wr250f conversion

[dougf]

Hi,

My first electric conversion. 2012 WR250F with the idea to maintain road rego. Thats still a work in progress. Parts used are;



Motor: QS180
Controller: APT 96800
Battery: 24s17p Molicel p42a ~6.3kWh
BMS: JK-B2A24S
Top speed: 100kph
0-100: 4-5sec
Weight: 136kg (heavy)
Range: 100km city and bush, usually between 65 and 75Wh/km without regen.


Ripped everything off and rewired from scratch.

Motor mount;





Controller mount, regret making it out of aluminium but it hasnt broken yet. it has tabs going into the top motor mount that are not visible in this photo;



The battery was the hardest bit and the most expensive. I made the mistake of engineering too much precision into everything but managed to make it work;









fiberglass vac bagged, didnt get many pics, sorry.



then filled, sanded (eh a bit lazy here but got the sharp bits off) and painted.

Things that went wrong. Had 3 main problems with the controller and motor. One, i needed to replace the encoder wiring with shielded wire because it was cutting the motor out. Along the same lines, I had the white connector on the back of the motor become intermittant after about 2hrs of riding offroad.



If you get this motor put some strain relief on this connector. Neutral cure silicone is good.

Last issue I had with the controller is the crimping job done by the supplier has proved to be poor. I had wires pull free of the crimps with very little force. Fortunately for me, I use the exact same connector for work and have the gear to recrimp properly.
The support from QS motors has been great. So long as your patient they will help you.

Other issues I had was the BMS unit failing after a few hrs. It uses 0603 0.1R resistors as fuses. 2A * 0.1R is 0.2V drop. 0.2V * 2A is 0.4W, way too much for 0603. It is possible that the resistor was a 0.01R, but a bit irrelevent given i had them fail. Anyway, replaced all the "fuses" with 0805, should be good. I also got the CAN version of this BMS, if anyone needs help with it let me know.



Otherwise really happy with the bike. Power is great, battery power peaks at around 31kW and tapers down until top speed. Gear ratio in my case is 11/60. Top speed 100kph. It makes a good dirtbike and a crap road bike. Ran a 16.61s 1/4 mile @ 100kph but Id maxed out before the 330ft . Going to do something about the shrouds and interested to see what people with more powerful controllers are able to squeeze out of these motors.


Doug

 

[SlowCo]

Very nice build! :thumb:
Get a set of SuperMoto wheels and tires with better gearing for the street. Or build a similar bike for the street so you don't have to switch wheels :wink:

 

[j bjork]

I am jealous of your machines :wink:
But I guess it is not from your hobby garage..

I am surprised by your numbers, 100kph top speed with 5.45:1 gearing?
I have higher top speed on the ktm with gearing's around 8:1.
That is with a 138 70h. Similar kv, a little lower voltage (22s battery) but a lot of field weakening.

But still pretty quick up to 100kph 4-5 seconds? I'm thinking the last few kph would be pretty slow to reach.

 

[dougf]

Very nice build! :thumb:
Get a set of SuperMoto wheels and tires with better gearing for the street. Or build a similar bike for the street so you don't have to switch wheels :wink:

Thanks mate. It turned out ok. Next one will be better. Some supermoto wheels would be fun, will definitely need to look into it with the current price of fuel.

I am jealous of your machines :wink:
But I guess it is not from your hobby garage..

I am surprised by your numbers, 100kph top speed with 5.45:1 gearing?
I have higher top speed on the ktm with gearing's around 8:1.
That is with a 138 70h. Similar kv, a little lower voltage (22s battery) but a lot of field weakening.

But still pretty quick up to 100kph 4-5 seconds? I'm thinking the last few kph would be pretty slow to reach.

The Mill is mine, the router is a work toy. Very useful bits of gear.

Its on the upper side near 5s, and thats with a full battery. Pulls like hell to about 60kph and gets a bit out of breath after that. My 60ft time was 2.019s which isnt terrible. I initially planned for a higher speed, more like 120kph. 60kv and 100V but I only get maybe 5200-5300rpm. I havent tested it yet but I believe I've actually gone faster before by gently ramping the throttle. Might have the same effect as field weakening. I have no field weakening turned on at the moment. QS motors are a bit careful about letting anyone program the drives, preferring to send what they call firmware updates out but which I believe are just configurations they have setup. I have some ability to change things but not everything. One day Il put it on the dyno and tweak. I think at that point though Id prefer a more feature rich drive. Right now it is more then enough for me.

Looking forward to see the results you get on your motor with the sevcon. I feel like this size motor should be capable of a bit more power short term. Riding enduro/ fire tracks it stays at ambient. Sitting at 80kph for 45min gets it to about 40deg. Havent tried on a MX track yet, if anything will get it hot it will be that.

Doug

 

[fechter]

Very nice work.

How did you attach the copper connections to the batteries?

 

[dougf]

Very nice work.

How did you attach the copper connections to the batteries?

I soldered them. I wanted to laser welding them but the welder I have has the wrong wavelength and just reflected off. I built 2 different spot welders to try do the job but ended up soldering because I was worried about accidentally blowing a hole in a cannister. The 0.3mm copper was too conductive. Thought about nickel plating it but at that point I had spent more time not making the battery then making it so I soldered them. Ive soldered many batterys for model planes in the past without issue. Different stakes when its 408 batterys vs 8. If I built it again now I would make PCB's with spring terminals and sandwich the cells together.

Doug

 

[fechter]

Everyone says that soldering to the cells is bad, but the few packs I've done that with are still working fine. It will be interesting to see the long term results.

 

[dougf]

Everyone says that soldering to the cells is bad, but the few packs I've done that with are still working fine. It will be interesting to see the long term results.

Yeah I did a fair bit of research on it and I believe lots of people have damaged cells by soldering. I weighed up my options and decided it was the best way for me to move the project forward. Ive soldered them in the past without issue. They remain in perfect balance during discharge and charge. Id certainly recommend a decent soldering iron. I used a 150w inductive iron with the biggest tip i had. But yes, If I have any problems with the pack I will update here.

An update on battery range. Ran the battery flat on sunday and managed 91.6km. My average speed was low at around 35kph and my total elevation change was about 1.2km. That is up and back down. I was taking it pretty easy to see what I could get. Total Ah out of the battery was ~68ah so pretty much bang on 6kWh used(only discharged to 2.8v per cell). ~65Wh/km which is the best Ive achieved off road. The small amount of regen I have must make a tiny difference.

Doug

 

[dougf]

Dont think I can recommend this BMS unit anymore. Had more fuses blow and when I went to plug it in for testing the connector exploded . I guess Ill have to make my own unless anyone has a recommendation. Only want the balancing and voltage monitoring, preferrably with canbus out.

Doug

 

[j bjork]

I am relatively happy with the ant bms, but it seems pretty easy to blow mosfets on it.
At least it happened to me at least twice on the 24s model I have on the ktm.
The problem dosent seem to be under use, but when turning on or connecting a charger.
The models I have has a button to turn on and off. There is another model that you need voltage to turn on, pretty annoying I think.
But a good anti theft I suppose.

I dont know if there are any canbuss you can use, but at least there are bluetooth for a phone and a display. (the display is not bluetooth)
The display has communication by 4 wires from the bms, there are a bunch of other unused connections that might have something useful

 

[dougf]

I am relatively happy with the ant bms, but it seems pretty easy to blow mosfets on it.
At least it happened to me at least twice on the 24s model I have on the ktm.
The problem dosent seem to be under use, but when turning on or connecting a charger.
The models I have has a button to turn on and off. There is another model that you need voltage to turn on, pretty annoying I think.
But a good anti theft I suppose.

I dont know if there are any canbuss you can use, but at least there are bluetooth for a phone and a display. (the display is not bluetooth)
The display has communication by 4 wires from the bms, there are a bunch of other unused connections that might have something useful

For now ive just run the wires out to a db37 and will check the levels with a multimeter. These cells are all new and stay in balance so it's not a big deal. Doesnt look like the ant units do canbus. Id prefer to have canbus, my charger is canbus controlled. I built a controller module to talk the this jk bms and control the charger. I was initially just going to make my own bms, maybe Ill revisit that. I believe some form of onboard BMS is required for road registration...dont know for sure. I'll speak to an engineer next week. Id prefer to just monitor during charging. The controller cutoff protects from over discharge.

The Orion BMS units look good but big and bulky, I wouldn't be able to sneak it onto the bike somewhere. Definitely not mounting another bms inside a battery pack. That was a mistake.

Doug

 

[dougf]

An update on the motor/controller combo. Did another 40km's today and pushed the bike pretty hard. Must admit it felt a bit gutless north of 60kph . Did a bit of following someone on a YZ250 and they would easily pull away from me on the faster trails. More then once i realised I was flat out. IMO not enough power to be a proper dirt bike. Will slap it on my dyno and have a go at tweaking the phase currents on this controller. That will need to wait a few weeks until I can dig the dyno out and set it up.

Also made some internal covers to protect the wiring.




Will update once I have a chance to do some tweaking on the dyno.

Doug

 

[j bjork]

It will be very interesting with dyno results :thumb:

The dyno itself is very interesting too, homemade?
I have some thoughts about a simple dyno from a car hub with brake and rim.
A load cell for the torque and use the abs sensor for speed.

 

[dougf]

It will be very interesting with dyno results :thumb:

The dyno itself is very interesting too, homemade?
I have some thoughts about a simple dyno from a car hub with brake and rim.
A load cell for the torque and use the abs sensor for speed.

Yes, homemade. Uses a eddy current brake and i had the drum made by a local engineering mob. Built the controller myself and wrote the dyno software in vb. Very V1 right now. As ever i fail to take pictures of things but found this one in my google drive. The plan is to drop it into the concrete pad when done and point the exhausts at a spray booth. Will be nice to dyno a bike that doesnt need fume extraction .



Doug

 

[dougf]

It was bugging me too much to wait till the dyno so I crunched some numbers. This is what QS told me the motor would do,



Something about the top speed was bugging me so I re-did the math based on gearing and j bjork is correct, the top speed is way low. Motor only spinning to about ~4500rpm MAX. That was significantly lower then my full charge voltage would suggest. It also shouldnt be a power limitation because the graph above leads me to believe the motor should make ~24kW all the way to redline but Im maxing out at 12kw Battery power @ 100kph....So I plotted the programmed phase current against rpm. The blue line is what I got,



The orange line is what I just set it to for a test ride. Hit 110kph with battery 82v . Thats more like it. Looks like there should be some good improvements to be made on the dyno. Will update after I unbury the dyno.

Doug

 

[j bjork]

Very interesting, do you know what voltage and amps the torque map are for?

I dont understand your phase amp settings.
In the map it looks like full torque up to about 1800rpm.
For that I suppose you need full (800?)pA up to 1800rpm, but you are already down to 570?
It looks like half torque at about 4000rpm, then I would think you should have 400pA not 100
And so on..

I am a noob at this, so dont take my comment as absolute truth.
But that is what seems logical to me.
I must take a look at my torque maps from the sevcon and see how they match

 

[dougf]

Not 100% sure about the phase current settings myself. Im guessing it's not as straight forward as amps == torque, RPM probably has a role and power factor probably does too. I dont have the completely unlocked version of the software to alter this drive freely. In my case I will make the dyno hold speed at the rpm points that relate to the phase amp points and turn up the current. Im guessing if pushed too hard the motor wont make extra torque, only heat. Im also guessing that sub 2k rpm this will be all she can give at 800pA but I should hopefully find a limit at the higher rpms.

I had a closer look at the BMS today, It's had a handful of fets fail short circuit. When I repaired the fuses the second time and plugged back in the connector arc'd where there was higher voltage. I manufacturer either doesnt know or doesnt care why it happened but I have a theory. Noticed the power on counter was way higher then the actual amount of times Ive powered it up. Looks like when I would turn the bike off, the bms would reboot over and over while the main caps in the controller drained down. Fets being incorrectly driven on during brown out could cause the damage. Will repair and use it as an external module.

Doug

 

[dougf]

Got the bike on my dyno today. I only did 3 pulls because I as mostly testing the dyno and I still had a knobby tyre on which was getting heavily shredded by the roller. I did one run with the setting's I posted last (not the factory ones) and I then set the phase amps in a direct corrolation to the torque map qs supplied me. As you can see, no difference on the dyno.


Brake kW on Left Y, Ph A on right Y, kph on X.

I was easing into the run at a bit of speed to prevent wheel slip and it was mostly the top end I cared about. What I did see was it basically pegged the battery at 300A for all runs meaning the battery current is the limit, not the phase amps. I believe this controller "might" go to 350A but really better off getting a better larger controller. I will need to fit an actual road tire and do some science. The sudden change in slope bothers me at around 90kph. I did just make some upgrades to the dyno's kph pickup, could be a problem with my code.

I did take the bike for a ride afterwards and it definitely has more top end since I upped the upper phase current limits. I hit 115kph indicated at one point which was basically all the motor could theoretically do with the battery down at ~85V. In suburbia its grins from ear to ear 8) .

Doug

 

[j bjork]

Nice to see some numbers :thumb:

Kw, I suppose it is on the dyno shaft?
So after losses in chain, tire to roller etc?

It will be interesting to see what happens on lower rpm, where the bA shouldn't be the limit.

Is there limits in the controller program that makes you unable to set higher bA?
I know I have seen discussions of different fw for APT, probably in the 96600 thread.

The far driver controllers looks interesting I think, I sure am curious to test the 1081800 :wink:
It should have MPTA, but no variable regen it seems.

 

[dougf]

Nice to see some numbers :thumb:

Kw, I suppose it is on the dyno shaft?
So after losses in chain, tire to roller etc?

It will be interesting to see what happens on lower rpm, where the bA shouldn't be the limit.

Is there limits in the controller program that makes you unable to set higher bA?
I know I have seen discussions of different fw for APT, probably in the 96600 thread.

The far driver controllers looks interesting I think, I sure am curious to test the 1081800 :wink:
It should have MPTA, but no variable regen it seems.

Yes, my dyno is a brake dyno with reaction cell. Read those numbers well salted, I believe i calibrated it by standing on the moment arm with the idea to calibrate it properly once I got it working. I'm also not taking into account the rotational inerta which will be using some of the power. I generally lock the speed and measure the power ouput at a constant speed. Thats most useful for tuning bikes. I will put the road tire back on at some point, that or buy a new rear wheel. A little part of me will die if I have to change the tires for 20min of dyno time. This would be bloody good fun as a supermoto though so the extra wheels might happen.

I cant see any obvious 300bA limit setting in the software I have. It might be set in the firmware. Like I said the drive is only rated for 350A I believe. Have looked thru the 96600 thread, some useful information in there.

Took it offroad again today with the increased phase current settings, feels pretty much the same...eh, maybe a bit more poke up top. I just wish it was 15kg's lighter and had 200km of range :lol: . I think with current battery tech, smaller pluggable batterys are the go. This is too heavy for hard enduro and not enough range to be a dual sport. Basically have to trailer it so having a big battery is just extra mass to lug around.

Doug

 

[dougf]

Did a top speed test last night, I hadnt tried since changing the phase current. GPS reported 120.25kph 8) . Im now maxing out the motor according to the voltage and kV which it wouldnt do before. I filmed the battery meter to look at the pack current and it matches the dyno, maxed out pretty much the whole time. Here's a little clip from when I turned around to go home.

https://www.youtube.com/shorts/K58__UWJKbU

I will still put some dyno work in. I believe the PC software for this controller can tell me the actual phase current, Il use that to set the curve to something reasonable.

Doug

 

[dougf]

More bad news with this apt 96800 controller. Connected to the drive today to tweak the throttle curve and noticed this,

Looks like after ref 6 the points are 8 bit max . This might explain some of the random oscilations just before the steady decrease of power on the dyno plot. I could perhaps get around this by changing the rpm range but that will throw off all the other rhings using that rpm.

I also changed to a magura throttle. It works ok but the apt reading is a bit noisy leading to some surgy behaviour while steady state.

Really not impressed with this controller. If anyone has recommendations...i enquired with 2 local vendors on a sevcon, one didnt reply and the other didnt wanna sell me one. Are the far drivers any better then these APT's? Or the kellys? Any thoughts would be appreciated.

Doug

 

[j bjork]

Get a sevcon, then you can teach me how to tune it :thumb:

 

[dougf]

Get a sevcon, then you can teach me how to tune it :thumb:

Ha! Seems like the best choice for quality and functionality. Its also lighter....il stick with this apt for now. It does work. I might need to get rid of it to pass engineering though as it can spin backwards when first powered up. Gonna push to get this regoed, its by far my favorite bike to ride right now. The low end torque is addicting 8) . Makes me wanna build a bigger road bike, or a drag bike.

Doug

 

[dougf]

Just to correct myself, I did some more digging into the settings on this controller and found this,



That's different to how I had worked out the RPM steps, so i replotted it and changed the iqref points to what they actually were (255) and this is how the dyno plot vs phase amp limit looks now,



That explains the shape. Granted, preceeding that the controller is basically maxed out on battery current anyway. I could up the max rpm to 10000 to gain full range of phase amps over the rpm range i actually use. Hopefully should help others trying to use these controllers.

Doug