KONA Entourage Build Thread - middrive (pic heavy)

the kelly KBS48121X looks quite small as well: 108x78. so it should fit easily in my 12fet infineon case which is 155x87. i don't want to change the case, as all mounting holes and cable housing is made for this controller (case).
i can't find a picture of a kelly controller opened, but i guess it should look more or less like the infineon/lyen ones, and have their mosfets on the long side standing up.
or should i look for a sine-wave controller? they should be able to reduce motor noise - but i'm not sure if what i'm hearing as high pitch noise is the motor or the planetary/sun gears.

gwhy! said:

...also I dread to say it, but you may even have a bad phase winding :?

Click to expand...

today i measured both of my motor's wiring resistance. attached a 1a load from my icharger and measured voltage drop through the windings. measure all 3 combinations and they are all the same. it was 35mohm for the 8t mac and 46mohm for the 10t.
i don't know what values to expect but as they are different between both motors but exactly the same compared to the phases of the same motor i guess i'm fine.
i revived the broken 9fet controller by replacing a broken 3077 and measured no load current of the installed motor. it was 7amps as well. so i removed the motor from the bike, and disassembled it. as the last time 2 of the planetary gears came loose i installed them upside down. this was a mistake. they were not supposed to be installed the way and the gears came off from the bearing. so i will have to install them in the correct position. the rubbed heavily on the rotor. this may have caused some resistance?!
the no load current with all gears etc removed was 2a. so this was fine.
i then wanted to check no load of my old motor (the 10t). connected any phase/hall combination, the motor stuttered a bit and no movement any more. i tested the halls and saw that there was a broken hall as well. this may have caused a fet to burn. normally this doesn't happen that easily, but maybe because the hall/phase combination was wrong even a few amps lead to fet destruction.
so i will see if i can find another 3077 spare and repair the controller. i also replaced all hall sensors and renewed the wiring. we'll see. keep my fingers crossed

izeman said:

the kelly KBS48121X looks quite small as well: 108x78. so it should fit easily in my 12fet infineon case which is 155x87. i don't want to change the case, as all mounting holes and cable housing is made for this controller (case).
i can't find a picture of a kelly controller opened, but i guess it should look more or less like the infineon/lyen ones, and have their mosfets on the long side standing up.
or should i look for a sine-wave controller? they should be able to reduce motor noise - but i'm not sure if what i'm hearing as high pitch noise is the motor or the planetary/sun gears.

Click to expand...

regarding length and width the KBSX should fit into the 12FET case, but the height will be the problem i guess. As far as i remember when i had the Kelly open the FETs are located on both sides. Big capacitors vertically in between. I could not see that much because it was filled up with epoxy. Maybe im able to tell you more next week about the size of the board and share a pic.

Sine wave controllers do need fine tuning to make them fit to the motor - otherwise efficiency will not be as good as it could be. Aside from the Adaptto Mini-E the Accelerated Systems BAC controller may be good for your needs, but they are very new and i do not know how they will work together with the MAC. For the Adaptto you will find pretty sure optimal settings.

i couldn't find a spare 3077 (only 4310 and 4410 en masse) so ic couldn't repair the second controller. but the hall replacement was a full success. i installed the "old" motor in the bike and no load current was down to 2.8a.
and the stuttering is gone. i don't know what's going on here, but it seems the controller and motor don't really go together. i have no idea why. this motor is a 10t now and the bike is a lot slower, so i can gear it up at the freewheel a bit.
quite happy now
EDIT: no it's not gone. still stutters. i will order some 3077 and see if the other controller handles better.

Just read all 10 pages, excellent build and clearly a lot of hours put in!

Well worth it.

yesterday weather was good so i had a 30km ride. average speed was about 35. heavy forest uphill and downhill riding. really steep hills. motor temp climbed up to 105c. which was absolutely fine and confirmed that the cooling concept does work great. even though it's not finished - the motor cover is not polished and has paint on it. and it's not completly in contact with the heat sink. so heat distribution may be enhanced greatly still. i bought a thin 200x200mm copper plate 0.8mm thick to act as heat bridge.
the motor noise is acceptable and noticeably became better.
the 9fet holds up to 2.2kw peak very fine. i set block time to zero which is fine as the bike is quite itchy one the throttle in low gears anyway.
i installed the 10t motor so top speed is quite low - around 50km/h. i will have to install a bigger gear in front as i want the top speed on flat at 60km/h+.
i still have the 42t rear cassette gear lying around but can't install it as the reallieur is too short.
so i think after installing some water cover plates under the ca to cover connectors and i'm done.
the bike is a real plug and play now.

i got a kelly KBS48101L for free and would like to test it. kelly webpage says it's capable of 35a cont and 100a peak. which is fine i guess as my bike does 45a max - but this 45a can last for several minutes. we'll see if the controller gets too hot.
another issue i may have is 40k electrical rpm. which may be too little for my 10t mac - it should be around 70k if i'm right.
e-rpm= rpm * pole pairs. rpm no-load speed = 470rpm *6 (as it's a mid drive conversion). mac has 16 pole pairs? so it's 470 * 6 * 16? or do i need to calculate with 32? because 16 pole pairs would equal to 37k e-rpm unloaded or 26k loaded?
kelly sells a "high-rpm" upgrade option. is this doable after market? or are they using different boards?
should i try it? it's quite some work to do. or will it fail anyway?

izeman said:

should i try it? it's quite some work to do. or will it fail anyway?

Click to expand...

470rpm no load speed at what voltage?

470rpm x 5 (gear ratio) x 16 pole pairs = 37600 ERPM

try it. it sould work :wink:

after many unsucessfull tries to install a amp/voltmeter in my 750w kingpan style charger i gave up. outside of the charger the meter worked fine, but one installed in the charger's case it's voltages readings start to fluctuate and is useless.
so i installed a turnigy wattmeter. it's harder to read from a distance, but it's very precise and it counts Ah put into the battery as well. the connector to the bike is selflocking and is rated for 16a. so no issues there charging with 14a.
the two pots can adjust the voltage from 3.85v to 4.20v and from 1a to 14a. so i can quickly set voltage to anything between storage and full charge. the batteries itself behave great so far. they all are 5mv apart so very balanced. internal resistance of the whole 12s pack is 2.2mohm as measured by the cylcle analyst.



View attachment 1

Very very nice Build, on of the best here

thanks

i was successful installing the kelly KBS48101L on my bike for a test ride. had some issues as it was always cutting out very early at half throttle and i didn't know why the last time i tried it, so i though: what a piece of junk and set it aside. today, as i was wondering what all those additional wires coming from the controller where good for, and all ppl talking about diagnostics leds i read the manual. now i know that my controller has leds as well - you just need to connect them - i thought the "better" controllers have them built into the case.
so with the leds installed my controller tolds me that a) my throttle was not connected and b) that my hall sensors where faulty/not connected. which both was correct, as the controller was only setup here on the desk and not installed to the bike so after programming it (all values maxed out) i was looking for the correct hall/phase combination. i found one but the bike was always cutting off power at half throttle. so i connected a turnigy wattmeter in line to see that no load current raised where quickly and jumped to 15a and then the controller shut off with blinking leds 4-2 (hall problem). so i tried some other combinations until i found the correct one. 3a no load speed (with the motor driving all gears and rear wheel). nice.
went for a short test ride, and now i know why ppl like them so much. a very smooth nice ride. very different from the infineon type one. especially with a mid drive and some drive train lag a current controlled torque throttle is MUCH MUCH nicer. there is no brutal kicking in like with my other controller. and it seems to be more quiet as well.
i will now have to find out how to set battery/motor current percentage correctly to achieve my wanted 45a/80a as i had it with the infineon. and then i will see if i can install the kelly internals into my 12fet infineon case so i can screw that to my bike.
wish me luck

too bad it's f**** cold outside. doing a quick high speed test ride at 60km/h with only a polo shirt and a vest is COOOOOLLLD!!!

i guess i have a winter's project now. i will install the little kelly (SO much smaller than a 12fet infineon) into the infineon's case so i make it a 1:1 replacement on my bike w/o drilling new holes, painting cases and stuff. and all wires will invisibly enter into the bike's battery box through the hole in the controller's box as it is now.


the kelly is really much smaller than the infineon

View attachment 4
those are the internals. a really stuffed controller with laquer on all components. really waterproof. well not anymore now that i opened it :?


fets are 48mm apart (outer side)


open width of the infineon is 74mm


so i will just need to make 2 bars 13x20x100mm (74-48)/2=13mm, screw the fets down on those bars and then screw those to the controller. this will add some thermal mass for short time load and the infineons case's surface is much bigger as well, so overheating should be no issue.

so it's taking shape. i found 12x20mm rod and used that. drilled a lot of holes and taped m3 threads.
all the pics are with only one or two screws installed to see if it fits, and it does. there is enough room for the inrush limiter on/off switch and the little dc/dc converter may fit in as well, and if not i will put it inside the frame box.
it took ages to remove all this silicone from the fets and wires. i like my controllers look clean inside as well :wink:

View attachment 2

note to myself: if you ever again service your brakes, remember to reinstall the safety bracket in the lever, or the axle will fall out and the break will not work
glad i had access to the lathe, and 15min later a had a new axle made. better than new, and made of stainless steel.

nice heatsinks (now) ;D

Merlin said:

nice heatsinks (now) ;D

Click to expand...

yes. they are just there to transfer the heat to the case. but surely will help buffer short heat spikes. unfortunately the controller is not working (anymore). i removed all unnecessary wires (braking wire, brake light supply, meter connection ....) to get rid of too much cluster.
i want to connect my CA-v3 so i installed a shunt (1.3mOhm) in the negative line after the inrush limiter. the controller ignition wire is connected to positive all the time. so if i flick the switch the inrush limiter turns on the controller and the CA at the very same time. this causes an error condition in the kelly as it sees no throttle input while the CA is booting. i need to delay the kelly startup by 2-3s. any idea how to do it?
but still after i turn the kelly on/off manually to clear the error (no red led blinking anymore) it doesn't turn the motor. no idea what is damaged.

shit....to bad ..iam not that eletric engineer you need to delay it.
maybe with a second switch between controller "start enginge" wire?

i don't want to have several switches. i don't even like that single one
i ordered one of these for $3.80. they need a 12v source and can switch 30v10a DC or 250v10a AC. so i hope it will so 50v and 1(?)A. i can feed it with the 12v coming from my front light dc/dc converter. hope it works. one more part that adds complexity and more wires

tried everything to make the kelly work again. no idea what's going wrong. measured between B+ and phases and B- and phases as stated on kelly's homepage. everything ok. green led turns on. red led turns on and turns off immediately. if i disconnect hall sensors it's putting out a failure signal. so the logic seems to work and the fets don't have any shorts either.
i directly connected a hall throttle to the logic board. but still no motor movement.
any ideas what i could do now? anything more is can test?

might be a silly question but it isn't a lvc or something simple?

No. LVC is set to minimum voltage possible. And the batteries are charged. And afaik the controller gives error code 3x1x blink if LVC is reached.

LVC (was) set to 18V already.

EDIT: this kelly controller is strange. i can program it, i get a GREEN ok led, and no blinking RED error signals. hall throttle goes nicely form 0.9v to 4.2v and the controller reports the "can't start motor within 2s error" if i turn the throttle. so i guess it's more working than not working. but there is ABSOLUTELY NO movement of the motor. what shall i do?

Did you write to the kelly support?

Merlin said:

Did you write to the kelly support?

Click to expand...

no. do you have any special contact data? or just the contact from their website?