Paul & Sabrina's cheap DIY 144v motor controller

[jyanof]

Originally Posted by Darxus

No. That's why I said split the battery pack up into six packs connected to the controller. So each bank of mosfets only gets one sixth of the current.

hm, I must not understand, then. Whichever mosfets are on at the same time will see the full motor current. This design used a set of 10 to achieve a 500A output.

[Darxus]

Originally Posted by jyanof

hm, I must not understand, then. Whichever mosfets are on at the same time will see the full motor current. This design used a set of 10 to achieve a 500A output.

For round numbers, pretend the controller is using 12 mosfets to do 600A output.

Instead of hooking 600A of batteries up to a single bank of mosfets, break the batteries up into 6 isolated packs of 100A. Hook each 100A pack of batteries up to a separate bank of mosfets. Evenly stagger the triggering of each of the six banks of mosfets. So each bank only gets 100A, and you're still outputting the full 600A (max). Did that make sense?

[jyanof]

Originally Posted by Darxus

Instead of hooking 600A of batteries up to a single bank of mosfets, break the batteries up into 6 isolated packs of 100A. Hook each 100A pack of batteries up to a separate bank of mosfets. Evenly stagger the triggering of each of the six banks of mosfets. So each bank only gets 100A, and you're still outputting the full 600A (max). Did that make sense?

Ok, I understand the setup. Keep in mind that the controller shares B+ with the M+ terminal. They are hardwired together, eliminating any isolation from multiple packs. There could still be separate banks of mosfets to switch connections from M- to the B- terminal of different packs, but unfortunately, the output would not add together as proposed.

When one bank switches on, it carries the motor current which will rise slightly until it switches off. At that point, the diodes carry the current until the next bank switches on. Each bank will see the full motor current whenever it is switched on. Thus, each bank would have to be sized for the full motor current.

In order for the current to be shared among all of the mosfets, they need to be switched on simultaneously.

[Georgia Tech]

hey guys whats the story on the Synchronous controller? The one that can handel 1000 amps..

[Darxus]

Originally Posted by jyanof

...the output would not add together as proposed.

When one bank switches on, it carries the motor current which will rise slightly until it switches off. At that point, the diodes carry the current until the next bank switches on.

No I'm not talking about switching the first bank off before the second bank switches on, just evenly staggering their cycles.

Originally Posted by jyanof

In order for the current to be shared among all of the mosfets, they need to be switched on simultaneously.

As they would be, when each bank is at 100% duty cycle.

[Greg Fordyce]

Hi Darxus,

The controller and motor together form a Buck converter. In this image from the wikipedia article the motor is the load (on the right) and the inductor is the motor's field coils. The controller is the switch, diode and capacitor.


The buck converter circuit will drop the pack voltage down to a lower motor voltage automatically. This voltage will be mainly determined by motor rpm, as the motors rpm increases so does its back emf which will increase the voltage required to maintain the higher rpm.

I think what you are describing is a multiphase buck converter shown in the same article.



There are several advantages to this according to the article, less current ripple and splitting the load current. But notice the inductors, one for each phase, so we can't get away with using the field coils in the motor, we need to add external inductors, however it does look like you could mount the caps on the motor. Now it may be possible to get around this problem in some clever way, but looking at the rest of the circuit makes me think that if you are going to all this trouble you may as well design a AC controller.

Greg

[Darxus]

Originally Posted by Greg Fordyce

I think what you are describing is a multiphase buck converter shown in the same article.

Yay, you understood me! Yes, that sounds exactly like what I was thinking.

Are you sure there must be a separate inductor for each phase, and that wasn't just the way they drew it or it's most commonly implemented when you're actually making a buck converter and not a motor controllor?

We could use a separate motor for each phase

I need to start ordering parts to blow up.

[Darxus]

It also just finally sunk in that an increase in switching frequency causes a decrease in loss due to switching ripple. Which I guess more than balances out the increased switching losses from switching the mosfets more often? I wonder where that breaks even, or if it does.

[apowers]

Originally Posted by Darxus

For round numbers, pretend the controller is using 12 mosfets to do 600A output.

Instead of hooking 600A of batteries up to a single bank of mosfets, break the batteries up into 6 isolated packs of 100A. Hook each 100A pack of batteries up to a separate bank of mosfets. Evenly stagger the triggering of each of the six banks of mosfets. So each bank only gets 100A, and you're still outputting the full 600A (max). Did that make sense?

Does the motor ever see 600 amps? Or is it 100 amps 6 times per stagger cycle.

[Darxus]

Originally Posted by apowers

Does the motor ever see 600 amps? Or is it 100 amps 6 times per stagger cycle.

Staggered, not separate. So when they're at 100% duty cycle, they're all on all the time, not 1/6th of the time each.