Kelly vs typical brushless controllers

[John in CR]

I've read bits and pieces here about the Kelly controllers working differently. Am I understanding correctly that the throttle has a different affect? For example with the controller I have, when I'm on a hill 50% throttle is the same as 100% throttle. Do the Kelly's work differently and give you the effect more like the gas pedal on a car, ie more throttle = more current?

Is there a way to have a current boost button with typical controllers? Could 2 identical controllers be wired to the same motor in parallel?

John

 

[hejo]

I bought two Kelly 50A expecting to run them in torque mode.
After lots of experimenting / firmware updates, my conclusion is that there is not so much difference.
I expected to have full control of current in torque mode , means :
If current limit is set to 30A (internal paramater in config), it should never exceed 30A, but it does!
If throttle is 2,5V (50%) it should send 15A into motor (of course under load)
if throttle is 1V (20%) it should send only 6A into motor...
It does not, instead it operates very similar to normal speed mode : 2,5V Throttle, causes 40A accelleration.
I have to admit that my 50A controller might be way too powerfull for my application (1-15A) which could be a reason for poor controller resolution.

 

[frodus]

Could 2 identical controllers be wired to the same motor in parallel?

John

nope, the PWM can't be sync'd between the two

 

[John in CR]

Could 2 identical controllers be wired to the same motor in parallel?

John

nope, the PWM can't be sync'd between the two

They'd get the same hall signals and magnets are in the same position relative to the stator, so why not? I would have thought the issue would have been current from one controller shooting up the phase wires to the other, or maybe just a rough start, but at least once going would the phases have to fire in sync?

John

 

[frodus]

Sinwaves/trapwaves for AC or brushless controllers are created via PWM and PFM at a high frequency. Its not just an analog sinewave, its created digitally. You have to have ALL of the FET's inside turning on and off at the same time. If they're not sync'd, they won't work. AFAIK there is no way you can sync between two DIFFERENT controllers at the switching frequencies we're talking about. If should be done all one one board with matched impedance circuit board runs from the driver to the gates of the FET's.

Now, if you have 2 motors and 1 controller, and the halls are on one motor, and the two motors are mechanically coupled and verified that the windings are in the exact same position, it will work IF AND ONLY IF the controller can compensate for the reduced resistance in the coils if they're in parallel.

 

[John in CR]

Sinwaves/trapwaves for AC or brushless controllers are created via PWM and PFM at a high frequency. Its not just an analog sinewave, its created digitally. You have to have ALL of the FET's inside turning on and off at the same time. If they're not sync'd, they won't work. AFAIK there is no way you can sync between two DIFFERENT controllers at the switching frequencies we're talking about. If should be done all one one board with matched impedance circuit board runs from the driver to the gates of the FET's.

Now, if you have 2 motors and 1 controller, and the halls are on one motor, and the two motors are mechanically coupled and verified that the windings are in the exact same position, it will work IF AND ONLY IF the controller can compensate for the reduced resistance in the coils if they're in parallel.

If they have to be sync'd then it must be something coming back up the phase wires from the other controllers that would be messing things up. Aren't the high frequency pulses just to emulate a different voltage? Let's say the motor is moving by the position where phase wire A is supposed to get pulses of + current. Both controllers will want to try to send these pulses to the same phase at almost exactly the same time. Why do those pulses have to be exactly in sync? The motor shouldn't care, since it's fooled into thinking that it's getting a single sinusoidal pulse.

It would be nice if it worked and somehow the motor was fooled into behaving like it was getting double the voltage.

John

 

[fechter]

Could 2 identical controllers be wired to the same motor in parallel?

John

nope, the PWM can't be sync'd between the two

They'd get the same hall signals and magnets are in the same position relative to the stator, so why not? I would have thought the issue would have been current from one controller shooting up the phase wires to the other, or maybe just a rough start, but at least once going would the phases have to fire in sync?

John

In theory, that seems like it should work. I know a guy that tried it a couple of times, and it always resulted in controller failure. I think the basic problem is that even though the hall signals are sync'd, the switching delay times between the two controllers were just enough different to cause shoot through (both high and low side on at the same time), resulting in a FET short and letting the smoke out. The other problem is the PWM would not be sync'd, so would tend to cause uneven current sharing and nasty transient spikes.

If you tapped into the feed for the gate drivers one one, and fed that to the input of the gate drivers on the other one, you might get them to sync up OK.

 

[ZapPat]

I would think that using 6 inductors (coils) should let you parallel two controllers. This would have the effect of buffering the high frequency PWM signals of each controller, which is what will be giving you problems. The slow commutation sequence should not be a problem as others have pointed out as long as the controllers are identical and all.

However this would not make the motor "see" any more voltage, but instead would just raise the maximum current capability.

Personnaly I would not mess with this unless I was using my own controller, since just buying a bigger Amp rating controller will do the same thing likely for less money and much less hassle.

 

[John in CR]

Thanks guys,

Fetcher's statement puts the nail in that coffin, ie since others tried and fried.

Pat I've been asking Knuckles to sell me some more controllers since November, so buying something bigger isn't always as easy as you'd think. Plus I can get motor+controller+wheel+tire+DC/DC converter for less than I can buy the Kelly I need. I guess I have to bite the bullet and try souping some of mine up a bit.
I'm on to the voltage doubler idea that this thread put in my mind.

John

 

[frodus]

Sinwaves/trapwaves for AC or brushless controllers are created via PWM and PFM at a high frequency. Its not just an analog sinewave, its created digitally. You have to have ALL of the FET's inside turning on and off at the same time. If they're not sync'd, they won't work. AFAIK there is no way you can sync between two DIFFERENT controllers at the switching frequencies we're talking about. If should be done all one one board with matched impedance circuit board runs from the driver to the gates of the FET's.

Now, if you have 2 motors and 1 controller, and the halls are on one motor, and the two motors are mechanically coupled and verified that the windings are in the exact same position, it will work IF AND ONLY IF the controller can compensate for the reduced resistance in the coils if they're in parallel.

If they have to be sync'd then it must be something coming back up the phase wires from the other controllers that would be messing things up. Aren't the high frequency pulses just to emulate a different voltage? Let's say the motor is moving by the position where phase wire A is supposed to get pulses of + current. Both controllers will want to try to send these pulses to the same phase at almost exactly the same time. Why do those pulses have to be exactly in sync? The motor shouldn't care, since it's fooled into thinking that it's getting a single sinusoidal pulse.

It would be nice if it worked and somehow the motor was fooled into behaving like it was getting double the voltage.

John

Its like this. Imagine two hoses coming from a faucet going to a Y, then a larger hose connected to the end which is the output. If one faucet is pulsing "ON" and the other is off, you get only the output of ONE of them. If you get BOTH of them "ON" then the output is double what it is normally. You only get double output if you have both ON at the same time. Since BLDC/AC controllers are are a sinusoidal/trap wave that crosses zero, if one is OFF, then you'll actually potentially shove current BACK through the other controller as regen. Its not going to work.... The FETS HAVE to be on at the same time, off at the same time. If they don't, you're going to have many issues, and "regen" and half current are just a couple of them. You can actually start blowing fets, which is one of the big reasons homebuilt controllers have lots of problems... FET timing.