Low power bike with Turnigy 80100 motor

[J0nathan]

My mum recently asked me to make an e-bike for her to use, just for short trips. A long time ago I bought one of the large turnigy 130kV motors, intending to use it on my own bike however I never got round to it. More recently someone gave me a bike conversion kit with a broken controller. Unfortunately the controller is built into the BLDC motor and potted with epoxy, so I've left that for now. The kit comprised of this motor connected to a planetary gearbox and chain, which attaches to the rear wheel, so I just made a new shaft for the turnigy motor and mounted it to the gearbox by making an appropriate aluminium adapter plate. As you can see in the images below, I machined some aluminium bars and another plate to help protect the motor in case the bike falls over. It's powered by two 5Ah Zippy 20C 6s batteries in parallel to give 22.2V and 10Ah.






It was all going well the few times I tested it - 20.1Wh/mile at an average speed of 12.7 mph and top speed 17.6 mph. Slow by most people's standards here, but plenty fast enough for my mum! The first time she tried it without me, the ESC went pop. I'm not entirely sure what caused it. Apparently a capacitor touched the bike frame and made a reasonably loud noise, so the most plausible explanations is that it short circuited with the motor via the ESC, however none of the motor wires appeared to be touching the bike frame so I'm not sure. The meter registered a peak of 64A, which is less than I've seen in normal operation. It's quite annoying since although it was a cheap ESC from a long time ago, it did seem quite robust as it uses TO-220 mosfets, not the usual masses of silly little things. I tried replacing the pair of mosfets that had clearly blown, but it didn't make a difference - the motor just twitches and draws quite a high current so I think one of the gate drivers is probably at fault.

It seems sensible to replace this with a sensored controller and from reading around on this forum it looks like the following parts would work well:
http://www.ebay.com/itm/48V-600W-brushless-controller-for-E-bike-scooter/300352534034
http://uk.farnell.com/jsp/displayProduct.jsp?sku=3111465&CMP=KNC-GUK-FUK-GEN-SKU-MDC&mckv=s8W87YpR5|pcrid|19854174608|kword|ss411a|match|p|plid|

I'll happily replace the mosfets with IRFB4110 (or whatever), if required. As you can see in the pictures I have separated the wires in the motor, so I can easily wire it in star or delta, so I think the best option would be to change it from delta (current setup), to star, then put the batteries in series instead of parallel. That should raise the top speed by about 15%, but more importantly should make the system more efficient since the currents will be lower. Can anyone tell me if this plan is reasonable and if the controller I linked to is a sensible choice?

 

[gwhy!1]

as your motor is only running on 22v I would suggest trying a e-bike sensorless controller, the ones I have played wiyth in the past can normally run upto around 4500rpm the e-bike controllers are normally quite cheap and upgradeable , the sensored controllers are good but you will have to retro fit hall sensors so a sensorless controller will make things easy ( and they work very well with these motors ) , I bought a e-bike sensorless 12fet from ebay and it was only 16 quid delivered and without upgrading teh fets I was running it at 80A ( soldered shunt ) the standard current of the controller was 30A... just a thought

edit:

I have been trying to find a link to the actual controllers that i have played with with no joy ( sorry ) of course you will have to hack the lv cut off to operate the controller on 24v as most of the 12fet controllers are 48v

 

[J0nathan]

Thank you for the reply, that sounds promising. I'm not at all concerned by having to add sensors to the motors, since from various posts on this forum it doesn't appear to be at all difficult. If I double the voltage by putting the batteries in series, as mentioned in the original post, then I guess I'd have to use a sensored controller as it seems the consesus here that the only sensorless controller that works with the higher voltages is the CC HV160, which I certainly cannot afford.

Do you have a link to the controller you bought?

 

[gwhy!1]

sorry i posted a edit in my last post ... i can tfind the controlls i have played with, yes soon as the motor speed gets up around 4500rpm then you will need sensors fitted to the motor.

where in the world are you ?

Edit: just re-read your OP so if you run in wye (start) this should give you around a kv of 76'ish so you may just about get away with using a e-bike sensorless controller at 48v as this will be around 3800rpm max, i think it will be well worth a try.

 

[Thud]

Hello & welcome to the forum.
love the handy work on the motor mounting & the support bushing on the output shaft....very nicely done.

Going to a sensored controller that will limit power out put is a step in the right direction.
sounds like an old school rc controller for the car guys with TO220's in it.

now for my input:
I fear you are woefully over geared, even for 24volts.
from here I am guessing 26"wheels. (?)
what is your current gearing?

assuming 26" wheels,
you will want to gear for a wheel rpm of 200 ish for 15mph. @24v nom. it will require a 14-1 ratio to get there.
I suspect you are seeing the lower/target speeds because the motor overloaded on the lower voltage.
I have a bit of experiance with these & 24 volts should work fine for this application.

my guess is the controller blew while she was riding at a less than full throttle setting....which is a worst case senario for the controller.
especialy an rc style unit.

if its a 24" wheel try 12-1 ratio...that will fall into a 15mph range give or take.

hope that helps.

 

[gwhy!1]

i dont think he metioned the total reduction ? he also has a planatry gear reduction or did i read the OP wrong .

 

[Thud]

I guess a small planitary could be hiding in there...(I hope so :wink: ) but if its a 12t to a 48 or 60t.....he is overgeared.
EDIT:
duerrrrr.....I just found it in the OP! (& i double read the whole thing.....or so i thought)
sorry for any confusion....I stand corrected....but look it over anyway.

 

[J0nathan]

I don't remember what the ratio of the gearbox is, but we can work it out given that I measured the max speed as 17.6mph (7.87 m/s). Wheel diameter is indeed 26", so 1.92m circumference, 130*23.5/(7.87/1.92*60)=12.4, so the ratio is about 12.4 :1. I'm confident that the controller didn't fail due to overheating since my mum said the motor didn't seem to work from the start when she used it, and the capacitor shorting on the frame soon after connection would explain that. When I tested the bike the mean power draw was just 257W, so around 11 amps and unsurprisingly given that low current, I didn't notice any rise in temperature from the controller or motor. There were some high peaks, but still well within the ratings of the mosfets.

It would be nice to be able to use the bike on 48V with the motor wired in delta, so if I want to use it I can go faster...that may rule out the sensor-less controller? I did buy a cheap turnigy redbrick ESC from Hobbyking and make a PCB to take the control board from that ESC and drive 6 TO247 mosfets. This should be fine for 48V, however unfortunately the ESC never arrived .

Added location to my profile.

 

[gwhy!1]

I don't remember what the ratio of the gearbox is, but we can work it out given that I measured the max speed as 17.6mph (7.87 m/s). Wheel diameter is indeed 26", so 1.92m circumference, 130*23.5/(7.87/1.92*60)=12.4, so the ratio is about 12.4 :1. I'm confident that the controller didn't fail due to overheating since my mum said the motor didn't seem to work from the start when she used it, and the capacitor shorting on the frame soon after connection would explain that. When I tested the bike the mean power draw was just 257W, so around 11 amps and unsurprisingly given that low current, I didn't notice any rise in temperature from the controller or motor. There were some high peaks, but still well within the ratings of the mosfets.

It would be nice to be able to use the bike on 48V with the motor wired in delta, so if I want to use it I can go faster...that may rule out the sensor-less controller? I did buy a cheap turnigy redbrick ESC from Hobbyking and make a PCB to take the control board from that ESC and drive 6 TO247 mosfets. This should be fine for 48V, however unfortunately the ESC never arrived .

Added location to my profile.

The other problem if you want to run on 48v in delta for a higher road speed would be you will need more power and to be honest your tiny gearbox may not be able to hold up to the increased speed and power, but if you did retro fit sensors this will open all options for you try , if you fit the sensor internally in the slots of the motor then you will not be able to run the motor in wye reliabley so I would advise externally fitting sensors so the sensors can be shifted enough so you can run the motor in either configuration , if you go for a e-bike controller this will make things so much more reliable than using a rc controller as e-bike controllers have far more over current protection and more protected against the weather. If you get a sensoed e-bike controller then the controller you linked to originally is the best option as it is well documented for moddification.

 

[wildharemtbkr]

I believe the gearbox/reduction unit is a currie, it uses a matex planetary reduction, 4.75:1 or 4.875:1, I can't quite remember, then I believe it has a 19 tooth #25 sprocket, with a one way bearing to a 54 (I think 48 was also an option) tooth on the wheel. Mine was 13.6:1 reduction. I ran one on my recumbent, with a brushed scooter motor at 48v, and blew the one way bearing within hours. Charged that a couple times, before I gave up. I currently have it on my norco with no one way, but haven't finished it.

 

[J0nathan]

Yes, I think it was made by currie.

Thanks for the confirmation that that's the right controller. There are similar looking ones from other eBay sellers, so is the one I linked to the cheapest one and do you normally get customs charges on them? I don't mind paying the VAT, it's the £10.50 fee from FedEx that's irritating. If so I'll buy it and make a mount for the sensors outside as suggested.

 

[gwhy!1]

Yes, I think it was made by currie.

Thanks for the confirmation that that's the right controller. There are similar looking ones from other eBay sellers, so is the one I linked to the cheapest one and do you normally get customs charges on them? I don't mind paying the VAT, it's the £10.50 fee from FedEx that's irritating. If so I'll buy it and make a mount for the sensors outside as suggested.

I have never paid customs charges for the controllers from e-crazyman, but as regards the cheapest supplier I have never found 'these' controllers cheaper, the controllers them selfs are very reasonable but the postage to ramp the price up

 

[J0nathan]

I keep forgetting to update this...

The controller suggested from eBay worked well - I did change the motor back to delta termination and the performance was quite reasonable. My sister was using the bike to commute to work, however quite recently the controller seems to have died. When I switch the controller on the motor emits 4 very quiet 'clicks' in quick succession (takes about 1s), then nothing happens, no matter where I move the throttle to. The motor doesn't twitch/move or anything. The battery is of course charged and above the cut-off voltage.

I've checked the following things:
- Hall sensors still get correct voltage and work
- Voltage signal from throttle is correct
- MOSFETs seem fine (or at least not shorted)
- Regulated voltages on controller seem normal

Has anyone seen this problem before? It's a bit tricky to diagnose when nothing's happening, so if it's not fixable I'll adapt the servo motor controller I recently designed and use that.

 

[J0nathan]

I forgot the mildly interesting bit - here's how I mounted the sensors on the motor:

Simple PCB, with slots to move it to adjust the timing...

But still no progress with getting the controller to work again

 

[gwhy!1]

I forgot the mildly interesting bit - here's how I mounted the sensors on the motor:

Simple PCB, with slots to move it to adjust the timing...

But still no progress with getting the controller to work again

I like the hall board 8) ,

what controller did you get in the end ( link ? ) . I have had some controllers that "self learn" the halls that just dont work , and I never did get to the bottom of why this was.

edit:
just re-read your OP and you have the e-crazyman controller, there should be no problem with this controller.. what spacing have you got for you sensors ? 30degrees ? if so did you flip the center hall ?

edit: just read that it was working then stop I should read all the stuff instead of try to skip over stuff.. ok it sounds like you have a problem with your controllers output and my guess is you have blown at least 1 fet. Did you tune the sensor position before you started to use it ?

 

[J0nathan]

I got the controller I linked to in the first post - the 30A 48V ecrazyman one. It worked for about 10-20 charges before suddenly not responding...

Edit: Sensors are 60 degree spaced, middle one flipped.

 

[gwhy!1]

if tuned correctly you should have been getting around 6 - 7A noload current from the motor at 48v WOT if in Delta.

 

[J0nathan]

Yes, the first time I set it up I used my lab PSU and adjusted the sensor position to get the lowest current. At that point it was in star - about 2.5A:



I do have a video of that test with oscilloscope attached, but it's pretty generic so probably not worth posting.

Just remembered, I vaguely remember accidentally disconnecting the power connector before disconnecting the wire that switches the controller on/off. Maybe that caused it? Is there a schematic for this controller around anywhere?

 

[gwhy!1]

Yes there is a circit somewhere have a look in the 6fet cutting out thread, if you cant find it post back and I will post it up again when im back at my pc. Did you retune when you swt hed back to delta, as the timing is slightly different.

 

[J0nathan]

I just had another look at the controller...had my hands full with another project all week, sadly not my equipment:



Anyway, the first thing I spotted was that one pair of MOSFETs (phase B to DC-) were short circuited. Not sure how I didn't spot that the first time! I de-soldered them and checked them with a meter - one seemed fine so I put that one back, just on the back of the PCB temporarily. I connected the controller to the current limited PSU again, just to be safe, and it worked fine. The existing MOSFETs are STB75NF75 and by a strange coincidence I had a tube of five STP75NS04Z MOSFETs that I aquired 7 years ago for no apparent reason, so I removed the single one I'd just put on and soldered in two new ones. That didn't work - no shorts and correct signals from throttle and halls, +15v and +5v regulator outputs correct, but no current applied to the motor. I tried putting the fet that worked in parallel with the two new ones (long shot I know) and that made no difference so I removed it, then promptly accidentally shorted a gate pin to the source which broke more fets... which I replaced, but I switched it on and they just immediately shorted again, although a pulse of current was evidently applied to the motor as it did make a noise.

I think it's time to give up with this controller and made the one I made work:



For this controller I used the STGIPS30C60 IGBT module as it's for my CNC router so using mains voltage is convenient, but the current rating should be just enough with a decent size heatsink. I programmed it to do vector control with an encoder, so I'll need to do a little bit of coding to get the hall sensors to work, but that's all. It's tempting to rewind the motor and just run it at a higher voltage to make the system more efficient, but that means buying more batteries.