Home Brew Challenge for Electric Geniuses (make a DIY electric motor controller)

[Coyote X]

If you have built a PWM controller before this is basically the same. The only difference is instead of using a single or a pair of mosfets you are wiring a lot of them in parallel. There are some good 35 amp mosfets out so with those wiring up 20 of them would make it able to handle pretty much anything the batteries can do without the batteries sustaining damage. I would say group the mosfets into 3 or 4 and wire each group with 8ga wire on the load carrying parts. Then keeping the wires short and neat put them all together and attach to the normal heavy current wire with a large solid copper screw clamp. The mosfets will need heat sinks to keep cool but it should be easy to find an old car stereo amp and use for the housing.

If you don't plan on using a transmission for reverse you could wire in maybe 10 extra mosfets for reverse, Just limit them to 25% duty cycle or something and they will probably be fine for the limited use they will see.

There is nothing really magic in a controller it just takes a bit of care making sure the mosfets and power wires can handle the current loads of the motor. Other than that it is the same as any small remote control car speed controller. With some of the components out now though it should be an easy thing to build really.

[Coyote X]

http://search.digikey.com/scripts/Dk...=VMO650-01F-ND

ok skip the big mosfet bank and use that one. 690A 100V are it's ratings

[Shifty]

I have built some PWM controllers that used parallel MOSFETs but this can get you in to trouble rather quickly. Timing is absolutely critical. If one FET turns on just a few microseconds before the rest it will experience a large current surge through it. Also there is the issue of driving all that gate capacitance. You need to quickly get the FETs from the off state to the saturated fully on state. Power loss and therefore heat generation are real problems when in between states. My bet is that using an IGBT may be the best way to go for the real large stuff. I have a broken IGBT to show students that can supply 300 amps.

[JohnnyV]

For me, doing the control part of it would be relatively easy. However, I have no experience designing circuits for medium/high voltage equipment. If someone has experience with that, I think we could put our heads together and bang out a solution.

Like Shifty said, the biggest concern will be the power MOSFETs and getting them to switch quickly enough, if if you're using multiple ones, in synch with one another.

I would like to see a requirement that the control method use feedback from the motor so you can verify you are running at the desired speed. This also will help with efficiency since you won't have oscillation in the motor speed assuming you set your parameters correctly. The easiest way to implement this is through a PID controller, which, with some effort, can be accomplished via a PIC microcontroller (or any other decent micro).

If going the feedback route, we would want the software for the micro to be easily adjustable for different motors - all motors have inherent parameters that must be taken into account when developing a control system for them.

[83VillageRepair]

This group might interest you. Its an open source motor controlller that got started during the battlebot days.

http://tech.groups.yahoo.com/group/osmc/

[3wheeled]

Originally Posted by JohnnyV

For me, doing the control part of it would be relatively easy. However, I have no experience designing circuits for medium/high voltage equipment. If someone has experience with that, I think we could put our heads together and bang out a solution.

Like Shifty said, the biggest concern will be the power MOSFETs and getting them to switch quickly enough, if if you're using multiple ones, in synch with one another.

I would like to see a requirement that the control method use feedback from the motor so you can verify you are running at the desired speed. This also will help with efficiency since you won't have oscillation in the motor speed assuming you set your parameters correctly. The easiest way to implement this is through a PID controller, which, with some effort, can be accomplished via a PIC microcontroller (or any other decent micro).

If going the feedback route, we would want the software for the micro to be easily adjustable for different motors - all motors have inherent parameters that must be taken into account when developing a control system for them.

Hi,

I think what I would do instead of using the speed as the error signal to the PID controller would be to use the current to the motor as the error signal. A motor is actually a current device, not a voltage device. That is why the PWM works so well in controlling the speed. The output FETs switch on and off at varing dutycycles depending on the desired average current through the motor windings. Think about the way a car works. The accelerator sets the fuel going to the engine not the speed. The driver adjusts the speed depending on driving conditions etc.

As far as output FET switching times...
I have designed Complimentary FET drivers that drive the gates of the power FETs. Made a HUGE difference in the heat dissipated in the power FETs.
The idea here is to turn the power FETs on and off as quickly as possible. Once the FETs are full on or full off the power dissipation is nill. It is the transition switching the kills you. Hence, just design a highspeed driver before the Power FETs

If you parallel the output FETs, you will most likely have to match them. Otherwise, if one FET has a slightly lower ON resistance than the others, it will tend to pass a higher % of the current which could easily excede the individual device rating and *poof* you end up letting the smoke out. Never do you want to let the smoke out; it is the smoke inside the device that makes the part work - let out the smoke and the part stops working

Thinking about how to build a controller without matched FETs makes my head hurt. That could get REALLY complicated. Although an interesting academic exercise.

Have fun!

Wallace

[jjackstone]

Hi guys,
Here is a link to an interesting controller I found a while back. May be just what you're looking for.
http://www.evhelp.com/Controller_Prototype.htm

[jjackstone]

And here is a link to the build article.
http://powerelectronics.com/mag/711PET22.pdf

[DKaine]

Hey guys, this thread seems to have gone dead for awhile. Has anyone put some work into their controllers? I have a few ideas fluttering around in the ol' brainbox, but mine are a bit old fashioned - Keep in mind I'm no EE, I just have some basic electrical/electronics knowledge to get me by...

What about a mechanical motor controller - a sort of cylinder (similar to the commutator inside the motor) that is geared to the motor driveshaft and spins, while a throttle cable adjusts an armature that sweeps across a varied-width contact... Hm, I'm not so good at explaining these things, but once I get home I'll sketch something up. Sound feasible?

[MetroMPG]

Not only is it feasible, drum contactors were commonly in use in EVs a century ago.

The drums comprised a network of connections that switched the battery pack in a progression of series/parallel steps to provide smooth power/acceleration.

I rode in an antique Milburn EV that had this type of setup.