boleee controllers? (18fet model 18_02)

anyone using an 18fet boleee controller?

bole18_2 , website has no mention, only the 18_0, guess mines a revision 2

Just had mine delivered, bought off ebay pre modded with a 48/60/72v switch, first thing i did was open it to photo all the insides for one day when something will happen, first thing to happen was a wire from the swicth came off the board bringing R25 with it....

now the wire is completely across r25 making it 0 ohms/shorted, so i dont know if thats right or bad fitting, also the pads on the pcb are off, as they were like 1mm square, now guessing were the connections go...

it looks like one side goes to ground, so with the wire as is would ground the other side, not prepared to do that yet, just incase this is a bad mod.

also controller has very little in the way of sensible fitting instructions, and theres no sign of an led on the pcb to flash codes, can see fitting this today being a mare....

http://www.boleee.cn/ProductList.asp?SortID=135&SortPath=0,135,

Ok I have made some progress, in fact the last 5 hours or so has been taking pictures, googling, tracing wires, and to an extent winning...

It seems the single medium red wire is indeed for 'power up' the controller by connecting it to +pack_volts

Upon connection I was greeted by a flashing led (which I had to fit myself as they didnt) - 8 flashes, which on my other controller is low volts, so guess the firmware/software is going to be similar. (no doubt may also do the 7 flashes for blown mosfets...)


I have also reconnected the wire that fell off to one side of r16, the other side goes to the cpu, nothing seems to have gone puft so i'm guessing i traced it back ok

Now i will try to understand the rest of them and post back.

what i cannot understand is why there is a three voltage selector switch. did you have that added as a custom feature and why?

i see the input power resistors are all the same so there doesn't appear to be anything that the switch does to compensate for the high voltage on the 72V setting. are the mosfets able to handle the 88V?

The switch is already fitted by the ebay seller, I didnt need it, but his price was lower than others without it, I just bought the controller form the cheapest, switch was a bonus.

It changes between 3 different resistors to +pack volts to those 3 big ones, it also changes 3 other resistors to the LVC circuit.

the additional resistors are heatshrinked near the back of the switch.

I will take the heatshrink off and draw out the circuit.

the mosfets are 4410z which i think are 95A-100V, the caps are 100v

Have a lot of the wiring worked out now, I wished these company's could get together and agree on the colours of the phase wires, it runs backwards from my other controller so now i have to rewire the thing. (green/yellow inverted)

I'm afraid to say that with those rather nasty FETs fitted this cheap 18 FET controller is likely to be significantly less capable than a decent 12 FET controller running IRFB4110's, so I'd not expect too much from it.

The IRFB4410 FETs are pretty poor, with an on resistance of 8 to 10 mohms, so even with three FETs in parallel in each bank they are still going to be around 33% worse in terms of resistive losses than a 12 FET with IRFB4110 FETs. Add in that the gate drive is often poorer on 18 FET controllers, due to them often sharing the same driver across all three paralleled FETs, and the switching losses may well be worse, too.

The chances are that even a good 6 FET controller fitted with decent 100V caps and IRFB4110s would match this cheap 18 FET for performance.

Ouch, not the kind of thing I'd want to hear!!

And yes the quoted for the fet is 9mohm, more m'ohms, more loss, more heat, more chance of failure, I dont think the voltage drop is gonna cause a great loss of speed, and as for a loss of instant power, I wont know yet...

At least unlike some other controllers the gates do have a resistor each.

But if the weakspot is fets, its fets, they can be swapped out for better ones later, I am planning on using this as a diy super controller, just it was easier to buy one to mod than to start from scratch writing software etc

planning on simply adding a 2nd row of fets 'upside down' and watercooling it and running 84v on a mcycle.

still dont have a hub motor for it yet......

flez1966 said:

Ouch, not the kind of thing I'd want to hear!!

And yes the quoted for the fet is 9mohm, more m'ohms, more loss, more heat, more chance of failure, I dont think the voltage drop is gonna cause a great loss of speed, and as for a loss of instant power, I wont know yet...

At least unlike some other controllers the gates do have a resistor each.

But if the weakspot is fets, its fets, they can be swapped out for better ones later, I am planning on using this as a diy super controller, just it was easier to buy one to mod than to start from scratch writing software etc

planning on simply adding a 2nd row of fets 'upside down' and watercooling it and running 84v on a mcycle.

still dont have a hub motor for it yet......

Click to expand...

Swapping the 4410's for 4110's will reduce the resistive losses by around 2/3rds, so worth doing. If there are individual gate resistors for each FET than that might mean that the gate drive is better than on some 18 FET controllers, provided the resistors are low enough in value to get reasonable switching times. The Xiechang 18 FET controller is particularly poor, as it uses a single 91 ohm resistor feeding three paralleled FET gates, so all three are sharing a measly 130 mA of drive current, making for very slow turn on times. Hopefully the gate drivers on this controller may be a bit better, but I would suggest that adding more FETs will be a recipe for failure. The 36 FET Xiechangs (and the 24 FET ones) double up on the gate drivers as even the cost-cutting Chinese realised that turn on times were far too sluggish with just a single driver trying to drive 6 gates.

My DIY controller uses 10 ohm gate resistors for each FET (so around 1.2 A of gate drive per FET) and is pretty much OK to 100 A or so with just 6 FETs, so I'd suggest you probably need to look at finding a way to get more gate drive if you want to parallel up more FETs. TBH, you'd probably have been better off buying a bare board 36 FET Xiechang as I doubt it'd have been more expensive. That would have given you a clean board, ready to add the FETs of your choice. Keywin Ge can usually supply bare boards if asked, certainly I've bought a few from him over the years, as have others on here.

I dont know what they charge for boards but this cost me $58 shipped to my door, I see where your going with the gate drive needing to be kept high enough to get good switching, I dont think I can get much out of whas in, I could double up on the final drive stages mind.

its all a bit hypothetical at the minute, by the time I get my hands on a motor that can take 3kw+ i may have replaced this with something else.....

At the moment i have other issues, we have now had 48hrs of rain, the bikes got water into everything....

also the new controller wont run on bike on road, fine on bench....

thought it was a bad connection, still not sure...

on a very soft-medium throttle cuts out after a few random seconds 2-8, on full throttle from a standstill cuts out after about 8-10s everytime but its pulling away well and is faster than my 12fet

the current shunt i have on the bike drops 76mV on the old controller, 122mV on the new one under hard acceleration

just a 5mph trundle causes a cutout and led flash, but lift wheel off ground, and it will run on forever, and pulling the brake on the wheel it will fight and keep going till you smell them heating up...

I thought it was batteries so recharged, still same.

everytime it throws an error of 7 flashes, but with no documentation could be anything, low batt is 8 flashes btw.

changed back to old controller a few miles from home, bust a bullet of the phase wire plugging it in, had to jam it in.

then inside throttle got wet so controller threw error everytime i turned it on, boring pedal home....

Got home, dried wiring, then it would not respond at all to throttle, seems wiring has been getting trapped under gear selector and it decided now was the time to go just to confuse me...

Everything now back to normal on old controller, but now dark and /still/ raining, not going back out just yet.... new controller can wait a few more days till summer is over.

An email to Keywin Ge would get you a price for a 24 or 36 FET Xiechang bare board (all components except the FETs and case). My guess is that the cost might be around the same, maybe less, than the controller you have.

The big cost on these controllers is the FETs, 36 off 4110's will set you back over £100, so the cost of the board tends to get lost in the noise a bit.

Ok now the bad news lol


Controller was still running fine on the bench and tripping out when ridden, I thought ''it must be voltage sag'' and that as i had to turn round to see the led flashing that i was mis-counting.

I extended the led to the handle bars, I saw 3 flashes, 4 flashes, 7 flashes, no flashes, and multiple variations...

as i kept riding after each trip out it seemed to be losing power, I decided I must have a weak batter in my pack of 5 sla's and decided to keep going to get to the bottom of it, then it stopped..... 7 flashes...

when i tried to set off, nothing from the throttle, and 7 flashes.,....

when i tried to pedal, huge cogging from the motor......

one phase has gone short to +pack .....

maybe its had a weak or non working fet from the start, stressing the other 2 in the triple, until both of those have gone under stress...

/hope this controller does not have a bad batch of fets....

"/hope this controller does not have a bad batch of fets...."

for $35? no way! it has to be real IRFB4110s!

plus you got the 3 way switch so you can change the battery voltage while on the fly. going too slow? then switch to 72V!

maybe you can mount the led on your helmet so you can watch them flash without having to look down.

at least you know the bike won't fall over with the 70lbs of SLA to hold it up.

jeremy is really trying to help you.

dnmun said:

"/hope this controller does not have a bad batch of fets...."

for $35? no way! it has to be real IRFB4110s!

at least you know the bike won't fall over with the 70lbs of SLA to hold it up.

jeremy is really trying to help you.

Click to expand...

Your sarcasm is weak, you need practice...

the controller is not built with fake 4110's, nor is it built with genuine 4110's, its built with 4410's, if these are genuine or fake is not yet questioned, and to be honest if fake I dont really care, as long as they WORK !

Jeremy has made some very inciteful posts and I agree with all of them, thats not disputed, albeit for one thing, the buying of a bare pcb/cpu, its a good idea but the cost difference is minimal and (a) I did not know I could and (b) I had already bought the controller I have here, the only part of it i wanted was the cpu, but by buying a prebuilt one i get to try it before stripping it down, I know it works before i start my work, if I bought a bare pcb/cpu and built on that and it didnt work, i would be left puzzling if the fault was with the pcb/cpu or what i had done to it

Some people can afford to buy lipo, some people can afford to start out only with sla, feel free to mock what I can afford all you want, I will move to lipo as soon as I can, and as soon as payment arrives for some work done some time ago drbass will be getting it sent on to him, as I do believe in buying from others were possible instead of just ebay

Looks like this controller is probalby the 18FET version of this 15FET:
http://www.endless-sphere.com/forums/viewtopic.php?f=4&t=44226
VERy similar layout and component choices, same case, etc.

I think most of these cheap controllers are the same with just little modifications to suit,

The one good side of that is what fixes one usually fixes another give or take