Sunday, December 2, 2012

The PWM Field Controller Is Installed

I was finally able to wire in my "proper" field control system. This system basically consists of two components:
-72V/100A Kelly motor controller (meant for small bike/scooter motors)
-Amp transducer to measure current in the armature negative cable

The LEM amp transducer that I bought is typical of most units like this: an output signal of 2.5V with zero current going through it. If current goes one way, the signal voltage increases and if current goes the other way the signal voltage decreases. I did some testing with a carbon pile to put a load on a few batteries and found that at 450A, the signal voltage is 3.9V.

This signal voltage is what I feed the Kelly motor controller to tell it how much current to give my field windings. Luckily, Kelly has provisions in the controller software to narrow down the signal voltage range to whatever you want, so I didn't have to change this signal from a 2.5-3.7V one to a 0-5V one. That would have been above my head anyway!

Here is a screenshot of the parameters I am talking about:
I have adjusted "Throttle Effective Starting Position" so that when the key is on and things are "idle", the field is using 8 amps. Since the field windings are exactly 1 ohm, this takes 8V to accomplish for a total of 64 watts. Pretty insignificant.

Why not not cut field current altogether when off the throttle? Because I have found that when freewheeling down a hill at high motor RPM and you turn the field on, a voltage is generated in the armature that confuses my Alltrax controller causing it to go brain dead for a few seconds. Keeping a small amount of current going through the field prevents this from happening.

I am still tinkering with "Throttle effective ending position" so that it is a linear relationship between field current and armature current instead of reaching max field current too early, when armature current isn't high enough yet.

Varying field current proportional to armature current has made a dramatic effect on how the throttle feels. With steady field current, the throttle felt exactly like a farm tractor, where 50% throttle = 50% motor speed. It made driving very awkward, especially when trying not to accelerate at full power. Now it feels much more like a car, where throttle position controls torque output. Much easier to drive.

Many EV enthusiasts advise to stay away from Kelly products for reliability reasons, but since I am only using this controller to less than 50%, I figure I should be okay. Makes sense to me, anyway.

As happy as this all sounds, there is a problem. During my first drive to work I found that at higher RPM and light throttle the motor developed a high speed surge which shook the entire car. It was sort of like turning the throttle on and off ten times a second.

Here's is my best attempt of an explanation of this strange phenomenon. At a given motor speed, when you increase field current, back EMF (the motor's resistance to accept current) increases, which causes armature current to drop. This drop, sensed by the amp transducer, tells the field controller to lower field current. This in turn lowers back EMF, causing the armature to increase. This oscillation happened many times per second.

What I did to remedy this was change the "Throttle up/down rate", which is basically a time delay for the field controller's response to its throttle input.
I originally had it set to 1, which was really bad. I upped it to 3, 6 and then finally 9, where it is much better. It still surges a bit at part throttle but at least it doesn't oscillate and vibrate the whole car.

There you have it. One more bit done. This weekend I am working on improving the heater now that it is getting cold outside.

No comments:

Post a Comment