KU93 with Bafang CST 250w (350w) inadequate for hills

Joined
Sep 12, 2014
Messages
50
Hi all,

My noisy, whiny, slow-ramping 23a ku93 9 fet controller, isn't quite powerful enough on throttle only to get my Bafang CST 250w (180mm) up some local hills, and I was wondering if the sine wave s12s with 12 mosfets would be any better. The s12s used to be 23a apparently but now it's 25a +/- 1a. Am I just pouring money down the toilet here,ie is there any improvement in power to be had from the s12s, or will the motor only be able to use 22a as some have said, and the extra amps be wasted, shortening my range, (and perhaps generating harmful heat)?

Is there another controller which will get more out of the CST 250w, or am I really looking at another hub motor? I'm considering going over the legal 250w nominal, to allow me to use a bike freely despite my knee problems, although am worried because I hope to qualify as a lawyer at some point in the near future, and any criminal convictions could prevent that. If another motor is the only answer, would anybody have any recommendations please?

Thanks :)

Edit - using a 36v battery, 14.5ah 10s x 5 Samsung ICR18650 29E LiNiMnCoO2 cells (3.7V 2.9Ah).
 
First off, the Bafang CST you are referring to is rated 350 Watts, not 250. It can handle up to three times that for short bursts, so the motor is not holding your system back.
BMS Battery does not recommend the KU93, they recommend S09P, S12P or S12S controllers.
Of those, the S12S is most commonly used, especially in Europe.
If you are feeling "techy", you can use it with the SLCD-3 display, but this display requires some extra installation and programing time to get everything working right.
Perhaps the SLCD-1 would be a better choice.
Although the KU93 isn't the recommended controller, it should power that motor just fine.
What is your battery?
 
The battery is a 36v, 14.5ah with 10s x 5 Samsung ICR18650 29E LiNiMnCoO2 cells (3.7V 2.9Ah).

This is marked and sold as 250w although as you say it's a 350w. D8veh tested one of these 350w CST/BPMs at 22 amps and at 30 amps and found there wasn't an increase in power at the latter current. I don't know what winding code that had though.

Could mine have a high RPM, hence the hill problem? It is marked CST36V250W 26. It was sold by Elifebike as being the 201rpm model - yet oddly with my 15a 6-fet minicontroller (which it was supplied with) I get 22mph out of it on the flat at 42v in throttle only, and 25mph+ easy on the flat with a bit of pedalling in PAS mode. Could that be a bit high for 201rpm? Although the KU93 9 fet will pull a max 15-16mph on the flat and goes down to 11mph on moderate hills - with 860w peak recorded on the wattmeter.
 
jonathan75 said:
The battery is a 36v, 14.5ah with 10s x 5 Samsung ICR18650 29E LiNiMnCoO2 cells (3.7V 2.9Ah).

This is marked and sold as 250w although as you say it's a 350w. D8veh tested one of these 350w CST/BPMs at 22 amps and at 30 amps and found there wasn't an increase in power at the latter current. I don't know what winding code that had though.

Could mine have a high RPM, hence the hill problem? It is marked CST36V250W 26. It was sold by Elifebike as being the 201rpm model - yet oddly with my 15a 6-fet minicontroller (which it was supplied with) I get 22mph out of it on the flat at 42v in throttle only, and 25mph+ easy on the flat with a bit of pedalling in PAS mode. Could that be a bit high for 201rpm? Although the KU93 9 fet will pull a max 15-16mph on the flat and goes down to 11mph on moderate hills - with 860w peak recorded on the wattmeter.
with my 15a 6-fet minicontroller (which it was supplied with) I get 22mph out of it on the flat at 42v in throttle only
I plugged those spec.s(with 26"wheel) into the Ebike CA sim. and got 265 rpm @36V.
Also, those results are exactly the same I used to get with my geared MXUS 260 (350 Watt).
It seems that something is wrong with the KU93.
A 270 motor speed CST is what BMS B. sells and it's a good compromise.
I think once you get a decent 22 to 25 Amp controller on it, it will climb moderate hills well.
 
Thanks. 265 rpm explains a lot.

I feel pretty sad about this as I deliberately made sure I chose the 201rpm model because my knees cannot cope with hills (or even flats) and the hills are more than moderate, many of them, where I live, and where I've been hoping to travel. Pretty angry at Stephen Deng and Elifebike.

But yeah I might get an s12s and see if it demonstrates that there is a problem with my KU93 :) So thank you for your suggestion :)
 
The late model KU93 unit, the one with the optional control panel, does have soft-start as you noted and unless you use the control panel it also defaults to the mid-speed setting. I don't use the control panel but I did install an internal jumper to enable full speed. I use one on my main bike and like how it performs with my small geared YOUE motor. It's smooth and quiet (well as quiet as any square wave controller I have) and works great in sensorless mode. Mine peaks at 21-22A. I do however use a 12S LiPo pack (50V max) with it which makes a world of difference compared to a 36V, especially an 18650 battery pack which sags much more than LiPo.

The older KU93 models did have instant start. I used one with the shunt soldered up to allow 26-27A peak on my Townie with a code 10 (300RPM) Bafang BPM motor (very similar to your CST) and 36V high C-rate battery from BMSBattery and found it performed very well. With a 48V downtube battery (w/Samsung 29E cells) it was quite strong. With it I recorded my best speed up an 8% hill that's on a route I frequently ride cresting that hill at 21 MPH (with pedaling). I recently switched to a 19A 9-FET "normal controller" from ELifeBike because 27A is a bit much for the downtube battery pack. This controller is like the old model KU93 with instant start. The bike is substantially less powerful now but the BPM generates nice torque and is still satisfying to ride.

So in short; make sure your KU93 is running at normal speed instead of the default middle setting and try soldering up the shunt for a few more amps. The best thing you could do however is switch to 48V, it makes a huge difference.

-R
 
My 350w CST was slower than the 260 rpm 500w one, but when I shorted the high-speed side of the three-speed switch on the KU93, it went up to 33mph (no-load), though I coldn't detect much difference on the road. Is your speed with or without those wires shorted?
 
jonathan75, I don't want to sound harsh, but I have gone back and read all your posts trying to get my head around your problem and with your multiple thread postings, each, it seems with different componets, has me thoughly confused.
For example, after buying a kit from fairly reliable vendor Elifebike, you order a controller from Alibaba.
Now you have a KU93 controller, which, I'm assuming came from BMS Battery(and the Battery with Samsung cells, which seems somewhat rare, must have come from Alibaba as well).
Is the KU93 controller the one you are trying to make work right now and did you order it with the 810 display?
If the answer is yes and no, then I think russell has explained the problem, you need to do an internal mod or get the 810 display.
Russell also mentioned the samsung cells, you should ck the continous discharge rate to make sure that a higher current controller won't trip the batteries BMS.

As far as the motor speed, I wouldn't fret too much over that.
The difference between a low speed geared motor and a mid speed is not so great as to render the mid speed useless for hill climbing. It's when a high speed motor is installed in big wheel the performace really suffers.
Even a 260 mini motor on 36V will climb steep(but short)hills.

I guess my point for any potential ebikers looking in, is to do the research first and also pick one vendor, research them and stick with them.
The folks at ELB and BMS Battery are getting better at commuicating what works together and can be a help when ordering.
 
jonathan75 said:
The battery is a 36v, 14.5ah with 10s x 5 Samsung ICR18650 29E LiNiMnCoO2 cells (3.7V 2.9Ah).

This is marked and sold as 250w although as you say it's a 350w. D8veh tested one of these 350w CST/BPMs at 22 amps and at 30 amps and found there wasn't an increase in power at the latter current. I don't know what winding code that had though.

Could mine have a high RPM, hence the hill problem? It is marked CST36V250W 26. It was sold by Elifebike as being the 201rpm model - yet oddly with my 15a 6-fet minicontroller (which it was supplied with) I get 22mph out of it on the flat at 42v in throttle only, and 25mph+ easy on the flat with a bit of pedalling in PAS mode. Could that be a bit high for 201rpm? Although the KU93 9 fet will pull a max 15-16mph on the flat and goes down to 11mph on moderate hills - with 860w peak recorded on the wattmeter.



Was re-reading your posts and see that the speed is lower with the KU93 which tells me you definitely don't have the KU93 running at 100% speed. If you do not have anything connected to the '3-speed switch' connection on the controller it will run at a lower speed (70-75%) which is exactly what you are seeing. Installing a 3-speed switch (or make a jumper or cut and twist the correct two wires together) is necessary to achieve full speed (note this wasn't always the case, disconnected used to be 100% on older controllers). If you use the 810 display then this is all you should have to do to get the proper speed. I don't use the 810 LED display and found I had to install another jumper with a resistor between the green and black pins on the controller plug meant for the 810 display in addition to the jumper between red and blue to power it up (I installed all jumpers internally so I'd have fewer wires exiting the controller but external is fine).

KU93_810.png

Once you get the KU93 set properly you should achieve the same speeds you did with the 6 FET controller. The added speed capability should help a little up hills but even 800W input isn't going to give great performance up hills if you don't pedal (even pedaling lightly helps a good deal). The BMS in most of the DT batteries can't handle more than 30A peaks and really 22-23A is pretty much spot on what you should run as a max with them. If you want to push the limit though you could add a touch of solder to one of the shunts in the controller to get a couple of extra amps and maybe another 100W or so.

-R

PS, If you need an 810 LED panel for 36V operation I have one I'm not using.
 
Thanks Russell, that's terrifically helpful. I bridged two of the pins on the 3-speed connector as you recommended, and it is indeed about 25% faster now, and seems to keep its speed quite a bit better on the hills. I think it could be a viable controller now.

Might be a little thirstier too, 19wh per mile averaging 13.5mph on a range of moderate gradients over 5 miles seems kind of high - hopefully when I mount the wattmeter on the handlebars I'll be able to work out what are its sweet spots for energy consumption.

Wow - what an accelleration lag, though: if you want to increase speed you have to plan about 5 seconds in advance. That's like a big boat. I suspect on PAS mode (which actually doesn't work, I think because the LED display isn't available) there is better accelleration when pedalling a little. Hope so anyway.

Your offer of an 810 was really kind and that'd be brilliant - thanks. I can PM you with my details.

Jonathan
 
The delay on my KU93 from a dead stop is maybe 2 seconds and when already in motion maybe a second. I only really notice it when I need to cross a street quickly, then it can seem like forever when traffic is bearing down fast.

Your power use does seem rather high for that average speed. My highest power consumption this season so far has been 19.0 Wh/mile with an average trip speed of 18.2 mph with my Townie cruiser with very little pedaling. With my other two E-bikes I usually consume 9-11 Wh/mile at an average trip speed of 18 mph with moderate+ pedaling effort. I suppose though if you ride something like my Townie and don't help at all up hills and accelerate at full throttle you could use that kind of power.

Is the motor hot after a normal ride? Is the controller hot? Aside from initially accelerating from a stop or lugging up a hill does the motor make much noise (it shouldn't)? If it is rather noisy all of the time check all of the motor connections, phase and hall sensors wires. You can also disconnect the hall wire plug (5-wires) and see how it runs sensorless. If it improves then there's likely a sensor problem.

If everything seems normal but the performance is not to your liking you'll want to increase the input power to the motor. With your present battery it would be wise to limit the max current to about 25A, but even a couple of extra amps can be felt. Either do the time tested shunt solder trick on the KU93 or buy a new controller, one with instant start this time. I would recommend one of the neat little 6-FET IRFB4110 or 3077 25A controllers from Cellman ( http://em3ev.com/store/index.php?route=common/home ). I presently use one of these (detuned to 23.5Ap) on my Gravity bike with a 48V downtube Samsung cell battery.

If you decide to stick with the KU93 and want the 36V 810 LED panel just send me your name and address and I'll get it mailed out to you.

-R
 
Russell said:
So in short; make sure your KU93 is running at normal speed instead of the default middle setting and try soldering up the shunt for a few more amps. The best thing you could do however is switch to 48V, it makes a huge difference.

-R

Thanks Russell. Does the follow scenario sound like likely overheating at 48v? Motomech helpfully pointed out it looks like my Bafang CST 350w 36v motor is at least 265rpm (not the 201 I specifically paid for, thanks Elifebike...) at 36v. I gather that means at 48v it would be 352rpm - I have a bad feeling that means overheating on hills on throttle-only. Here in North London it's very hilly and there are often traffic jams on the narrow & antiquated roads, which means you often have to move very slowly through them. It is better for my knees (or nerves in them) that almost always don't pedal.

Your comments about checking for noise etc were also very helpful - it has indeed been noisy, between 0-4mph (even at no load) a pretty brutal resonant booming whine, and then makes a slightly irritating undulating high-pitch whine at speeds above that. I haven't yet checked sensorless mode, will do shortly.
 
Moving up to 48V will increase the overall motor power and torque output and yes it will increase the maximum speed of the motor. It is YOU however that control the speed via the throttle so if you don't want to go faster then don't. If the throttle becomes too sensitive after upping the voltage you can install a 3-speed switch to limit top speed. You will still get the power boost at lower speeds.

Small low-power geared hub motors are great for assist . Up hills however they simply can not output enough power and torque by themselves to climb steep hills. I love my 250W-350W geared motors running on 48V at 21-23A (1000-1100Wp)but then I pedal all of the time with at least moderate effort. Using them alone without pedaling however is a different story. They feel over-taxed and under-powered, because they are. My cruiser with the larger 500W Bafang BPM, again running on 48V/20A, just begins to feel like I could ride it like a scooter, but I still need to assist it up steep grades. Running it at higher currents and higher peak power of around 1300W+ is even better.

Basically what I'm saying is if you can't pedal you are going to need an E-bike with significantly higher power than your existing set-up. Your downtube 36V battery, controller and motor are all limiting factors however as I stated previously you could increase your current a few amps either by the solder shunt trick or purchasing a new controller. That won't make much of a difference however so the next logical step would be a 48V battery and 25A+ controller. That will provide a meaningful boost. Whether your CST can handle it depends really on your terrain. If you are constantly going up long grades then maybe not. Do remember that ANY pedaling you do up hills can help considerably by providing just that little extra torque. If the CST does bite the dust then upgrade to a larger motor.

-R
 
Russell said:
Small low-power geared hub motors are great for assist . Up hills however they simply can not output enough power and torque by themselves to climb steep hills.

Be careful with statements like that. I would call the Xiongda a small motor. The 48v version is like a winch when it comes to hill-climbing.
 
d8veh said:
I would call the Xiongda a small motor. The 48v version is like a winch when it comes to hill-climbing.

Slightly off-topic of me but will the 48v Xiongda work with a 36v battery? I realise there will be a low voltage cutoff issue but can that be overcome through altering the 48v/36v setting in the LCD display? Any other issues?

Ty
 
You could use a 48v motor with the 36v controller. It'll be very slow though, at about 15 mph max. In high gear and 8 mph in low gear.
 
d8veh said:
You could use a 48v motor with the 36v controller. It'll be very slow though, at about 15 mph max. In high gear and 8 mph in low gear.

Thank you. 15 is perfect actually - could the 36v controller run off the same lcd unit as the 48, so that I could unplug and replug when swapping 36v and 48v batteries? I've ordered a 48 but would like to be able to use the 36v too.
 
d8veh said:
I don't understand what you want to achieve.

Apologies, that wasn't very clear of me. I'd like to carry and use on my bike alternately on a given tour, both batteries I now own, a 36v and 48v. So I'd like to be able to switch over from one to the other when one runs low.
 
It's not really practical unless you have controllers and displays that are automatic dual voltage. Even then, they can get a bit mixed up if you change fom 36v to 48v. It seems that they have some sort of memory so that they can deal with the difference between a disharged 48v battery (38v) and a charged 36v one (42v).
 
d8veh said:
It's not really practical unless you have controllers and displays that are automatic dual voltage. Even then, they can get a bit mixed up if you change fom 36v to 48v. It seems that they have some sort of memory so that they can deal with the difference between a disharged 48v battery (38v) and a charged 36v one (42v).

Oh dear - I think you're right. I've looked at the LCD-3 instructions and there doesn't seem an easy way around that. It seems to automatically detect voltage. It would be used with an S12S controller (for my non-Xiongda motor setup, at least). What a pain. I can't think of an easy way around this, except for making sure my batteries are the same voltage. A dual-controller setup seems possible to enable me to ride with batteries of different voltages, with my existing 36v controller, but would involve either a very large custom-made switch with lots of wires coming out of it, or having to stop by the roadside every 25 miles to unplug all the wires and plug them into the other controller. I would have to make some quick release connectors for the motor wires and for my brake cutoffs.

Does that sound very impractical? This is to enable me to make a European bike tour. Perhaps I should sell my 36v so I have two 48v.

Thanks D8veh.
 
The late model KU93 (the one with the 4-pin LED plug) is dual-voltage 36V/48V. The only problem you would run into is if you use the 810 display which is calibrated for either 36 or 48V. I don't use the display with mine so there was no problem going back and forth between my 36 and 48V batteries.

You could also use a basic 36V controller with either battery as long as it has 63V or higher main caps. The only downside is you don't get LVC protection from the controller when using the 48V pack but I assume your packs have their own BMS. If you want a display with this setup you can use a Cycle Analyst which doesn't care what voltage you run. A simple wattmeter in combination with a regular cycle computer is another option.

-R
 
Russell said:

Thanks Russell - after reading your post I opened up the 36v 18a minicontroller and it turned out it does have 63v main capacitors. Just been on a 20-mile run, first run with the new 48v battery, charged fully. Worked very well. Ate up the hills. Is your 11.6ah battery still going OK? I think this is exactly the same one.

The display was flashing empty on some hills towards maybe 16 miles, and I'd not exceeded 16mph during my run, and tried to avoid max throttle at all times. Flashing empty seemed odd because the minimum voltage off the wattmeter read 45.5,which doesn't sound that low compared with the 40v which some LVCs are set for on controllers for 48v batteries.

Interested in what you wrote about limiting amps helping to extend usable range. Going to open up a 10a 36v controller in a mo to see if it will run the 48v battery, could get me a few more miles.
 
My 11.6Ah downtube battery (w/Samsung 29E cells) performs better than I expected. The most I ever drained it was 10.9 Ah/497 Wh using a 15A controller. The minimum voltage was 37.85V with a resting voltage of 43.04V at the end of the ride. The Battery's BMS never did trip even though the pack was clearly almost completely empty.

If your wattmeter is correct then your pack wasn't anywhere near empty. What wattmeter are you using? Did you record the resting voltage of the pack along with the Vmin at the end of the ride?

A 10A controller is going to feel very weak over the entire ride so I wouldn't bother just to try to get more range. The best thing you can do for range anxiety to know the battery's capacity and stay well within it. One way you do that is with a battery so big you never have to worry. The other way is to use an accurate wattmeter so you can keep tabs on the battery and turn for home before it gets too low. I would even suggest purposely riding until the pack is fully drained one time so you can get the capacity data for it. That will give you confidence on subsequent rides. A log of your rides with pertinent data from your wattmeter (Vrest,Vmin,Ah used,Wh used) and speedometer (miles,average speed) is useful too.

One more thing, every so often leave your battery on the charger a good long time after the green LED turns on. This will allow the cells to balance as best they can assuring maximum capacity.

-R
 
Back
Top