New Kelly pseudo FOC controller series

[bowlofsalad]

ERPM is electrical rotations per minute. The wheel might rotate at 2000 RPM or something like that, but the motor is performing an ERPM much higher. ERPM relates to two other things so far as I know, pole count and reduction ratio. Perhaps a mac has a reduction ratio of 5 and 16 poles. I could be completely incorrect here, but how I think approximate ERPM is calculated is RPM x reduction ration x pole pairs. So in your case, perhaps it's 2000 x 5 x 8 which tells us 80,000 ERPM.

In summation, search.php?keywords=ERPM&t=71942&sf=msgonly Per searching ERPM in this kelly FOC thread it looks like the ERPM for their KLS controllers is a maximum of around 70,000, which may explain your issue. Perhaps find the motors ERPM and see if it's indeed below 70,000.

 

[izeman]

tested it with my klein mantra bike and HS3540 installed. though this motor is meant to run at 80V to get to speed i connected the 50V controller to it. did the auto_setup and what can i say: ABSOLUTELY NICE. smooth, totally silent and full of torque. a pleasure to ride. no pumping rpm not stalling, no overshoot, nothing.
this MAC motor must be something really special. and i'm out of ideas what to do next. i don't think that kelly can make some special firmware for the MAC. i hope they tested those motors with the controller?!
it could be erpm limiting. looks like, but fany told me that those new controllers use a different kind of sensing and can go up to 70kerpm. and with 2000rpm*16 i'm only at 32kerpm. so that should be no problem.
i wish fany could join as again. i hope chinese holidays don't take too long ...

edit: bowl_of_salad: 2000rpm is the bare motor w/o reduction. so no need to multiply by 5, and erpm is calculated rpm*polepairs, or rpm*poles/2.

 

[izeman]

i now recall what is important for me, and was told by fany:

I sent the new App to release this parameter so that SplinterOz can configure it at 1000.
Then we get the higher speed. If some customers still can not get the expected speed,I can send the special APP for him to configure the parameter,especially for the motor with higher poles number.

maybe this can help me? i need the special app. splinteroz: could you please tell me what this parameter did for your setup and if it helped?

 

[SplinterOz]

i now recall what is important for me, and was told by fany:

I sent the new App to release this parameter so that SplinterOz can configure it at 1000.
Then we get the higher speed. If some customers still can not get the expected speed,I can send the special APP for him to configure the parameter,especially for the motor with higher poles number.

maybe this can help me? i need the special app. splinteroz: could you please tell me what this parameter did for your setup and if it helped?

It allowed me to change the "Max Output Fre" to 1000. Sorry I have been offline for a couple of days.
I can say that I had issues like yours and it did seem to fix them.

I thought Fany updated the software on the site.

 

[madin88]

hopefully this additional setting will help spinning the mac properly.
thanks for testing the controller with the DD motor. i'm happy to hear you got good torque from stand still.

btw: adpatto controllers do have a setting called "BMC hall fix: yes / no" and from what i know this motor should be electrically identical to the MAC. they recommend honeywell ss413a for the halls

 

[izeman]

hopefully this additional setting will help spinning the mac properly.
thanks for testing the controller with the DD motor. i'm happy to hear you got good torque from stand still.

sorry for not taking a video. but you make audio yourself: stick your fingers in your ears. what you hear is the sound of the motor. nothing. yes. you can hear it turn with the wheel lifted, but that's completely gone when riding. wind, wheels and others sounds are a multitude higher than anything coming from the hub.

 

[izeman]

I thought Fany updated the software on the site.

BIG FACEPALM! i didn't realize they updated the software as well. i thought it was the manual only. downloaded and yes, you can set MAX FREQUENCY with it. arghhh...
i will test asap and am quite confident that it will solve the issues.
maybe kelly should add VERSION NUMBERS to their software as well so ppls will know that something changed. and i hope they will open all other greyed out settings as well. maybe we can tune the controllers even more to match our motors.

 

[izeman]

i'm starting to hate it
changed MAX_FRE to 1000, did auto_config again, set max rpm to 3000, and NOTHING. the motor behaves exactly the same as before.
so there must be something else. eg. riding it in first gear it pulls very powerful up to 32km/h in no time and then settles down to 26km/h. so it needs the power to accelerate to 32km/h (this can't be explained with an overshoot cause by intertia of speeding up a lifted wheel). and it can for sure keep that speed. but something reduces rpm again. i hope fany is joining in some time soon ....

 

[SplinterOz]

Ok so the Colossus 14kw motor using a KSL-D (400amp peak model) in a midget Go Kart.
20S of lipo around 10ah.

[youtube]VkqMabJUvOg[/youtube]

 

[izeman]

f*ck! i WANT one this looks like some serious fun! how does a 10Ah lipo pack handle 400A? this pack must be freaking hot.
i tested a 10Ah (only 25C hobbyking pack) with 3000W lately and it got quite warm very fast.

 

[riba2233]

400 A is only motor current, battery current could be and probably is lot smaller.

 

[izeman]

400 A is only motor current, battery current could be and probably is lot smaller.

lol. sure. but even 400A phase current probably means ~100A to 150A battery current. but on the other hand, 400A is MAX. splinteroz didn't say how they set the controller up.
i just dropped fany an email to drop in. china is celebrating it's national day on oct-1st, followed by a week of celebration.

 

[SplinterOz]

Controller is set to all max. so 400amp peak phase current.
This is not sustainable on that motor, but fun for short bursts.

The lipo pack is supposed to be good for 25C so 250amp battery peak and these were "direct sourced cells" and seemed to be rated accurately.
BTW none of us were game to look at the Cycle Analyst to actually read the amps.


And yes it is SOOOO fun!!

For the people willing to look at Facebook https://www.facebook.com/tony.castley

 

[fany]

In this case it could be inertia, I'm not sure...

Maybe fany should write here if these controllers can do field weakening. If they do, I would be very interested

Sorry,we don't provide the field weakening control option to increase the motor speed.Because the field weakening is related with different motors.
And it only can increase the speed a little bit.It is not too much.

 

[fany]

i did some more testing and comparing to the old trapezoid kelly controller.

as you can see the old controller runs at a higher rpm. and is around 15-20% faster at no load speed. changing max rpm, Kp and Ki parameter, Err rpm or anything else doesn't change a bit. and i have no idea else what to do. there is just nothing that you could adjust.

As the same goes,the max speed of KLS is lower than KEB or KBS.The speed is less about 15-20% than trapezoidal controller.

Have you already got the special user program from one of our customers?

 

[riba2233]

In this case it could be inertia, I'm not sure...

Maybe fany should write here if these controllers can do field weakening. If they do, I would be very interested

Sorry,we don't provide the field weakening control option to increase the motor speed.Because the field weakening is related with different motors.
And it only can increase the speed a little bit.It is not too much.

You are wrong Field weakening can be done and is done with BLDC motors, it's used widely in productions electric vehicles, and in electric cars. Also, it can gain a lot of speed, even up to 4 times the original speed. So maybe you should speak with your engineers and see what do they have to say. Because your competition has that option (sabvoton, adaptto, sevcon, ASI...)

 

[fany]

ERPM is electrical rotations per minute. The wheel might rotate at 2000 RPM or something like that, but the motor is performing an ERPM much higher. ERPM relates to two other things so far as I know, pole count and reduction ratio. Perhaps a mac has a reduction ratio of 5 and 16 poles. I could be completely incorrect here, but how I think approximate ERPM is calculated is RPM x reduction ration x pole pairs. So in your case, perhaps it's 2000 x 5 x 8 which tells us 80,000 ERPM.

In summation, search.php?keywords=ERPM&t=71942&sf=msgonly Per searching ERPM in this kelly FOC thread it looks like the ERPM for their KLS controllers is a maximum of around 70,000, which may explain your issue. Perhaps find the motors ERPM and see if it's indeed below 70,000.

Actually this calculation should be correct if the gearbox is inside the motor.
Usually MAC motor got the 5:1 inside the motor.
So the ERPM=2000*5*8=80000.

 

[izeman]

ERPM is electrical rotations per minute. The wheel might rotate at 2000 RPM or something like that, but the motor is performing an ERPM much higher. ERPM relates to two other things so far as I know, pole count and reduction ratio. Perhaps a mac has a reduction ratio of 5 and 16 poles. I could be completely incorrect here, but how I think approximate ERPM is calculated is RPM x reduction ration x pole pairs. So in your case, perhaps it's 2000 x 5 x 8 which tells us 80,000 ERPM.

In summation, search.php?keywords=ERPM&t=71942&sf=msgonly Per searching ERPM in this kelly FOC thread it looks like the ERPM for their KLS controllers is a maximum of around 70,000, which may explain your issue. Perhaps find the motors ERPM and see if it's indeed below 70,000.

Actually this calculation should be correct if the gearbox is inside the motor.
Usually MAC motor got the 5:1 inside the motor.
So the ERPM=2000*5*8=80000.

we are talking about 2.000rpm MOTOR and not axle. so the reduction can be taken out of the equation.
and it's 32 poles, so 16 pole pairs, and the equation should be: 2.000 * 16 = 32.000. so far below your stated 70kerpm. correct?

 

[izeman]

my MAIN question is WHY DOES IT REDUCE rpm after it reaches the top rpm??? it pully up to let's say 2.200 and then reduces the speed to 1.950 again. it also reduces power output almost completely. so speed goes up to 32km/h, then after a 2-3s drops to 26km/h. power drops to 150w, speed drops to 25km/h, controller boost power of 2.500w, speed goes to 26km/h, and so on. why doesn't it reach 32km/h again and stays there, providing some hundred watts needed for maintaining speed?

 

[madin88]

2000 motor RPM x 16 pole pairs = 32000 ERPM
correct!

Fany, you really should talk with your enginieers about field weakening. Without this function it is not as attractive as FOC controllers from other manufactures (at least for most members of this board).

 

[izeman]

first of all i'd like to make my motor run at it's normal speed

 

[madin88]

first of all i'd like to make my motor run at it's normal speed

if we want to have same speed with FOC controllers, we necessarily need to make use of field weakening. without it will have 15-20% less rpm.
i believe the reason behind is RMS phase voltage of a sine waveform is lower as RMS voltage of a square waveform. less volt = less speed.

 

[izeman]

first of all i'd like to make my motor run at it's normal speed

if we want to have same speed with FOC controllers, we necessarily need to make use of field weakening. without it will have 15-20% less rpm.
i believe the reason behind is RMS phase voltage of a sine waveform is lower as RMS voltage of a square waveform. less volt = less speed.

but how would you explain that the bike accelerates to a much higher speed than it has no-load-speed with lifted wheel?
no load speed is 26km/h, it accelerates to 32km/h and then falls to 26km/h again. so the controller seems capable to drive the motor much faster, it just doesn't do it over a longer time.

 

[riba2233]

Maybe the controller does field weakening "by accident" for short time until some loop settles down? IDK lol :lol:

 

[izeman]

who knows. i hope fany has forwarded my questions to engineering. the MAC is a not too special motor, so one could assume that the controller has been tested with a similar motor.