KONA Entourage Build Thread - middrive (pic heavy)

i have some serious stutter issues with this bike (mac motor and eb309 controller).
i suspected my CA to be thr problem but no think that other things may be the problem.
i had a broken fet (had one before) and replaced it with a 3077. only then i realized that all (but the one replaced before and the now replaced one) other fets are 3307. i don't know about those fets but i could imagine that this could cause the issues i see.
i also see a no load current of 7a (if my CA reads correctly - will check with a turnigy watt meter as well). i can't imagine that there is a hall/phase combination issue as the bike runs nicely at very low throttle and nicely on wot.
if you have any ideas what else could be wrong please let me know!
 
izeman said:
i have some serious stutter issues with this bike (mac motor and eb309 controller).
i suspected my CA to be thr problem but no think that other things may be the problem.
i had a broken fet (had one before) and replaced it with a 3077. only then i realized that all (but the one replaced before and the now replaced one) other fets are 3307. i don't know about those fets but i could imagine that this could cause the issues i see.
i also see a no load current of 7a (if my CA reads correctly - will check with a turnigy watt meter as well). i can't imagine that there is a hall/phase combination issue as the bike runs nicely at very low throttle and nicely on wot.
if you have any ideas what else could be wrong please let me know!

3307' only good for a max phase limit of 90A

if the no load of 7A is correct then this may well be your source of stuttering as the controller has a hardware non-programmable over phase trip that can cause these symptoms and the 7A can be due to a incorrect hall/phase wire combo, so hopefully the no load of 7A will be confirmed then you need to find the correct Hall/phase wire combo to put the no load in the ball park figure of 1-2A.
 
Nice welding mate! I was jsut thinking about powdercoating epoxy joints yesterday - useful info in this thread. Thanks for that!
 
thanks. i like it if information is shared.
and the result of powder coating is just incredible. a short notice: powder coating can cover little scratches quite well (if you can't feel them with your thumb's nail) but it also makes little bumps even bigger, so take care to have a smooth surface!
 
That's for sure - when it comes to coating. You'll see some similar frames at my links - I'm inspired to try a captive hubmotor build by this.
 
i thought i was an old pro now that i built several ebikes over the last years, but i must say you never stop to learn.
i thought that if you find a good hall/phase combination then you're done. but now i know that there are three (out of 36) and only one is the right one (the one with the lowest no load current).
unfortunately even testing all 36 combinations reveiled those 3 good ones, but no load current is quite high for all of them @ 7-8a. which may be caused by a bad bearing or a not perfectly alligned axle, but still very high.
and very unsatisfactory. the motor still stutters (not massively with Pgain of 10) but still not nice. wot is very smooth.
i have no idea what to do next. i could prepare another controller (solder the right connectors to try it with this bike) but after that my ideas are endless.
at least i sorted out the CA as cause for this issue by directly connecting a throttle to the controller. as it didn't solve my problem i now know how wonderful the CA handles dead throttle zones and power application (with 'power or amp controlled' throttle).
 
did you confirm that the CA was reading the correct current ?.


Edit:
And i would expect that the stuttering is defo to do with the high no-load current..

also I dread to say it, but you may even have a bad phase winding :?
 
yes. current was correct. i double checked with a turnigy wattmeter in line which showed even slightly higher amps (0.5a+).
how would i check for a bad phase winding? i wouldn't mind if the motor was not ok, as i have a spare one and i plan to convert a third mac from scratch.
if adopting the spare 12fet wouldn't be so much hassle i would have done it already.
 
izeman said:
yes. current was correct. i double checked with a turnigy wattmeter in line which showed even slightly higher amps (0.5a+).
how would i check for a bad phase winding? i wouldn't mind if the motor was not ok, as i have a spare one and i plan to convert a third mac from scratch.
if adopting the spare 12fet wouldn't be so much hassle i would have done it already.

you can measure the resistance of each phase using a constant current method or if you have a scope you can measure the output wave forms from each phase when spinning the motor up via another means, ideally you will need to put a load on each phase when you do this. All 3 waveforms should be identical and of the same magnitude.

edit: i think a motor fault is unlikely tho it may just be to do with your controller as you have have a few mishaps with it.. I would first try a another controller.

If you have another MAC motor cant you crudely connect that to your controller and measure its no load current
 
izeman said:
the no load current worries me most as it pure heat (350w) that stays inside the motor.
i don't have an oscilloscope. how would you measure the resistance other but with an ohmmeter?

because the resistance is so low a normal ohm meter will not be able to measure it so you will need to use a constant current method

http://www.youtube.com/watch?v=0vMvCVyOp9g

or you could buy a milliohm meter.

but try a different controller first or use you other motor to test controller
 
sure. that's easy. i could have figured that out myself. it's just the way the shunts in the controller work ;)
and i don't need to know the resistance of the winding. i just need to see if all of them are the same. correct?
man. so many things to do. i measured my others motor's no-load current and it was below 2a. but this was the motor only. not installed in the case, no planetary gears and stuff. but if i connect this motor to my bike's controller now and measure again and get 2a+ then i know it's the controller.
i have 3 spare 12fet controllers (sensorded and sensorless), a 6fet, but no more 9fet - only broken ones. :( damn.
 
i did install the additional 10ah today to make the pack the final 20ah. my range still is not very promising. to be realistic it's around 35km of fun. but not much more. that's not bad, but if you imagine an average speed of 25km/h then it's just a bit more than an hour. i was hoping for 2h at least :(

IMG_0122.JPG
and the bike is a real eyecatcher. EVERY biker i met wanted to know how fast, how much, how heavy, what distance, blabla ... and i only got very positive comments.
i rode a bit with a non electrified friend today and on our way back is started raining like in the amazonas rainforest. this is the view from top of the hill where the weather was still dry, but you can see a lot of rain in the distance already.


we were soaking wet within seconds. the trails became slipery, and our sight was minimal as the raindrops hit our eyes. and then the misery began: i did NOT make the bike really rainproof, as i wanted to go riding so badly. the first thing i noticed was a throttle signal w/o any throttle from my side. so i turned of the controller as it would have been a 100% crash if the throttle signal would have raised by itself on a slippery downhill trail.
so i had to pedal. pedal this heavy bike with only one virtual 58 teeth front gear (42 teeth with a 16/22 gearing to the rear cassette) is not what you want to do everyday. as is said it was our way back, so most of the road was downhill anyway, but even the shortest uphill parts where not easy to master. there were puddles 20cm (8in) deep and 3m wide, but we had to come home somehow.
as my motor's freewheel doesn't disengage fully it had the CA turn on by the "regen" voltage input of the motor several times. so it was not only the bike i had to move, i was powering the motor as well. yippie!!
when i was back home i realized that the controller was lit again w/o the bike moving. so there must have been a short (or power creeping) inside the controller and the fet-power-switch as well making the bike's controller turn on even with the main switch in the OFF position. luckily the CA still suffers from the "blank-screen-power-on" bug so even as it was powered it didn't book making the bike move.
it was too unsafe to leave the bike in this state, so i opened the battery box (i need a screwdriver to do it) and physically disconnected the bms's power from the battery. so there was no way the bike could start rolling away on it's own ;)
so what would i need to do to make the throttle waterproof? it's one of those "standard chinese" half twist hall type throttles. fill it with di-electric grease? vaseline maybe?
the controller itself needs to be made waterproof and this can be done, so i don't fear the "bike turn on automagic" anymore.
 
did you sort out the problem with the high no-current ? ... why dont you just use the 6 fet that you have :D

edit:
yes either fill with grease or paint the exposed hall leads and solder joints with nail varnish or paint.
 
no i didn't have the time to prepare a spare controller, but hopefully i can do it today. all connectors need to be changed and some wire extension has to be added.
what i noticed yesterday was that the motor's/gear's noise got less during the trip. it lost the high wining pitch sound, a bit nicer now and not so enervating.
i also ordered a CA replacement case from grin. being able to replace the case for ~$20 is a big plus.
 
Could you use a fly by wire system similar to a motorcycle? Throttle cables that link to an actuator inside the battery compartment ? I'm sure you could canabalise a fuel injected bike throttle from a breaker ? :D
 
Dunkenb said:
Could you use a fly by wire system similar to a motorcycle? Throttle cables that link to an actuator inside the battery compartment ? I'm sure you could canabalise a fuel injected bike throttle from a breaker ? :D

or if you did want to go down that route you could use one of these http://endless-sphere.com/forums/viewtopic.php?f=31&t=58801 :mrgreen:
 
those are nice gwhy!. i know that thread. but i'd like to keep my half-twist.
i disconnected all wires. disconnected the throttle, and something strange happened: the motor reved up. up to full speed. when i connected the throttle again it came down. i filled the connector with vaseline and it came to a full stop. but not always. it took some time (and drying out maybe). i filled all connectors with vaseline as they were all wet inside.
i tested my (not sure to be working) other 9fet and no. it was not working. i can program the controller but it didn't turn the wheel.
 
my next try will be with a 12fet. unfortunately those are all sensorless. so if the test is positive (lower no load current) i can't be sure if it is
a) the 12fet working better in general
b) the sensorless controller working, as the problem are the hall sensors.

the 9fet at least is totally up to the job. i programmed 45a battery and 120a phase, and even after a quite long full speed climb the controller was only hand warm. it's so strange that with very low speed everything is fine, with full speed everything is fine. it's just very hard and stuttering to move keep some low speed with low amps. it seems that there are current spikes that aren't filtered out (reduced) fast enough.
 
tried several controllers now. first one was a 9fet sensored, second one a 6fet sensored and now a 12fet unsensored. no movement off the wheel at all with any of those. that was not very surprising as o don't like to throw things away once the don't work anymore as i always think that i could repair/re-use them in the future. so i COULD be that all of them really are broken.
but maybe i just used a broken throttle connected to them (can't use the bike's throttle as this is connected to the CA).
so i may need to do some further investigation. unfortunately it's raining all the time and starting to get dark earlier now that autumn is at the door. it may take some time to find the culprit.
ps. at least i got all parts for my quadcopter build - so i can build something :)
 
hall sensors are very sensitive regarding moisture. all solder joints and the open locations at the sensors legs have to be isolated with some kind of rubber spray or similar stuff if you do not want to have an accident due to throttle malfunction in wet conditions. also the switches in the chinese throttle are normally not protected against water. I can tell you several stories about this and what can happen.. - don't ask why i know :)

as for the controller i have a suggestion: Kelly KBSX with optional waterproof (the case is mould with epoxy resin, so it should be real waterproof). These controllers work very well with the MAC Motors - even a bit above the rated ERPM. The programming features are far better than the few ones Infineon controllers do have. I like to work with them.
 
madin88 said:
as for the controller i have a suggestion: Kelly KBSX with optional waterproof (the case is mould with epoxy resin, so it should be real waterproof). These controllers work very well with the MAC Motors - even a bit above the rated ERPM. The programming features are far better than the few ones Infineon controllers do have. I like to work with them.
you got one to spare? i'd like to test them.
edit: afaik cyclone ships their kits with kelly controllers, don't they? and can i use them with my CA?
 
izeman said:
you got one to spare? i'd like to test them.
edit: afaik cyclone ships their kits with kelly controllers, don't they? and can i use them with my CA?

unfortunately i do not have a controller for spare..
yes there should be no problem in combination with the CA. you only have to build your own wire and an external shunt is necessary. the throttle type and working range can be set in the programming software. AND: no 3-speed switch (as far as i know only 50% and 100% so a 2-speed would be possible..)
 
Back
Top