Motor comparisons [CA120 et al.]

[Lebowski]

Maybe you should try to build a controller based on my controller IC ? Arlo got the colossus running with my IC...

 

[crossbreak]

Nice project you've got running there. I have to read further to get into this. Thx for the tip!

 

[circuit]

This is what I would buy from. You think this one would be any better? http://www.devi-motion.com/webshop/controller/p-1/D1000479--hpc300h-motor-controller--72v-300a.html

Take a look here:
http://e-motion.lt/2010/09/17/golden-motor-kokibe/

I tried it with Turnigy CA80100 motor. The controller blew up within 3 minutes with no load. I guess this is due to low motor's inductance.
The construction has a potential (dsPIC + phase current sensors), but actually is a huge mess.
And GM's return policy is not as straight-forward. Be warned.

 

[h0tr0d]

http://www.freeair.cz/freeairen/index.php?option=com_virtuemart&page=shop.browse&category_id=49&Itemid=53&TreeId=32

Controller capable of low inductance motors (I believe) and BLDC's up to 40kW!!!

in a ebike or motorcycle the load is at its highest at 0 rpm

Never did like this "typical" sentence... the load is highest to the controller because at 0 rpm the BEMF is 0V. Since the motor voltage is battery V - BEMF, current "grows" faster at 0 rpm, killing controllers.

Mechanical load is highest at top speed since aero drag and rolling resistance increase with speed.

With any low inductance motor, anything below 100khz PWM freq becomes a disaster...

 

[Martini]

Hello Arlo ,

I pay 5000rmb = +/- 600Euro , motor is 5Kg shaft 15mm diam with high quality magnet and hall sensor inside the motor
+ shipment costs.

Now I am in busuiness trip and I can not test and run it .

I will use A123 prismatic 21S3P + Kelly controller KHB72701C 350 / 700A

WHAT THINK THE PEOPLE OF THE FORUM OF MY CHOICE ?

for order please contact gs@gsmodels.com Cherry will help you

Gianni

 

[circuit]

WHAT THINK THE PEOPLE OF THE FORUM OF MY CHOICE ?

Cells - OK
Kelly - um, well, ok (nothing better available)
Motor - where is that inductance rating?

 

[Arlo1]

This is what I would buy from. You think this one would be any better? http://www.devi-motion.com/webshop/controller/p-1/D1000479--hpc300h-motor-controller--72v-300a.html

Take a look here:
http://e-motion.lt/2010/09/17/golden-motor-kokibe/

I tried it with Turnigy CA80100 motor. The controller blew up within 3 minutes with no load. I guess this is due to low motor's inductance.
The construction has a potential (dsPIC + phase current sensors), but actually is a huge mess.
And GM's return policy is not as straight-forward. Be warned.

I can't belive it only took an 80-100 to kill that controller. Glad I skiped that controller.

I have herd of some of the bad ass motors can also melt down the kelly controllers. Although splinter has a kelly on his Colossus and so far so good.

 

[circuit]

Yes, I've seen many dead kellies. Had two DC models lying around. However we are now using kelly BLDC 400A 144V controller in a small car and, so far, it works.

 

[Martini]

Hi

What about MGM COMPRO ? manual is nice

someone has been used this driver ?

Gianni

 

[circuit]

This one?

http://mgm-compro.com/index.php?tid=brushless-speed-controllers-tmm-40063-3-x-series-32-bit-system

Looks interesting.
However, it is rated at "63V". This is interesting, because it is a standard voltage of electrolytic caps. If they rate their products at maximum, it is not a god sign. Voltages on "63V" bus may go well above that during operation.

EDIT: wow. They took the time to cut-out the background of this picture. Nice!

 

[h0tr0d]

this one with phase current sensing:
http://www.sinusleistungssteller.de/Menue_SLS.html

only 12kW, probably Collossus ready

 

[Arlo1]

this one with phase current sensing:
http://www.sinusleistungssteller.de/Menue_SLS.html

only 12kW, probably Collossus ready

voltage is to low.

 

[crossbreak]

Question is, what kills the FETs? Is it
a too slow feedback of the current control
Loop which leads to a very short current spike or is it the rapid current rise during
PWM cycle? Raising capacity seems to help somewhat, which tells me it's more likely that current control is too slow, allowing pwm duty cycles during startup that kills the fets. Any thoughts here?

 

[Arlo1]

Question is, what kills the FETs? Is it
a too slow feedback of the current control
Loop which leads to a very short current spike or is it the rapid current rise during
PWM cycle? Raising capacity seems to help somewhat, which tells me it's more likely that current control is too slow, allowing pwm duty cycles during startup that kills the fets. Any thoughts here?

Thats a loaded question. Temp, current, voltage, gate voltage, etc.
Read this data sheet then understand it and look at all the parameters if you have any parameter exceeded then the fet dies.

 

[Lebowski]

Question is, what kills the FETs? Is it
a too slow feedback of the current control
Loop which leads to a very short current spike or is it the rapid current rise during
PWM cycle? Raising capacity seems to help somewhat, which tells me it's more likely that current control is too slow, allowing pwm duty cycles during startup that kills the fets. Any thoughts here?

Me thinks it's parasitic inductance which causes ringing. A peak in the ringing then overvolts the FET and blows it. Therefore my preference for snubbers.

 

[Arlo1]

Question is, what kills the FETs? Is it
a too slow feedback of the current control
Loop which leads to a very short current spike or is it the rapid current rise during
PWM cycle? Raising capacity seems to help somewhat, which tells me it's more likely that current control is too slow, allowing pwm duty cycles during startup that kills the fets. Any thoughts here?

Me thinks it's parasitic inductance which causes ringing. A peak in the ringing then overvolts the FET and blows it. Therefore my preference for snubbers.

I think he is asking a blanket question Lebowski. He is not directing the question at my experiments.
As for RC snubbers if you need them then you need them but you should try to design a system to work without them it will be more efficient and cheaper to manufacture.

 

[crossbreak]

I think he is asking a blanket question Lebowski. He is not directing the question at my experiments.

Your're right. I'm trying to understand the fundamentals of designing a special circuit for low induction motors. Using big caps, or use the components well below their rating, is the last thing I would want to try. ATM i'm trying to find out more about multistage FET circuits. Sounds promising IMO.

 

[circuit]

For low inductance motors, I'd go for high frequency (100-300kHz) drive and use three phase LC filter to feed the motor with pure wave.

 

[crossbreak]

That's reasonable. High frequency means also high switching loss. The PWM frequency should be rpm dependent.

 

[John in CR]

The solution is simple. Just get one with another turn or 2 on the windings. It's still the same motor, just easier to drive, and increase the voltage to make up the difference if you need the rpm.

 

[circuit]

The solution is simple. Just get one with another turn or 2 on the windings. It's still the same motor, just easier to drive, and increase the voltage to make up the difference if you need the rpm.

Not so simple. Let's say we increase turn count twice. Inductance increases 4x. So ripple current decreases 4x. May look ok. But RPM decreases 2x and we need to up the voltage twice. Now, ripple current is 2x lower than before. May still look OK. However in 2x voltage system you get 1/2 current. So this gives us the same level of ripple current in % of maximum available current.
And so increasing turn count wont solve anything.

 

[John in CR]

The motor only needs half the current, so the controller lives to see another day instead of going pop goes the weasel.

 

[Arlo1]

The motor only needs half the current, so the controller lives to see another day instead of going pop goes the weasel.

I think there is something to be said for all this. It is true that there needs to be a certain number of winds to make it work nice.

My argument is in simple terms like this.
If adding more turns and upping the voltage is not a good idea...
1 why are the OEMs running such hi voltages??? Zero is over 100v now?
2 why would we not just rewind everything to run on 4.2v and make all our batteries 1s so we can charge them safely?

For my tests I was planning on rewinding for more inductance and upping the voltage until I discovered I can run more voltage with the OEM wind and make more HP... SO Im kinda making it hard on my self but it will be worth it in the end.

 

[circuit]

My argument is in simple terms like this.
If adding more turns and upping the voltage is not a good idea...
1 why are the OEMs running such hi voltages??? Zero is over 100v now?
2 why would we not just rewind everything to run on 4.2v and make all our batteries 1s so we can charge them safely?

For my tests I was planning on rewinding for more inductance and upping the voltage until I discovered I can run more voltage with the OEM wind and make more HP... SO Im kinda making it hard on my self but it will be worth it in the end.

OEMs run relatively high voltages because at these power levels it gets ridiculously difficult to handle high currents... One has to balance between voltage and current to get optimal combo.

 

[crossbreak]

I thought of redetermine the rotomax 80cc to WYE and run it on 74V/20s lipo. I bought an 80100 and redetermined it in WYE for a test setup. It ran fine with a 12Fet KU123 sensorless controller, which is made for 30amps. I only ran "15amps" by cutting one of the two paralleled shunts. My Wattmeter showed almost 25amps on a simulated startup with the "15amps" setting, even in WYE inductance was too low to control current. I abandoned the project cause this motor is just too long to fit in a bike, so I sold it. Now I'm trying a new approach with this rotomax 80cc. The cheapest solution is to buy a 24FET/85amps controller and run it an only 50amps for safety. But this somehow sucks I completely agree with circuit about decisions on reasonable voltages.