Shutoff problem Q100H with KU93 at higher RPMs

eagle_eye

10 mW
Joined
Feb 22, 2013
Messages
24
Location
The mountainous part of Austria
Hi all
I've got a new Q100H hub front motor (-> http://www.bmsbattery.com/front-driving/630-250w350w-q-85100sx-motor-e-bike-kit.html) from bmsbattery plus a KU93 controller with a Wuxing brand type half grip throttle (-> http://www.bmsbattery.com/accessory/390-thumb-level-throttle.html).
The batt. packs are 36V and 48V nominal and can deliver 30A plus continuously.

I'm observing strange behavior in medium and highest speed mode (3-speed-switch) - when slowly going to and beyond roughly 2/3-throttle, the controller suddenly shuts off the hub motor. It happens in no load and low load situation, drawing rather minor currents only. It seems like the controller can not handle higher rpm of the Q100H, its shutting off LVC style. After turning the throttle to zero and then from zero up the motor works again until going to higher RPMs. It cant be LVC shutoff cause the batt.packs is not going low voltage - its happening even with no load and fully charged batt.packs.

I also did the KU93 'shunt mod' soldering out one of the two shunts to limit the current to ~ 12A max. but its not changing the shutoff behavior described.

So what could be the cause that the KU93 controller suddenly shuts off the Q100H hub motor at higher speeds ? Anyone using a Q100H hub with KU93 controller and 36V/48V batt.pack going up to high speed RPMs w/o the motor shutting off problem ? It cant be possible that the KU93 controller is too slow for the Q100H, can it ?!

Thanks for any help
 
The controller cannot handle the high eRPM of the motor. It's got something like a freaking 12:1 gear reduction ratio!!
That is why your problem gets worse the faster you go.

BMSbattery has some recommended controllers for that motor. The KUxx series is not on that list.
Alternately, any newer infineon-type controller ( from lyen or em3ev ) could very well work fine with that motor at high RPMs, because i know for certain that those will control a MAC/BMC ( very high pole count ) motor with about twice the pole count as yours - one of the most difficult geared motors for any controller to deal with.

ps; how does that motor perform on 48v? and what size wheel are you using?
 
Hello

Have you connected the hall sensors? If not, try doing so. IIRC the KU93 can run in both sensorless and sensored configurations, and from experience (though not with this controller) rpm limits are lower in sensorless and much higher when using sensored configuration.
 
If it is this one ;
http://www.bmsbattery.com/controller/363-universal-brushless-hub-motor-controller.html

The tittle say high speed motor the same as the KU63 I don't see why it shouldn't work.

Actually the KU93 controller is one of the few controllers I've heard of that can handle the MAC motor in sensorless mode.
 
Many thanks for the useful infos guys!

Did forget to mention I have the 260rpm version of the Q100H hub (-> http://www.bmsbattery.com/front-driving/630-250w350w-q-85100sx-motor-e-bike-kit.html )
and I'm using it laced into a 26" MTB rim (Rigida/Ryde Big Bull) for the front wheel ..

Indeed the newer Q100H has a significantly higher reduction ratio ( 12.6 ) than the Q100 ( 8.1 ) - that is over 50% more (!) So if the KU93 electronics is fast enough for conventionally Q100 hub that does not necessarily mean its sufficient i.e. fast enough for Q100H.

Was asking the bmsbattery service whether KU93 controller is compatible with the Q100H hub and they just said no, giving no detailed reason why not. I was assuming that its just because of the max current limit of the KU93 (22A ) being to high and since I was planning the do the 'shunt mod' to reduce max current to ~ 11-12A I thought ok. But maybe its not only the max current limit but the KU93 being too slow internally for Q100H also - heck, these China bmsbattery customer service folks tend to really starve you pertaining tech infos :-/

So I just asked them via mail whether KU63 controller would be fully compatible with the Q100H and if not then why not ... we'll wait and see

@ neptronix:

BMSbattery doesn't seem to have an appropriate 48V enabled controller for the Q100H..

In medium RPM range the Q100H works quite nicely with the KU93 controller, no problem with 48V batt.packs. Also at more load uphills its working nicely and relatively torquey while staying surprisingly cool and its never shutting off - both with 36V and 48V. BUT on level terrain I experienced that with 48V batt. the motor shutoff with increased speed and motor RPM occurs at medium 2nd speed-switch level already, with 36V batt.pack it only happens at the highest 3rd speed-switch level. So it really seems that the Q100H trips the KU93 'High speed' controller at higher motor RPM :-| :roll:

@ cheys03:

I'm intending to use it fully connected with all hall sensors. After installing the hub motor I did some re-wiring and soldering to shorten the wires for phase and hall sensors appropriately and to my knowledge everything should have been re-connected as intended. But I am checking the hall sensor wiring again just to make sure..

@ migueralliart:

Yep its exactly that KU93 controller - dubbed 'High speed controller' :roll:
They also state in the item description "It is suitable for all our motors. It is compatible with sensorless and hall sensor motor. Voltage: 36V and 48V are compatible." Also reading more stories of Q100 hub with the KU93 controller working w/o problems me was thinking KU93 together with the Q100H should work ok in every aspect.


It would be useful if anyone could possibly report the successful usage of the Q100H 260rpm hub together with a KU93 controller, preferable powered with 48V batt.pack - and at high RPM :)

On the other hand what options are there for an alternative controller for a Q100H with attributes like
[*]reasonably priced
[*]max current limit ~12A
[*]for 36V and 48V batt.packs
[*]preferably rather small in dimension ()

Thanks and regards
Al.
 
I would try a lyen/em3ev 6FET controller as an alternative. If i can drive my MAC to 45mph without stuttering ( 16 pole motor, 5:1 reduction; surely your motor has half the poles ), then it's very likely that you'll be able to drive the Q100H well.

Do you have the 268rpm at 36v model, by chance?

If so, on 48v, you are going to be in the low-mid 300rpm area. Let's say 320rpm under load. so 320rpm x 12.6 = 4032..
That's almost RC motor territory!!

Now, i hit 45mph on my MAC with a cell_man/em3ev 12FET.. that's 560RPM at the wheel x 5 = 2800rpm, but since i have easily twice the poles as this motor, the controller would be working at least twice as hard, so let's think about that as 5,600rpm.

I think a lyen/em3ev 6fet is certainly worth a try. At the least, it would be useful for another motor in the future. Great controller.. programmable and solid.
 
The Ku93 isa high-speed controller, and should have no problem with the Q100H. It's fully automatic, so will switch between sensor and sensorless mode depending on how it detects the hall sensors. Certain problems with them will make it switch to sensorless mode. Unfortunately you can't lock it in sensored mode to see if there's a problem.

Check the hall sensor connections carefully at the controller end to see if a pin has backed out or any other problem. Make sure that the multi-pin connector at the motor is all the way in as far as the guide line.

I don't understand the bit about soldering out a shunt. Normally, you cut it in the middle to disable it and halve the current. soldering it with give you 30 amps total, which will melt the motor.

I'm not convinced by that reduction ratio. The speed feels and sounds the same as a normal Q100. I thought the increased speed came from the windings. changing from 8:1 to 12: 1 would decrease the wheel speed, not increase it. Something doesn't add up.
 
You can finely adjust the shunt via tiny dabs of solder, or by shaving it somewhat. Does cutting it in half really cut the amps in half? have you tested that? otherwise i would imagine that the controller would just fail to operate, unless there are two shunts and you cut just one.

What does high speed mean, exactly? too bad they don't tell you the eRPM limit for each controller, then you'd know for sure.

Perhaps the Q100H/Q128H have less poles, which may explain the similarity in sound?

I've strongly been considering ordering one of these motors.
 
They recommend The S-series controllers.

If you really want to give it over 20 amps, then how about the s12s it's a sine wave controller of 21.5amps. There is also the similarly powered s09p but as it is square wave the shock loading will be audible.

Going out on a limb here.... The sinewave controllers smoother delivery is not going to kill gears off like a square wave one. That much is fact. How much better for them is it though? Could 22a of sine be about as bad for the gears as 17a of square?
Does it even matter, if the windings were going to be the weakness not the gears.

iirc the older q100 would stand 750w and I can't see this dissipating heat any quicker. That 30% torque increase has been called out by a few mathematicians. It's a peak gain made by taking power from elsewhere. The motor efficiency stated is virtually the same. However I would also be tempted to push the envelope too, if I had one. I would get the s12s. It says 21.5a but I dunno if that is peak as I would hope, or average. If it is peak, as it should be, then average is very likely to be lower, even at full power. Didn't the old one hold 17a? I don't think 21a is going to cause any behaviour that an expecting eye won't catch in time. The next controller down is the s06s and s06p but I think you want to start high and cut back, rather than start low and get soldering.

If I really fancied myself I would get an s06s and swap the fets then mod for more currant using a resistor not solder. I could find myself getting quite excited by that. A controller that fits inside these commercially available battery boxes. I really like that my frog-box has space for a controller in the bit that stays on the bike. It was not even advertised as such. Like the big slabs, they too have space for a 6 fet controller. They also have a specific frame box available for the 6 fet. It's just blinking marvelous, I'm going to chuck my ker-plunk s12s away and get one myself, right away.
 
I doubt that the couple of watts incorrectly fired into the stator make for any appreciable extra load on the gears.
The reality is that high amounts of amps from a stall is most damaging. Shock load being sudden harsh loads, not small amounts of vibration as the motor moves, actually has a real effect.

I've not heard of anyone killing gears on these motors, so let's get to that discussion when it happens.

Why would square wave commutation hurt a series of copper wires? Ever heard of that happening at all?

If I really fancied myself I would get an s06s and swap the fets then mod for more currant using a resistor not solder. I could find myself getting quite excited by that.

What would happen is the exact opposite of what you're thinking. A shunt senses voltage drop over a conductor. When you add solder, you reduce the resistance and thus the controller is fooled into putting out more amps. If you were to increase the resistance, you would get less amps.

Also, you can't expect a 6fet to run like a 12 fet with a shunt mod. It will create more heat from the FETs and other components. Maybe a few extra amps might be okay, but not more than that on such a small controller.
 
ps.. i have a random question. on bmsbattery's site, they show a weird 9 pin connector.. did your hub motor actually come with that, or a standard hall/phase type connector that comes with every other motor?
 
neptronix said:
I doubt that the couple of watts incorrectly fired into the stator make for any appreciable extra load on the gears.
The reality is that high amounts of amps from a stall is most damaging. Shock load being sudden harsh loads, not small amounts of vibration as the motor moves, actually has a real effect.

I've not heard of anyone killing gears on these motors, so let's get to that discussion when it happens.

Why would square wave commutation hurt a series of copper wires? Ever heard of that happening at all?

If I really fancied myself I would get an s06s and swap the fets then mod for more currant using a resistor not solder. I could find myself getting quite excited by that.

What would happen is the exact opposite of what you're thinking. A shunt senses voltage drop over a conductor. When you add solder, you reduce the resistance and thus the controller is fooled into putting out more amps. If you were to increase the resistance, you would get less amps.

Also, you can't expect a 6fet to run like a 12 fet with a shunt mod. It will create more heat from the FETs and other components. Maybe a few extra amps might be okay, but not more than that on such a small controller.

It's good to put the gears idea aside. I also think people soften them with heaps of power, then shock load them doing silly things like letting it spin up as they get some air. Thus causing almost instant deceleration upon landing. I was typing pissed, as most days.
Wave effects on the windings are not on my mind. Although they could get more shock loading from the square wave, it is really a timing issue that controls this. As yet I have seen no evidence of it being a problem, but now I think about it maybe I have seen winding weaknesses exposed that might of stayed hidden with sine use. Mid wind unexplained shorts that could well be driven by coil movement. Remember the force upon the magnets is on the coils. They get pushed about. The audible vibration is playing a part.


Your presuming the resistor goes in series. It would have to be a very big one to do that. A 5w in parallel should lift the 15a upwards of 20a and the new fets lower resistance means less heat from them per amp. I'm unsure if they are a third less resistive, but the standard ones will carry 18a so I'm sure I can get past 20a. I have the docs here to calculate the smd resistor values to go past 20a with a stock 6 fet, so swapping them for more suitable one's is really a forgone conclusion.

Don't people make a 6fet that is 25a, Or is that totally different?
 
Hi,
I am from Germany, sorry for my bad english.

The new KU93 are different from the old ones. I had the
same problems here. I tested them with all Q100 motors,
front, H, CSt. All motors 328 rpm. The new KU93 stop
all motors.
I tested with a 201 rpm rear. The new KU 93 was working.

I have some pics of the new inside here, but I don´t know
how to upload
 
neptronix said:
ps.. i have a random question. on bmsbattery's site, they show a weird 9 pin connector.. did your hub motor actually come with that, or a standard hall/phase type connector that comes with every other motor?

If it helps I have a Qtype motor from BMSbattery purchased over a year ago. The weird 9 pin connector is both phase wires and hall wires. I couldn't tell you the pin layout though. D8veh may be more familiar with it though. :arrow: You can get a better picture of the wiring once you get closer to the controller and the JST plugs. So if you wanted to diagnose wiring issues, you could do it there.

A KU93 controller with a "12A" limit sounds wrong. That's the rinky dinky KU63 (I have one). My guestimate on the KU93 is 15-20A limit. We need to clarify the shunt modification you did on the controller. You added more solder? Or did you remove more? From your original post, it reads like you removed, but that doesn't sound possible. Take pictures and post them?
 
wepfi said:
How to upload pics from notebook? I have some!
http://endless-sphere.com/forums/viewtopic.php?f=13&t=14748

^stickied from the Board index^
 
melodious said:
You added more solder? Or did you remove more? From your original post, it reads like you removed, but that doesn't sound possible. Take pictures and post them?

He didn't do anything with solder except melt it to totally remove one of his two shunts. While some controllers have one shunt, others have two in parallel. Pulling one out halves the power.
 
Ok ---- I did not understand how to upload pics
from my computer ! Only from public sites.
But:
I had contact to ecitypower.
I sent the pics from inside
new and old KU93 to them.
Xie from company has no idea !
 
neptronix said:
You can finely adjust the shunt via tiny dabs of solder, or by shaving it somewhat. Does cutting it in half really cut the amps in half? have you tested that? otherwise i would imagine that the controller would just fail to operate, unless there are two shunts and you cut just one.

What does high speed mean, exactly? too bad they don't tell you the eRPM limit for each controller, then you'd know for sure.

Perhaps the Q100H/Q128H have less poles, which may explain the similarity in sound?

I've strongly been considering ordering one of these motors.
The two shunts are in parallel, so cutting one right through doubles the resistance. The controllers measures the voltage drop, which will be double, so the current will be halved.

High speed means high enough for a 328 rpm motor. I don't know what the absolute maximum is.
 
Thanks for the input guys

neptronix said:
ps.. i have a random question. on bmsbattery's site, they show a weird 9 pin connector.. did your hub motor actually come with that, or a standard hall/phase type connector that comes with every other motor?

out of the 100H hub comes a common cable approx. 30cm long - then there is the ominous 9 pin connector (male-female) (seems to be water tight :wink: ) and then more cable approx. ~ 1m long ending with the 3-pase wires and single hall sensor wires.

..looking just like that

qswxk-front-drive-brushless-hub-motor-for-ebike.jpg


melodious said:
A KU93 controller with a "12A" limit sounds wrong. That's the rinky dinky KU63 (I have one). My guestimate on the KU93 is 15-20A limit. We need to clarify the shunt modification you did on the controller. You added more solder? Or did you remove more? From your original post, it reads like you removed, but that doesn't sound possible. Take pictures and post them?

The current KU93 I have does have 2 shunts in parallel

index.php


friendly1uk is exactly right
I removed one of them - the outer one - by carefully soldering it out of the controller circuit board. I had to apply some pressure locally on the shunt-pins to get it out so one needs to be careful when doing this. This should reduce the max. current limit by 50% so the nominal max current limit of ~ 22A should be cut roughly by half.

I need minor power for that Q100H front hub only because its on a AW-drive setup with Q100H in front and a MAC 255rpm 10T hub at the rear - with its cell_man/em3ev 9-FET Infineon controller. They are on a common throttle and batt.pack but each has its own controller and 3-speed switch. With both 3-speed-switches I can accentuate the power priority between the two motors quite comfortable. The MAC is quite strong already but the Q100H can significantly reduce strain on the MAC when going uphill for longer. Also if driving with low-power in level terrain at the city the front Q100H can do virtually do the job alone sparing the rear MAC almost completely.

One quick shot of the newborn AWD beast :D

AWD-q100h_260-mac_T10.jpg


and please dont mind the further interior
:mrgreen:
 
I had the very same problem with my ku93 controller on a Cute 100 201 front motor (12 amp limit with one shunt cut) - cutting out completely at over 95% of full throttle. I replaced the controller with a 48V 15Amp controller from ECrazyman off of EBAY and it seemed to solve the problem...although this new ECrazy controller has its own minor issue - surgy on lower throttle amounts for the first several minutes after I hook up the battery. Then it is is smooth and runs well at max throttle. At first I thought the issue with KU93 was with the motor...but it was a timing issue..The ECrazys are cheap - you might try one.

Chris
 
Thanks for the note on the connectors.

I got Jack Xie to produce a ~260RPM wind of the Q128H motor, and when it is available, i will order one, and also a 201rpm Q100H. Should be available two weeks from now.

Will test both motors on my 9fet 3077 EB3 cell_man ( infineon type ) controller on 36 and 48v, maybe even higher, to see what they can do.

I'm pretty confident that the EB3 will tolerate these motors. If they can tolerate the hardest to control geared hub i know of at 560rpm with twice the pole count as these ( at least! ), i'm pretty confident that it can handle these.

EB3/infineon-type controllers are not cheap, but being so programmable and reliable, i wouldn't mess with anything else.. just my 2c.
 
The new KU93 are completely different from the old
ones (on the pics), for example the resistor behind the shunts now
is at the right side.
Also the circuit board is different.
On the bms page now they announce a learning mode!

And if anyone can explaind me correctly how to upload
pics from my computer, I will post old and new inside!
 
Statement from bmsbatt. on incompatibility KU93 - Q100H ..
:roll:

-------- Original Message --------
Subject: Re: KU93 controller - Q100H hub motor
Date: Fri, 11 Apr 2014 17:58:03 +0800
From: Caroline Yang <bmsbattery@gmail.com>

Hi,

KU93 is too big for Q100H motor .

于 2014-04-11 06:54, Al. wrote:
>
> Hello
>
> Can you please explain to me the more detailed technical reason(s) why
> a KU93 controller is not compatible (or usable) with a Q100H hub motor ?
>
> Thank you
>
--
Best regards

Sincerely yours,
Caroline Yang 杨月兰
Company Name: Ecity Power Co., Ltd
 
Hello Chris,
thanks for the advice on ecrazy controller. 15A max current limit would fit the bill. But as mentioned before the q100h having a significant higher reduction-ratio than the q100 so the question would be if that poses a problem again with the eCrazy controller :?:

And do you possibly know if the 48V controller can be used with 36V batt.pack also i.e. with lowered LVC to <30V ? Factory setting seems to be 41.5 Volts
48V 350W brushless controller for E - bike
http://www.ebay.de/itm/48V-350W-brushless-controller-for-E-bike-scooter-/260851912117?category=11332

Cheers

chvidgov.bc.ca said:
I had the very same problem with my ku93 controller on a Cute 100 201 front motor (12 amp limit with one shunt cut) - cutting out completely at over 95% of full throttle. I replaced the controller with a 48V 15Amp controller from ECrazyman off of EBAY and it seemed to solve the problem...although this new ECrazy controller has its own minor issue - surgy on lower throttle amounts for the first several minutes after I hook up the battery. Then it is is smooth and runs well at max throttle. At first I thought the issue with KU93 was with the motor...but it was a timing issue..The ECrazys are cheap - you might try one.

Chris
 
Back
Top