Why would a controller cut out going uphill? (probably not battery related)

redmouse

10 W
Joined
Aug 3, 2017
Messages
65
Hi folks, so I got this cheap chinese "Enduro" frame ebike.
The problem is it's cutting out under heavy load, but only after couple of minutes of riding.

Typically it happens like this:
- I head out into a steep hill
- first 300m is fine, pulls hard
- then it stops going (but the display is still on and battery voltage is not low and still showing)
- I have to realease the throttle for 2s and then apply again, and it goes for another few seconds before cutting out again
- once I get to a less steep part or a flat, it goes just fine

I don't think it's the battery, because I have a voltmeter connected directly to it, and lowest it dropped was 45V - never to zero which would suggest the BMS cutting out. Also the low voltage cutoff on the controller itself is supposed to be 40V.
Also I tested this battery on another bike and ran it pretty low without any cutting out occuring.

I thought it could be heat protection kicking in - but - after it happens, the controller, motor and cables are barely warm to the touch (I'd guess right around body temperature). I'm not sure typical ebike controllers even have a heat sensor inside?

What would you suggest I try or look for?

The componets my bike has:
- "Enduro" chinese frame
- noname chinese hubmotor with 45mm stator, 26" rim
- noname controller, designated KT48ZWSRKD-RQ14L, rated voltage 48V, says 30A continuous, 60A peak, LVC 40+-1V. The controller is pretty long, but thin. I'm guessing it's 16-20 mosfets? Physically it feels pretty small for big bike like this and the pwer and phase wires feel a bit on the thin side as well. But what do I know. It doesn't get hot.
- battery is Li-ion 48V / 30AH / 1440Wh (it's pretty big, I'm sure it can handle measly 60A)
 
More likely it's the battery than anything else.

If the controller is programmable, then you can find out what the LVC is by setting it. It could have been mis-set from the factory.

If the power is cutting out at a certain voltage, then it's likely the LVC trippin'.

The power cut may so fast that the voltmeter doesn't update fast enough to catch it. Other indicators may have to do.
 
If you have a 52v battery and it's hitting 45v on hills, you probably have a battery pack that's inadequate to drive the vehicle. Or you have a good battery pack with multiple bad or weak cells.

You'd have to start tearing things apart to find out. Any warranty from the seller?
 
No, The bike's battery is 48V, but I have one extra separate battery from another bike that is 52V and I can use it on this one. Presumably it will never drop as low as the 48V one, so if it keeps cutting out regardless, I will know it's not the LVC.
 
One possibility is the controller / motor phase / hall combo is wrong; a false positive. Sometimes these will actually "work" and even appear to be normal, except that under enough load the system can bog down and overheat; the controller could be shutting down from that. Typically both motor and controller will get pretty hot in these cases, and be unusually warm or hot even in "normal" load situations.

So if voltage doesn't end up being the problem, you might check the phase/hall combo.


Similarly, a poor connection at one of the phases could cause a similar issue (and maybe even a poor connection at a hall).
 
Thanks guys for the advice.

So I definitely ruled out the battery or voltage in general: I did my test loop with the original 48V battery that was about half charged, rode until the bike started cutting out. Then immidiately switched to fully charged 52V battery (58.8V hot off the charger) and it had absolutely no effect - still kept cutting out.

I checked the controller temperature and it was barely warm to the touch, so were the power and phase wires.

I'm gonna open up the controller to see if there's anything apparently wrong, but I don't have high hopes. It's definitely not programmable, so nothing really I can do except replacing it.

The phase and power connections are very solid btw, so I think that's not it.
 
If the controller works with another battery, i don't think there is any problem with the controller at all.

We can't blame the LVC of the controller because the other battery's voltage is higher.

You've ruled out everything but the other battery. Unless the controller has a high voltage cutout, which i'd call extremely unlikely, as it would never spin the wheel up at all if that was the case.

China likes to cut corners at batteries the most. It's the first part i'd suspect.
 
Ok, more progress. I opened up the controller and it's apparently a KTE-18S5-2RX-A from www.szktdz.com. And it looks like it's programmable after all - I found this thread with a guy writing a software for it:
https://endless-sphere.com/forums/viewtopic.php?t=87870

I never done anything like this, that should be interesting :O
But I have to order a cable first, so that'll be couple of weeks before it gets here from Choina.

neptronix: I think we have some misunderstanding here. The controller malfunctions the same way with both batteries, and it seems to be correlated only with duration of riding under load. It's completely unrelated to battery voltage or charge.
 
I must have read your post wrong. Sorry about that then.

Yeah i knew that was a kunteng. Thankfully our forum has had a lot of people hacking on those lately :thumb:
 
No worries, thank you for trying to help :)

----


So I managed to dig up this cable I had laying around:
PL2303 PL2303HX USB to UART TTL Cable module 4p 4 pin RS232 Converter
http://s.aliexpress.com/JVvqE3UZ?fromSns=Copy to Clipboard

Can I use it for programming?

The cable has 4 pins: red, black, green, white

My controller has these wires going to the display: red (ignition), black, blue, yellow, green

How do I connect it? Anyone can recommend a good learning source?
 
ok, looks like I can't. Need to order "ST-link V2"
https://opensourceebikefirmware.bitbucket.io/windows_instructions/index1.html
 
Meanwhile, I tried relocating the controller outside for better cooling (it was under a cover). Combined with cold weather it seems to improve things slightly. It still doesn't even remotely overheat (it feels pretty much stone cold even after an uphill ride).

But it makes me wonder - could there be a component that may be overreacting to heat?

Next I will try heating up the controller artificially with a heatgun, and see if it affects things.
 
It's quite possible that it has a thermistor in it for over temperature protection, which may be reading an incorrectly high temperature.

NTC thermistors decrease resistance as they heat up, so a partial short somewhere would cause a high reading.

They're usually a ceramic disc with 2 leads and the letters NTC printed on them. Not always though.
 
Thanks! I'll try and look for something like that :)
I also took some pictures of the controller, posting here in case anyone sees something:

https://imgur.com/a/I7q7pZT

I was wondering maybe it's the brown-orange thing touching the yellow lead in the very first picture?
Wait, no, that's a 104J 100V capacitor.
 
Everything I see in your thread says the controller is reacting to overheating, or at least what it thinks is overheating.

My first hub motors, even before being on ES, so a long time ago, were brushed motors (BD36) and they had a real small controller that shut down if pushed too hard on a hill. The power had to be cut to the controller and then powered up again after a few minutes of cooling before the controller would work again. I probably experienced this over half of a dozen times.

One of those original BD36 motor and controller systems is still in use on a one wheel pusher trailer, in a 16 inch wheel, behind a hand cycle, and he can push it hard enough on a long climb to make it overheat and cut out.
 
Yep, just confirmed it's definitely controller temperature related.

Took a heatgun and made the controller nice and toasty, and it started cutting out right from the go.

Now to figure out where is the temperature sensing/sensitive component and make it behave.
 
More development:

I replaced all power wires in the controller with some beefy af thicc gauge wires (10 or 8 AWG) and beefed up all the traces. Didn't even manage to screw up.

Unfortunately it didn't substantially help the problem, controller is still cutting out.
Interesting thing is it doesn't cut out when I just blast full throttle uphil. It starts only when I'm taking off from a stanstill up a very steep hill on full throttle. Once I slowly pickup speed it takes a full throttle no problem.
Once again, the hottest controller got was about 45°C, probably lower. That's not a high temperature, right?

I'm guessing there is some system that cuts the power when there is high current and low rpms?

Things I consider trying next:
1. Reprogramming the controller (need advice with that - see next post)
2. Removing solder from the shunts to reduce max current
3. Lacing the motor into a smaller rim (maybe 16" moped)
 
I need advice with programming the controller. Can anyone tell which pins on the board connect to the programing cable of STlinkV2? Better pictures of the board here (with power wires removed, so easier to see:
https://imgur.com/a/w2XMfQr

On the website there are wiring instructions, but the example controller has different layout than mine.
https://opensourceebikefirmware.bitbucket.io/windows_instructions/index2.html
 
This has been a good discussion and I don't think you really have an issue now that you understand when it happens. Just like when you drive a car you don't go full throttle when you take off and on your bike when you go full throttle taking off up a hill all the juice that isn't turning the wheel is just going directly to heat anyway. Once you quit going full throttle from a dead stop and just give it enough throttle to accelerate you can pick up your speed just as fast without triggering the cutout mechanism.

Of course what I just said is pointless if you don't wish to finesse your takeoff skills and rather just go full throttle. In which case you probably need to find a controller that is built to take the abuse. :D
 
redmouse said:
Unfortunately it didn't substantially help the problem, controller is still cutting out.
Interesting thing is it doesn't cut out when I just blast full throttle uphil. It starts only when I'm taking off from a stanstill up a very steep hill on full throttle. Once I slowly pickup speed it takes a full throttle no problem.

Bingo. Your controller is momentarily overshooting it's current limit and the battery is responding by cutting power or you are hitting the LVC because your battery was never up to the task in the first place.

It's likely that it happens so fast that you can't catch it with a multimeter. But a cycle analyst or even $25 hobbyking watt meter would probably catch it.

You can shave the shunt on the controller if it has one to reduce current. However, changing a phase amp to battery amp ratio would probably help - though i have no familiarity with programming those controllers.
 
Well, whatever the explanation is, the bike is unusable in this state. It's not only annoying but also potentially dangerous becuse I can end up stuck in the middle of intersection with heavy bike that I can't pedal.

I'm going to try removing solder from the shunts next. It has more than enough power when it's cold, so reducing it a bit will hopefully help delay the crappy behavior without weakening it too much.
 
Yup, i see how that could be a big problem.

You should get a cycle analyst or hobbyking watt meter for your own uses in the future. It'll be like a very accurate gas gauge, tachometer, and battery -> controller diagnostic tool all in one.

Either device records peak amps / peak watts, and minimum volts... either of these figures would have answered the question before you asked it.

Remember that shaving the shunt will affect how many amps you program into the controller VS how many are actually delivered.
 
Thanks, that wattmeter sounds like a useful tool to have - I will put it on my wishlist, and once I stop being a cheap bastard I will buy it :D

Meanwhile, I took the bike on my commute and it handled everything like a champ, except one hill. So I'm thinking, if I don't manage to hack the protection system itself, I just have to boost the cooling.

What do you guys think - should I try active cooling? Cut a hole in the controller covers and put a fan on it that will blow air through. Problem is, the controller is mounted traditionally at the bottom where all the crap flies from the front wheel. So even with some shields it will be pretty nasty environment.

Another idea I like is - I got a leftover massive case off an old burnt out controller from a scooter. It's probably 3 kilos of aluminium, which I'm not super excited about, but it could be the ticket. The original casing feels like 500g, so this should be a substantial upgrade.
 
WTd0W47.jpg

You're a big guy...
 
Back
Top