BMC 50A controller & 1000W V3 motor

[ebikes-sf]

Testing setup:
xtracycle, BMC V3 1KW motor, BMC 50A controller, 50V LiFePO4 pack, stand alone cycle analyst.

Unfortunately testing lasted only a couple of minutes. I started with going up about 12-15% grade hill at full throttle. Throttle response was smooth. Current draw was about 44A on average (2000+ watts, double of what the motor is rated for) and speed around 10-12mph (no pedaling). Suddenly the motor quit running and I noticed melted insulation on motor wires. Controller was reasonably warm and seemed to be ok.

Controller has two limiting switches (details below) which I didn't get a chance to test. Up on connecting controller to battery, it arcs quite a bit and can melt corners of battery to controller connectors.

Will do more testing as soon as I fix the motor.




I have BMC motors, controllers and kits for sale, more details here:
http://endless-sphere.com/forums/viewtopic.php?f=9&t=11024

[TPA]

This is the same ol' same ol' for BMC motors.

The phase wires going in to the motor can't handle the current load. If I'm remembering right it was Oatnet and Mau' dib that did the testing on these and were stumped at how to get more current in.

It seems that a couple of the controllers will work with the 1000W motor, but the phase wires will not allow all the performance it is capable of.

I'm tempted to buy one of these, take it to the machine shop where I work and see if I can get bigger phase wires to fit by whatever means necessary.

[ebikes-sf]

My primiary focus of testing is on ablilty to climb hills (I live in San Francisco). Based on my experience with V2's, hall sensor tend to get damaged when the motor is overheated, this is more common for V2 torque models since they take more current at low motor speed than speed V2S, and they're usually operated at more power than rated 600 Watts. It seems construction of V3 motor is not much different from V2, so V3 isn't able to disperse heat generated from coils any better than V2.

Im my opinion, 1KW V3 motor should not run at much more than 1KW for reliable service independant of terrain or outside temepratures. Running it at 2500W (48V x 50A), even for short periods of time, is really fun but very abusive for the motor. Making phase wires thicker will only expose other bottlenecks like burned coils, blown hall sensors, burning lubricant, etc...

[dogman]

The motors need a thermometer sensor inside, so you know when to stop melting things. All motors overheat under some conditions, but for gearmotors, feeling the covers by hand is not good enough. I think having the torque to climb the hill leads riders to overheat faster than they would if they had less torque and speed up the hill. Having the phase wires melt before the motor melts is not a great solution to the heat problem. A warning buzzer, followed by a shut off if you continue would be better.

I'm amazed that companies increasing wattage continue to ignore the issue of motor heating.

[ebikes-sf]

Yes, thermistor would be great, only one small extra wire is needed. Heinzmann geared hubs are equipped with thermistors. I spoke to Chandu from BMC and mentioned about installing thermistor and supporting direct plugin Cycle Analyst, he liked the ideas but said if BMC decides to do that, it may take a while to implement. Chandu also mentioned that since inside shaft diameter doesn't allow thicker phase wires to run through, they're working on ways to enlarge the wires, probably in version 4. He also recommended cutting stock phase wires as close to the shaft as possible and replaing them with 10ga wires (which I'm going to do).

I opened up BMC 50A controller and hooked up direct plug-in CA to its shunt and wired the throttle override, it was actually relatively easy - I will share details in my next post. So now I'm able to limit the current using CA. Still have to get V3 motor fixed though, if damage is beyond melted insulation on outside and requires opening up a motor, I will try to install a thermistor and wire it to cutoff throttle signal when temperature goes over the limit.

[dogman]

Put in the thremistor, or a sensor from a remote thermometer. If you can actually watch the rate of heat buildiup in the motor, you will learn a lot more about what wattages, what loads, what ambient temps, do to the heating of the motor. Riding with a temp readout taught me a lot about how to ride cooler.

If nothing else, put the sensor on the axle stub, and add about 40F to the readout there. Thermometers for indoor outdoor are 10 bucks in the car section of wallmart. Look next to the compases and other car junk. You can read the hub exterior temps with a infared, but the 10 buck thermometer lets you do it on the fly.

[jag]

It seems construction of V3 motor is not much different from V2, so V3 isn't able to disperse heat generated from coils any better than V2.

The K and Ri differences between the V2 Speed and Torque are about 1.4:
1.28/.89 = 1.4382
.099/.07 = 1.4143
Hence a V2 Speed at 48V and 30A max set on the controller is performing pretty much the same as a V2 Torque at 72V and 20A.

It would be interesting if someoby can test if the V3 at 36V will give the same top speed as the V2 at 48V.

Or equivalently if someone could measure the internal resistance of the V3 and V2, or countthe windings, the current voltage scaling factor for equivalent performance could be calculated. Easiest would be perhaps to measure Ri: Put a known current (from 1 to 10A is likely safe) through the windings using a lab supply or current limited charger, measure voltage over windings. Divide the V3 and V2 voltages to get the conversion factor.

If the V3 is good at 36V there would be a BMC motor to match each common battery voltage 36, 48 and 72V and give good everyday performance (30mph + on flats, strong uphill)

In particular it would be a nice option to combine the BMC v3 with the new A123 prismatic 20Ah cells, since 12 of those can be put in a very compact (23x17x9cm - Perfect for the frame triangle above cranks.) and light (5.8kg) package and give a respectable 800Wh. I was thinking about the A123 prismatic's for my 72V setups, but cost and weight of 24 such would be too high. 12 cells at 36V I think is already more capacity than most of us use.

[oatnet]

He also recommended cutting stock phase wires as close to the shaft as possible and replaing them with 10ga wires (which I'm going to do).

Good, that means that the 8ga I replaced my V3 phase wires with in September will be overkill for the controller... which is just about right for my riding style.

I noticed that the phase wires on the controller each had (2) wires of the same gauge as the phase wire on the motor. Even so, I cut back the phase and battery wires on the controller to reduce clutter and resistance.

-JD

[ebikes-sf]

Nice setup JD. How is it handling 50 A controller?

I am thinking of improving wires inside inside motor shaft. Instead of using multi stranded wires with thick plastic insualtion, eliminate the gaps between strands by using solid wire with thin enamel insulation. Something like this:

[nicobie]

Hey Ilia,

I wish I could remember who did it, but somebody used flat ribbon wire inside of shrink tubing and got decent phase wire size in a BMC.

[Sacman]

Hey Ilia,

I wish I could remember who did it, but somebody used flat ribbon wire inside of shrink tubing and got decent phase wire size in a BMC.

That was Geoff57 who used flat-braided wires to upgrade the phase wires on his motor back in July 2008. He was working on a Puma Mk2 motor which is almost the same as our BMC motors that we have today. Here are links to his 2 threads...

http://endless-sphere.com/forums/viewto ... 024#p68713
http://endless-sphere.com/forums/viewto ... 110#p68110

What a coincidence... I'm finally putting the time to upgrade the melted wires in my BMC motor, and I log back into ES (after a long sebatical) to review just this same info ... and I find you guys talking about it. I'll keep you guys posted on how it goes for me... maybe on a separate thread.

[TPA]

Any closup pics of the axle hole?

[fechter]

Solid wire is going to be pretty hard to stuff through the axle, especially where it goes around the corner.

Splicing thicker wire as soon as it exits the axle is a good approach. Much of the heat from the skinny part will transfer to the thicker part.

What's really needed is a larger diameter bearing on the wire side so the wires can go in a channel on the outside of the axle like a Crystalyte X5 (thicker axle). This would eliminate the hollow axle, so would be stronger, and stuffing the wires would be way easier.

A thermistor is good. I installed one on mine, but haven't got it running yet. This is the only way to get good perfomance and still be fairly idiot-resistant.

[Sacman]

Solid wire is going to be pretty hard to stuff through the axle, especially where it goes around the corner.
What's really needed is a larger diameter bearing on the wire side so the wires can go in a channel on the outside of the axle like a Crystalyte X5 (thicker axle). This would eliminate the hollow axle, so would be stronger, and stuffing the wires would be way easier.

Or.... drill out the other end of the axle so wires can be fed thru both ends.

[ebikes-sf]

Finally made some progress. Here is diagram of how to hook up direct plugin Cycle Abalyst to the BMC 50A controller. This mod allows CA to use controller's shunt to measure current and override throttle signal when current or speed limit is exceeded. It works quite well, no oscillation.



I also tested routing throttle signal through Cycle Analyst, as outlined in CA manual section 9.3, however, due to the specifics of this 50A controller it resulted in constant oscillation at any motor speed. Justin explain that the controller must have a slight throttle response delay which results in oscillation.

I've pretty much ginven up on the idea of routing thicker wires through the axle. fetcher is right, feeding thicker solid wire is possible but very hard. Drilling through the other side of the axle would solve the problem but I don't have setup to do it.

So I will do what oatnet suggested (and what Chandu at BMC suggests) - use thicker wire from axle to controller, now with abllity to limit current it should be ok.

Next step is a road test.

[jag]

Finally made some progress. Here is diagram of how to hook up direct plugin Cycle Abalyst to the BMC 50A controller. This mod allows CA to use controller's shunt to measure current and override throttle signal when current or speed limit is exceeded. It works quite well, no oscillation.


Next step is a road test.

Hi Ilia,

Thanks for the update and look inside the BMC controller. Did you take pic of the other side also? What FET's are in there?

In the previous test I believe you used a 48V battery. I would be very interested in 36V performance (speed on the flats, hill climbing and any subjective impressions)

Cheers,
Martin

[ebikes-sf]

Martin, here is a pic of the other side, controller uses 18 x IRFB 4110. I will be road testing at 48v and 36v on Saturday and share results around Sun/Mon.



ilia

[Sacman]

Hi guys,

I'd like to share my test results with this BMC V3 Motor & 50A Controller too.

After a full year of recovering from a broken shoulder I'm sure glad to be getting out riding my ebike again. I did some upgrading and installed the BMC V3 1000W Motor & 50A Controller last month. I also constructed more powerful battery packs (Headway 36v 40Ah). I've been riding this setup for the past 2 weeks now and I've been very pleased with it. I started out with a short 6 mile ride but have been increasing my distance each time out and my last ride this past Sunday was a very long 86 miles. So I thought I’d share some my results running the BMC V3 Motor & 50A Controller at 36 volts.



CONTROLLER FEATURES – The controller has colored jumper wires that I’ve extended up to the front handle bars where I have switches mounted. Closing or Opening these Jumper Wires on the controller gives the following:
- Hi/Lo Speed Control
- Cruise Control
- Soft/Hard Start

SPEED CONTROL:
At 36 volts this motor/controller combination has MORE THAN ENOUGH speed for my commuting purposes. Full throttle on LOW Speed gets me up to 24mph. Full throttle on HI Speed gets me topping out at 32mph. I consider this pretty good because I'm a pretty heavy guy (weighing 280 pounds) plus I've got 40 pounds of batteries. Most of the time I ride on LOW Speed since my comfortable riding speed is between 17mph and 25mph anyway. So I really have no need to run at 48v to go any faster.

CRUISE CONTROL:
I’m pleased to say that BMC’s cruise control acts more like a “real cruise control” than Crystallyte’s does. Once the cruise control is set, the speed stays pretty constant even when riding thru moderate (5%) grade changes in slope. As I transition from flat to a 5% uphill grade I can actually see amperage and wattage on the Cycle Analyst increase to more than double to maintain constant speed. Likewise, power drops down when going downhill. In contrast, the “fake” cruise control on the Crystallyte acts more like a “throttle latch” rather than a cruise control. It doesn’t maintain speed at all.

SOFT START:
I really like riding with this feature turned on. I’ve noticed that throttle response on regular “Hard Start” is very twitchy when using these BMC V3 and V2T motors. You have to be easy on the throttle when you accelerate from a dead stop otherwise the bike lurches forward violently. The motor also lets out an unpleasant “moan & groan” noise until your speed increases above 5mph then the sound quickly tapers down. With the Soft Start turned on the bike doesn’t lurch forward and I don’t hear the awful moaning noise when accelerating hard from a dead stop. Just smooth, steady and strong acceleration. There is, however, a short ½ second delay between applying full throttle and the bike actually accelerating, but I got used to that short delay quickly.

CLIMBING HILLS:
I can’t give a thorough evaluation about this yet. I really need to do more testing with this setup on more hills. I have tried climbing a fairly “monster hill” on my commute that varies between 10% and 16% (according to MapMyRide.com) for a good ¾ mile. My system is capable of 1,950 watts (39V x 50A), but I was taking it easy and not giving full throttle. I was confident to only give enough throttle to sustain 1,200 watts during that test climb because I was many miles away from home and last thing I wanted to do was fry more phase wires or burn up my motor and controller and be stuck so far away from home. But at 1,200 watts I was still able to steadily climb that hill at 12mph with some mild pedaling. I can also say that this V3 motor climbs hills very similar to my V2T motor (the one I melted phase wires in last year). As I get more confidence with this system I’ll try steeper, longer hills with more throttle and let you guys know what I find.

[Mark_A_W]

My primiary focus of testing is on ablilty to climb hills (I live in San Francisco). Based on my experience with V2's, hall sensor tend to get damaged when the motor is overheated, this is more common for V2 torque models since they take more current at low motor speed than speed V2S, and they're usually operated at more power than rated 600 Watts. It seems construction of V3 motor is not much different from V2, so V3 isn't able to disperse heat generated from coils any better than V2.

Im my opinion, 1KW V3 motor should not run at much more than 1KW for reliable service independant of terrain or outside temepratures. Running it at 2500W (48V x 50A), even for short periods of time, is really fun but very abusive for the motor. Making phase wires thicker will only expose other bottlenecks like burned coils, blown hall sensors, burning lubricant, etc...

As far as I know, the only major construction improvement in V3 is that the gears/freewheel clutch assembly is now bolted down instead of just sitting on the axle. On the picture below, note that V3 assembly (right) has 3 mounting holes that V2 (left) doesn't have.

Ebikes (Illia?)

I can't see how the freewheel clutch assy (the planet gear carrier) can be bolted to anything through those holes around the outside of the carrier. The carrier needs to spin around the axle when you are freewheeling, and the hub case spins around the now stationary carrier when the motor is driving.

Maybe I'm missing something...


Oh, oh, and do you sell spare green BMC composite gears?


Oh, oh, oh, can the BMC controller be run at 60v and ~30 amps. (Shunt mod or reprogrammed?)

[jag]

Hi guys,
I thought I’d share some my results running the BMC V3 Motor & 50A Controller at 36 volts.
:
Full throttle on LOW Speed gets me up to 24mph. Full throttle on HI Speed gets me topping out at 32mph. I consider this pretty good because I'm a pretty heavy guy (weighing 280 pounds) plus I've got 40 pounds of batteries. Most of the time I ride on LOW Speed since my comfortable riding speed is between 17mph and 25mph anyway. So I really have no need to run at 48v to go any faster.
:
My system is capable of 1,950 watts (39V x 50A), but I was taking it easy and not giving full throttle. I was confident to only give enough throttle to sustain 1,200 watts during that test climb because I was many miles away from home and last thing I wanted to do was fry more phase wires or burn up my motor and controller and be stuck so far away from home. But at 1,200 watts I was still able to steadily climb that hill at 12mph with some mild pedaling. I can also say that this V3 motor climbs hills very similar to my V2T motor (the one I melted phase wires in last year). As I get more confidence with this system I’ll try steeper, longer hills with more throttle and let you guys know what I find.

Thanks for the detailed review of performance and features. Looks like you have a nice, clean setup.

Do you use the original phase wires? Or did you put in bigger wiring? What did you use for on-off switch? (The BMC controller unlike the Crystallyte seems not to have a switch on the controller case)

32mph at 36V is good news for those who plan to use a low voltage pack. Of course it is always tempting to go 48V...

I recall you used 72V nominal for your previous v2T setup. One thing to note is that when hill climbing at a particular hill and speed, say your example 12mph up a 12-16% hill and the motor uses a fixed wattage (here 1200W), the current in the v3 phase wires will be approximately twice as high as in the v2T. In each case the torque produced by the motor is the same, the magnetic field strength in the motors the same, but v2T has about twice as many wiring turns (would be nice if someone counted exactly the writing turns in the v2's and v3), so it takes 1/2 the current through the v2T phase wires.

[Sacman]

Do you use the original phase wires? Or did you put in bigger wiring? What did you use for on-off switch? (The BMC controller unlike the Crystallyte seems not to have a switch on the controller case)

All the phase wiring is still original, stock size. Although I agree with the benefits of going to thicker wire (and I'm inclined to change them later on), right now I'm still testing and I'd like to see what the original equipment is capable of first before I do any mods. I have no on-off switch yet. I just have Anderson connectors between the battery pack and controler. I hve a Crystallyte Keyswitch with 10 guage wire that I do plan to use later on.

I recall you used 72V nominal for your previous v2T setup. One thing to note is that when hill climbing at a particular hill and speed, say your example 12mph up a 12-16% hill and the motor uses a fixed wattage (here 1200W), the current in the v3 phase wires will be approximately twice as high as in the v2T. In each case the torque produced by the motor is the same, the magnetic field strength in the motors the same, but v2T has about twice as many wiring turns (would be nice if someone counted exactly the writing turns in the v2's and v3), so it takes 1/2 the current through the v2T phase wires.

You remember right. I was using 72v nominal with the V2T Motor but it was a cheap 10Ah ductape battery pack with lots of voltage sag (about 12volts). So when I was pushing 1800+ watts thru the V2T I was actually running at only 60v. I don't have that voltage sag problem with the current 40Ah battery pack I'm using now. But your point about "twice the current at half the voltage" leads me to ask the question.... What really causes the wire to heat up and melt? Is it current (amperage)? or is it power (wattage)?

[Sacman]

Finally made some progress. Here is diagram of how to hook up direct plugin Cycle Abalyst to the BMC 50A controller. This mod allows CA to use controller's shunt to measure current and override throttle signal when current or speed limit is exceeded. It works quite well, no oscillation.

I also tested routing throttle signal through Cycle Analyst, as outlined in CA manual section 9.3, however, due to the specifics of this 50A controller it resulted in constant oscillation at any motor speed. Justin explain that the controller must have a slight throttle response delay which results in oscillation.

Ilia,

Thanks for this diagram and the good info on hooking up the direct plugin CA to this BMC 50A controller. I'm curious and got a couple of questions about your hookup:
1) What RShunt value are you using on direct plugin CA to calibrate it to this BMC controller? (correct RShunt value is needed to display accurate amps & power)
2) How do you measure your speed? Are you using a direct plugin CA with a speedo sensor?

[ebikes-sf]

I can't see how the freewheel clutch assy (the planet gear carrier) can be bolted to anything through those holes around the outside of the carrier. The carrier needs to spin around the axle when you are freewheeling, and the hub case spins around the now stationary carrier when the motor is driving.
Maybe I'm missing something...

Sorry for false information. I removed my comment about V3 gear assembly being bolted down (posted that before I took those motors apart), V2 and V3 motor assemblies are identical:

The only differece is v3 has thicker coils and supposedly improved freeweel/clutch assembly.

Oh, oh, and do you sell spare green BMC composite gears?

Yes, $65 V2 gears, $105 for V3. I don't know why such big difference in price, hopefully V3 clutch is really that much stronger.

Oh, oh, oh, can the BMC controller be run at 60v and ~30 amps. (Shunt mod or reprogrammed?)

I don't know if it can be reprogrammed. I can think of a few options to limit current . Either use CA mod as i demonstrated above. Or use controllers limiting jumpers (see previous page). Or change shunt resistance.

[Mark_A_W]

While I have you, got any spare green gears for sale?

[ebikes-sf]

Sacman, thank you for your nice post!

1) What RShunt value are you using on direct plugin CA to calibrate it to this BMC controller? (correct RShunt value is needed to display accurate amps & power)

It actually differs between individual controllers. CA must calibrated to specific controller, easiest way is to use stand alone CA and calibrate to that.

2) How do you measure your speed? Are you using a direct plugin CA with a speedo sensor?

Right, using direct plugin CA with a speedo sensor.