So what's the point if you just want to sell your motors and controllers?

 

[daniel.weck]

This is after all something that requires a lot of additional work but doesn't give any concrete improvements. It will not go faster, go for longer per charge or make the motor smaller/lighter. So what's the point if you just want to sell your motors and controllers?

My thoughts exactly. Thus why I'm interested to try it myself. See, I've just cycled-up a long progressive hill today, with medium traffic conditions and lots of start/stop moves. The Bafang's 400W peak with my current battery+controller is totally fine, but I wish I had the torque-sensing feature, because the cadence-based PAS only makes the motor kick-in after a few pedal revolutions, which is frustrating when trying to avoid being knocked-over by busy/oblivious drivers... As a result, I tend to use the thumb throttle to quickly get out of the way, which works fine but is totally counter-intuitive (I'm riding an assisted bicycle, not a moped ! )

Cheers, Dan

 

[Herb]

Can I suggest the adding of extra magnets to the pedal ring, this would fool the sensor into thinking move revolutions had taken place in any given time.

May help

Herb