1000W BMC testing

[dirtdad]

The new 1000W BMC hub motor (V3) has proven to have a lot of potential that we (BMC and retailers) cannot figure out how to unleash...yet anyway.

We have tested with the Crystalyte 20 and 40A controllers with disappointing results. Speeds in the low 20s even at 72V, and controllers that heat up very fast. Soft start and current limiting help with overheating. I dont understand it, but BMC and others think the motors low resistance may cause the controllers to be heating up. I thought high resistance would result in heat...but there is more going on in this case with PWM and all.

This evening I just tested with the 72V 65A Infineon controller (18 beautiful 4310 fets). Good news and bad news. This controller definitely unleashed the motor - for about 1/4 mile. When I say unleashed, I mean un-freaking-leashed. At partial throttle I was accelerating HARD at 40mph before I let up because it was dark out and I had no lights except to read my speed. Under those conditions and in that short a distance the phase wires got very warm. And then the controller threw the thermal breaker. After letting is sit for a few minutes it was operating again.

The Infineon uses a high clock speed processor, we think this may be key to getting it to drive the V3 to is full potential. I dont know how it compares to the Crystalyte controller in that regard. Calling Knuckles... The Infineon does not appear to be the solution for now, but I think it does help us figure this out.

I dont know if the Infineon I used has soft start or not. It is the latest and greatest from Keywin. They used to have soft start. If it does, it is not soft enough. The best modulation I could muster with micrometer throttle movement resulted in 3 modes - nothing, a very light output, and BANG, afterburner kick in the pants, after which you still had 95% of the throttle left and no sign of let up in power as you turned the throttle. Not a good throttle curve to say the least. I am afraid I will toast a clutch like this. During the 1/4 mile I got it to work, it put the X5 motor to shame. I mean like not even close. I will try a 36V Infineon next.

Bottom line: this motor is a beast that we have yet to tame, but its potential seems enormous. A couple of us will be testing with the Kelly 120A (peak) 50A sustained controller next, and BMC is working on their own controller.

 

[Drunkskunk]

Taking a guess from what I know of RC motors, Since the Clyte controllers are ment for low RPM motors, the BMC might not be throwing back enough EMF to choke the amp draw. Wrong clock speed, wrong pulse width, too low resistance, wrong delta on the EMF ramp up, and wrong timing.
I could never make an ESC set up for a 700kV motor run a 5000kv without trouble.

What about trying one of the RC ESCs the non hub guys are modifying?
The HV110 was ment for higher RPMS, all it needs is some form of current limiter.

Are there any plans for a torque version of the V3?

 

[fechter]

I dont know if the Infineon I used has soft start or not. It is the latest and greatest from Keywin. They used to have soft start. If it does, it is not soft enough. The best modulation I could muster with micrometer throttle movement resulted in 3 modes - nothing, a very light output, and BANG, afterburner kick in the pants, after which you still had 95% of the throttle left and no sign of let up in power as you turned the throttle. Not a good throttle curve to say the least. I am afraid I will toast a clutch like this. During the 1/4 mile I got it to work, it put the X5 motor to shame. I mean like not even close. I will try a 36V Infineon next.

That does not sound right. Possibly a bad throttle? Otherwise it sounds like something in the throttle circuit is messed up. Try measuring the throttle signal wire against ground while advancing the throttle (you can disconnect the phase wires or lift the wheel). It should go roughly from 1 volt to 4 volts as you advance the throttle.

 

[Knuckles]

dirtdad

Generally this sounds like good news. Definitely check the throttle however. Maybe try another throttle.

The issues seem to be ….
1. Hard Start and Soft Start.
2. Processor Frequency and Performance.
3. Looking for “Jitter”.
4. Phase Wire Heating.


1. Hard Start and Soft Start.
Chances are you have the “original” hard start Infineons (12 and 18 fet) from Keywin.
Keywin ONLY started provided soft start because Mark in UK (TeamHybrid) needed soft start for his PUMA wheelchair motors. Hard start was just too CRAZY to comfortably control the PUMA in a 3-wheeled (wheelchair) vehicle. Hard start, however, is preferred for the 2-wheeled GANG that like high performance and Jack-Rabbit starts. BTW … The NEW “116” processors are (so far) all hard start.

2. Processor Frequency and Performance.
Looking at Kelly BLDC controllers …
http://www.kellycontroller.com/mot/Brushless-DC-Motor-Controller.html
They claim …
“Support any number of poles”
“Up to 60000 electric RPM (electric RPM = mechanical RPM x motor poles)”
“Frequency of Operation: 16.6 kHz”

The Infineon is a 27 kHz processor (the new “116” is also about 27 kHz)
It is not only the processor speed that determines performance.
It’s the “QUALITY” of the programmed PWM software that is just as (if not more so) important.

3. Looking for “Jitter”.
Put the motor-controller “system” on an oscilloscope and look for the tell-tail “jitter”.
Also. Conduct a no-load test and measure the overall battery current draw under no-load.
The onset of ‘Jitter” will cause a sudden increase in no-load current draw as the throttle is slowly increased.

4. Phase Wire Heating.
Thin phase wires (Standard 16-gauge silver conductor on most Chinese motors) will get hot under a heavy current load.
“Jitter” will also cause current spikes because the controller can’t correctly commutate the necessary high electric RPM.
Is Jitter the problem or are the phase wires just to wimpy to handle the amps?

 

[BVH]

Is this the new BMC 1000 Watt Black Lightning - geared motor or a new gearless model?

 

[morph999]

Wow, sounds awesome. Dirtdad, can you tell us the speed at say 36v...and then at 48v.

The 5303 goes about 36 mph at 48v and about 28 mph at 36v. I'd like to see a comparison on those specific volts if possible.

 

[dirtdad]

This is 1000W geared. BMC will confirm no plans to make a gearless model. I will test more voltages when we have a working controller, which we really dont...yet.

Throttle tests OK. A different throttle behaves the same, both throttles work fine on a C'lyte ctlr.

I exhaustively tested all 36 phase/sensor wire combinations. 3 combos gave this same result. No combination that drove the motor behaved any differently.

These combos draw high amps at no load, peaking at about 14A, and drawing about 9A sustained.

 

[docnjoj]

Hey dirtdad. I had a similar experience with an older Crystalyte 36-72 volt 20 amp analogue controller and a Bafang. It would pull 10-12 amps no-load start up current no matter which of Knuckles 36 combos I tried. I wasted a whole day trying them out. Got an Infineon 48 volt 30 amp controller and Shazam! 3/4 amp start-up and very smooth quiet power on the Bafang. Irealize the BMC is a much more powerful motor (envy, greed) but those start up currents seem excessive.
Just my 2 centavos!
otherDoc

 

[dirtdad]

Yes. I just tested a little 20A Clyte digital 36-72V. Pulls 12A through it no load. I took it for a ride earlier with that controller. It also threw an internal thermal breaker after mild use. I presume a thermal breaker. It acted just like it. But the controller was not that hot to the touch. This looks similar to the case that user "Maud 'dib" is also reporting on this forum on this motor with a 48V infineon.

http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=11124&hilit=1000W

 

[Sacman]

Yes. I just tested a little 20A Clyte digital 36-72V. Pulls 12A through it no load. I took it for a ride earlier with that controller. It also threw an internal thermal breaker after mild use. I presume a thermal breaker. It acted just like it. But the controller was not that hot to the touch. This looks similar to the case that user "Maud 'dib" is also reporting on this forum on this motor with a 48V infineon.

http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=11124&hilit=1000W

Dirtdad,

It might not be a thermal breaker that caused your controller to cut out. I do remember that the 20A 36-72V Crystalyte controller consistently shutting down on me without it getting hot (but only when going up steep hills). But then I could turn it off and after 30 seconds I turn it back on and it was fine again.

Your controller failures seem consistent with Muad Dib's in that you were both able to run it fine for awhile before the controllers shut down. Only difference is that the 3 controllers (1 Infineon + 2 Crystalytes) you tried just shut down, whereas Muad Dib actually fried components on his controller.

What really get's me is that you are still reading unusually high current draw (12-14 amps) on all 3 of your controllers with no throttle even tho you have already tried all 36 combinations for the hall sensor and phase wires. I could be wrong but it's starting to appear very suspicious that the source of the problem is in the new BMC 1000w geared motor.

 

[jag]

Dirtdad,

You also have a Crystalyte 5-series right? So I take it you found the the BMC 1000W much harder on the controller than the big Crystalyte?

Controller design for a wide range of motor characteristics can be tricky business. Like Knuckles said, going above max frequency is one possible problem; but that ought to show at high speed, not as difficulty running smoothly at low speed.

Getting the on-off timings right on FET's (and semiconductors in general) is also tricky business. Maybe at low RPM's the combination of low internal resistance and little back-EMF creates out-of-spec currents and slew rates, leading to FET's being ON-ON for brief periods. This could explain why the controller draws lots of amps despite no or little motor load.

Are detailed specs available for the BMC's? Winding configuration? Number of turns? Internal resistance? Poles?

Or by posting some photos comparing the internals between the 600W and 1000W poles and maybe turns can be counted, internal resistance estimated given wire gauge (thickness) etc.

I had a summer internship designing the PCB and building prototypes for a 3phase 380V industrial motors controller some years ago. Getting it to work robustly over the range of motors (sizes from few hundred W to kiloWatts) was difficult. The wave modulation was digitally controlled and the guy who did the circuit design (not me. I barely understood it) had done a lot of optimization to account for semiconductor properties, minimize on board resonances, emitted RF etc, and maximize efficiency.

There was an article in Newarks Electronics magazine (Feb or March issue) on waveform design for motor controllers. Article claimed significant gains in efficiency were possible over current practice, and thus less waste heat in controller, more power to motor. Couldn't find an online version, but when I get to a scanner I'll try to scan it.

 

[dirtdad]

Dirtdad,
You also have a Crystalyte 5-series right? So I take it you found the the BMC 1000W much harder on the controller than the big Crystalyte?

No. We dont have a controller that works at all yet. Once we get one that works, then we will know.

Getting the on-off timings right on FET's (and semiconductors in general) is also tricky business. Maybe at low RPM's the combination of low internal resistance and little back-EMF creates out-of-spec currents and slew rates, leading to FET's being ON-ON for brief periods. This could explain why the controller draws lots of amps despite no or little motor load.

It would also explain the heat. We are thinking along these lines.

Are detailed specs available for the BMC's? Winding configuration? Number of turns? Internal resistance? Poles?

I can probably get them. If I can, I will post them here.

 

[Iroise2000]

Hallo,

I live in swiss and just receive last week my BMC V3 from electric-bikes.com in a 26" wheel. They are very professional, and the pakage was well made and fast secure shipping.
Today i've made my first test !!!!, just a word : awesome !!!!!
I have a 48V 20Ah ping battery V2 + a new kelly controller (KEB72451,150A 30 sec and continus phase 75A,72V,4.5KW /Regen). This controller is software programmable, so i put max 50% continus phase, 37A, and slow throttle.
Whoua, the first run was super, and i goes up, but after 5 minutes, all stops !, i thought it was ping BMS, but in fact when i was at home, i see that 1 phase of the anderson connector as melted !
So i cut all the anderson connector and join all with soldering...

So in flat, the bike goes at about a max 35.2 mph or 56.7 km/h, pretty fast for me ! I had first a 1000W golden motor, but can go at about 45km/h. The motor is very silent, just at starting there is a noise, and after 10 km/h very silent. The kelly controller is very smooth.
For moment i haven't made a measure of the peak Ah, i must connect my wattmetter.

So for me, this first test is very very nice !, i'm so happy. I cannot imagine what it can be with more volt..., but 35 mph is much more than OK !

If yu have questions, don't hesitate. I see that someone had problem with this motor, but with kelly controller, it seem to go nice. But only 25 km for moment.

Yann
Swiss

 

[BVH]

Thanks Iroise2000 for your post. I should have my new rig with a Version 3 BMC next week and was getting kind of concerned about the somewhat negative data in this and the other Ver 3 thread.

The Kelly looks like a first rate controller. Has anyone here used one? This is the first post I've seen referencing one. Software configurable sounds like a great feature. If I've missed a thread on them, please direct me.

 

[grandmasterE]

Dirtdad,
You also have a Crystalyte 5-series right? So I take it you found the the BMC 1000W much harder on the controller than the big Crystalyte?

No. We dont have a controller that works at all yet. Once we get one that works, then we will know.

Getting the on-off timings right on FET's (and semiconductors in general) is also tricky business. Maybe at low RPM's the combination of low internal resistance and little back-EMF creates out-of-spec currents and slew rates, leading to FET's being ON-ON for brief periods. This could explain why the controller draws lots of amps despite no or little motor load.

It would also explain the heat. We are thinking along these lines.

Are detailed specs available for the BMC's? Winding configuration? Number of turns? Internal resistance? Poles?

I can probably get them. If I can, I will post them here.

Any luck tracking down the specs yet?

E.

 

[ebikes-sf]

I experienced somewhat similar issue with Infineon 30A controller and 600W-S BMC motor where motor ran very rough. But when I tried same with 400w BMC, it ran very smooth. The only difference between 2 motors (from controller's side) is the coils' resistance (don't know if number of poles is different).

 

[fabiograssi]

Hallo,

I live in swiss and just receive last week my BMC V3 from electric-bikes.com in a 26" wheel. They are very professional, and the pakage was well made and fast secure shipping.
Today i've made my first test !!!!, just a word : awesome !!!!!
I have a 48V 20Ah ping battery V2 + a new kelly controller (KEB72451,150A 30 sec and continus phase 75A,72V,4.5KW /Regen). This controller is software programmable, so i put max 50% continus phase, 37A, and slow throttle.
Whoua, the first run was super, and i goes up, but after 5 minutes, all stops !, i thought it was ping BMS, but in fact when i was at home, i see that 1 phase of the anderson connector as melted !
So i cut all the anderson connector and join all with soldering...

So in flat, the bike goes at about a max 35.2 mph or 56.7 km/h, pretty fast for me ! I had first a 1000W golden motor, but can go at about 45km/h. The motor is very silent, just at starting there is a noise, and after 10 km/h very silent. The kelly controller is very smooth.
For moment i haven't made a measure of the peak Ah, i must connect my wattmetter.

So for me, this first test is very very nice !, i'm so happy. I cannot imagine what it can be with more volt..., but 35 mph is much more than OK !

If yu have questions, don't hesitate. I see that someone had problem with this motor, but with kelly controller, it seem to go nice. But only 25 km for moment.

Yann
Swiss

Dear Yann:

Can you update your impressions? Did you ride more? How many Km's now?

 

[dirtdad]

I am setting up a Kelly KBS72121 conroller, it is one of their new micro controllers. I am a bit skeptical due to its size, but I have it working on the bench no load. They claim it is 24-72V, 50A sustained, 120A peak. That would take more fets and caps alone than you can put in that small case, never mind heat dissipation. But that is why I am testing.


Wiring colors are a direct match with the BMC

No load on the bench is still a bit high, about 4A sustained at full throttle. But with that big tire, heavy slime tube, and DH rim it is not quite a no load test. And 4A is much lower than the Infineon or Crysalyte controllers. And it stays very cool during no load testing whereas the other controllers got hot even doing this test, then overheated quickly on the road.

Nice controller so far. It came with the serial adapters to program it, but I needed to provide my own usb-serial cable and driver, which I did already have. The software installed nicely, is well documented, and in good English. The manual is nice and complete and also written in good English. It looks like a Chinese company has taken the extra effort for US/English support.

This is the regen controller, all of the mini controllers are. With the freewheeling hub I cannot use regen. But there are a lot of programmable features in the controller for regen, it has a lot of potential but needs further testing to prove itself.

 

[voicecoils]

That 4A current no-load sounds more reasonable, still surprisingly high though.

Do you have access to any of the beefier Kelly controllers? Methods on this forum is used a monster Kelly for 200A+ on his toasted X5.


I have that same 'truing stand' :lol:

 

[Iroise2000]

Hi,

So just a week with my new BMC V3 !, I've made about 140 km with no problem !, i'm so happy.

Today saturday, i put anderson connector on my wattmeter and i've made a 10.82 km ride. I run pretty fast with stop and start and a small hill with 3%.
The registered max Ah is 39.85. The cunsumption for the 10.82 km is 4.55 A, so it is 0.42A for one kilometer.
With my 20Ah 48Volt ping battery,i estimate i can ride with the same speed (not economic !!!!!!! - 44km/h moyenne) about 47 km, pretty good for me, to be safe i will put 40km max.



On this graph (gps garmin), yu can see the speed for the 10.8 km ride. Yu can also see the up and down in percent. Yu see i was very often more than 50 km/h !!!!!!
Ohhhh : on the small uphill, i ran no problem 40-50 km/h (without pedalling).

For moment, the only problem i have : my ping battery is 10kg on the back in a box, it is ok, but i will prefer a lipo in the triangle...

Just 2 photos of the bike :





i have the KEB72451 Kelly controller. With the parameter i put in (50% for the max current to be safe with my ping battery), it is just a little warm, as the motor, but i can touch it without problem with hands, so i think 40-50 degrees).

Yann

 

[fabiograssi]

...With my 20Ah 48Volt ping battery...
...i have the KEB72451 Kelly controller. With the parameter i put in (50% for the max current to be safe with my ping battery)...
Yann

Dear Yann:

I have no knowledge about this controller. In the kelly website, I saw that is a 72 volts controller. Can be used at 48 volts too? I'm asking because in the mentioned website I found some controllers that accepts voltages between 24 and 72 volts, like the new mini ones.

And what about the current? 50% of 150A is 75A. The ping pack can handle with it without problems? I think 2C is more indicated for this pack (40A), or I'm wrong?

 

[Sacman]

I wonder what controller Chris at Hi-Power Cycles is using with his 1000w BMC motor when he made this video

[youtube]7oLQdXw9K_M[/youtube]

 

[Iroise2000]

...With my 20Ah 48Volt ping battery...
...i have the KEB72451 Kelly controller. With the parameter i put in (50% for the max current to be safe with my ping battery)...
Yann

Dear Yann:

I have no knowledge about this controller. In the kelly website, I saw that is a 72 volts controller. Can be used at 48 volts too? I'm asking because in the mentioned website I found some controllers that accepts voltages between 24 and 72 volts, like the new mini ones.

And what about the current? 50% of 150A is 75A. The ping pack can handle with it without problems? I think 2C is more indicated for this pack (40A), or I'm wrong?

Hi,
it is the Kelly KEB72451, yu can use it with 24-to 72 volts. No problem with 48V. I've bought this one, so one time i will try more voltage. In the instruction, there is written yu can handle to 1.25 nominal voltage, so 72x1.25 = 90 Volts. My ping battery when full charged has 59 volt, and it goes to 52-53 rapidly.
But my controller has regen, there is the same without regen for BMC, it is the KEB72450. Regen is disabled in the software program.
The 72450 or 72451 has continus 75A and 150 in peek. So 50% is about 40 A continus, it's just ok for my ping !!!! at 2C.
For the ride, my wattmeter schow no more than 40Ah in peak !, i can't imagine how it goes with a lipo pack 66V with 15C-25C !!!!!

Yann

 

[fechter]

No load on the bench is still a bit high, about 4A sustained at full throttle. But with that big tire, heavy slime tube, and DH rim it is not quite a no load test. And 4A is much lower than the Infineon or Crysalyte controllers. And it stays very cool during no load testing whereas the other controllers got hot even doing this test, then overheated quickly on the road.

4A sounds about right to me. I don't think you can get it much lower than that.

The new micro Kelly looks suspiciously similar to the Infineon controllers.

 

[dirtdad]

I have all of the Infineons and the Kelly mini seen above. I ordered the Kelly with the "waterproof" option without thinking about it. It is pretty well sealed up with silicone and I suspect it is also potted so I am not going to open this one (see above) to look at the insides. That would be the best way to compare. If I get more Kelly controllers I may open one then.

There are a lot of external differences between the two. Kelly programming software is totally different from the Infineon software and allows you to program different parameters. The Kelly comes standard with a voltage regulator that allows a wide range of voltages, something that you have to mod the Infineon to do. Kelly packaging is different from the Infineon: the form factor is smaller and different from any of the 3 Infineon cases and has completely different bolt patterns. Plus there are the differences in heat and no load current draw noted above. That does not mean they couldn't still have a lot in common like the main processor, etc.

I still think the Infineon could handle the BMC 1000W but we may need to put the combo on a scope to see what is going on.

I think HiPowered used an analog Crystalyte, this is a pic of the controller they used. I got this information customer of theirs that bought the 1000W motor and the controller shown. Hi Powered are selling the 1000W with this controller.