Paul & Sabrina's cheap DIY 144v motor controller

[MPaulHolmes]

I'll probably test it this weekend. But I'm kind of on a building roll right now, and it always takes a couple hours to get into the "coding roll". Maybe I'll finish assembling them first. I still need some dang blasted enclosures. dang nab it.

[wakinyantanka]

Originally Posted by DJBecker

That configuration wouldn't be balanced.

The layout should be based around equal path length. The current shouldn't have an easier path to some devices.

This requires that the motor connect to opposite ends of the controller. Otherwise the freewheel current will mostly prefer the "close" diode. When that burns out because it's carrying all of the current, the current will prefer the next device and burn that out in turn.

That is if the device burns out. Sometimes diodes fail by shorting out. When the controller only has a motor current sensor, it can't detect the case when a freewheel device shorts out. Instead it does the opposite of protecting itself. Since most of the battery current flows through the short instead of the motor, it increases the drive PWM in response. And keeps increasing it until every push-side device is destroyed.

(We've replaced all of the push-side MOSFETs only to have them immediately blow up in a spectacular way with the first test. Twice. We now remove and test every diode after any power section failure.)

So if I understand this correctly, make sure the total length of the three busbars is the same to avoid stress on the mosfets and diodes. And by total I mean that B- and M- equal the length of B+ when placed in relation to B+.

[MPaulHolmes]

I think it's more the flow of current shouldn't have a preferred path to any one component. When the mosfets are off, the positive current goes in from M- bar, has it's pick of any of the 10 diodes to go through, and goes back out through B+. There's no reason one diode would get more current than the next. But with M- and B+ on the same side, the shortest route back out of the controller would be through the closest diodes to the M- and B+ bar ends with the mounting holes on them, so those diodes could get more current, but in practice it seems not to matter much.

Now, for the mosfets, the current comes in from M-, and exits B-. With M- and B- on opposite ends, there's no preferred route for the positive current to go. eenie meenie miney mo. I choose mosfet # fo. haha. But if M- and B- are on the same side, the mosfets nearest to the end might get more current?

[wakinyantanka]

Originally Posted by MPaulHolmes

I think it's more the flow of current shouldn't have a preferred path to any one component. When the mosfets are off, the positive current goes in from M- bar, has it's pick of any of the 10 diodes to go through, and goes back out through B+. There's no reason one diode would get more current than the next. But with M- and B+ on the same side, the shortest route back out of the controller would be through the closest diodes to the M- and B+ bar ends with the mounting holes on them, so those diodes could get more current, but in practice it seems not to matter much.

Now, for the mosfets, the comes in from M-, and exits B-. With M- and B- on opposite ends, there's no preferred route for the positive current to go. eenie meenie miney mo. I choose mosfet # fo. haha. But if M- and B- are on the same side, the mosfets nearest to the end might get more current?

Ok I'm kinda lost. Paul you've seen my design is this something I should be concerned with? I don't know if you remember but I have B- and B+ on one end and B+, (or M+), and M- on the other. The length of B+ is equaled by M- and B-.

[mrbigh]

Originally Posted by wakinyantanka

Ok I'm kinda lost. Paul you've seen my design is this something I should be concerned with? I don't know if you remember but I have B- and B+ on one end and B+, (or M+), and M- on the other. The length of B+ is equaled by M- and B-.

And that configuration is just fine and will work as a champ.

[harlequin2]

Don't worry about it! The current flows down the busbars at about .4 nanoseconds per metre! And that means there is no measurable difference in which device gets what first. The turn-on and turn off times are orders of magnitude greater and completely swamp any variations in busbar length.

[MPaulHolmes]

Originally Posted by harlequin2

The current flows down the busbars at about .4 nanoseconds per metre!

I could run that fast if I wanted to.

[harlequin2]

Originally Posted by MPaulHolmes

I could run that fast if I wanted to.

I bet you'd have to have something large and ugly chasing you.
BTW, pcbs arrived o Friday, thanks.

[DJBecker]

Originally Posted by harlequin2

Don't worry about it! The current flows down the busbars at about .4 nanoseconds per metre! And that means there is no measurable difference in which device gets what first. The turn-on and turn off times are orders of magnitude greater and completely swamp any variations in busbar length.

It's not the propagation delay (which is considerably longer than your estimate, but still inconsequential). It's the resistance, and a little bit the inductive loop area. At a few hundred amps, even the trivial additional resistance for the extra few inches of bus bar makes a big difference in how balanced the current is.

[harlequin2]

Originally Posted by DJBecker

It's not the propagation delay (which is considerably longer than your estimate, but still inconsequential). It's the resistance, and a little bit the inductive loop area. At a few hundred amps, even the trivial additional resistance for the extra few inches of bus bar makes a big difference in how balanced the current is.

Oh yes, how much longer than my "estimate"? Have you worked out the LR? As I said before, theory is a wonderful thing!
I have tried to measure some of these things and failed - too short, so ignore them.