Step on the Brake!
by Darren Ashby

Imagine careening down a hill on your electric scooter. "Gosh," you think to yourself, "it would be nice to use some of the energy I am wasting to slow this vehicle down." There ought to be a way to make it recharge the batteries. "I'm an engineer" you think, "why don't I go design a regenerative brake." Just such a thought has come into my head and I had been able to ignoreⁱ it quite effectively, until now.

Not too long ago I was asked to design a motor control with a regenerative braking circuit. Having done several controls, but none with regenerative braking, I started by perusing the Internet. I don't follow Star Trek's creed to boldly go where no man has gone before on a whim. That is to say that if someone has been there already I would sure like to know the path he or she took. Once the end of that has been found, I will then venture into the unknown. Several hours of searching were somewhat futile. A simple and concise explanation and possibly a schematic (particularly for a PM DC motor) were all I needed. There were reamsⁱⁱ of information explaining what it does but not much was there showing exactly how it was done. Alas, my effort to find the simple explanation was to no availⁱⁱⁱ . Those who read my column often know that I take such a lack as a personal affront that I must correct. Thus the following is what I have pieced together in my own mind, distilled down to my level of intelligence (the longer I spend in management, the lower this level seems to be...), then ousted to my readers in a form I hope is easy to understand. So after I looked at the best idea since raw toast, and the nice read about the Honda Insight, this is what came out.

No more secrets!

One place I found said that regenerative braking is the well kept secret of motor control. However, when I learned the truth, I think it is just poorly explained. Lets start with the following, a diagram of a simple PWM controller for a DC PM motor.

A PWM is fed into a switch, such as a MOSFET, at a frequency that is high enough to keep current flowing in the inductor inside the motor, not at all unlike a switching power supply. When the PWM shuts off, the current flows through the diode, also known as a freewheeler. But the question that I kept asking myself was how do you get a motor, which is spinning at a lower voltage than the output of the battery, to push current back into the battery?

Let's start with a small change to the circuit above. We will replace the diode with a synchronous switch that goes off when the primary goes on, and visa versa. For the purpose of this discussion we will ignore the fact that the FETs need particular driving methodologies for the high side and the low side of a motor. There is plenty of information like that at such places as www.irf.com .

I had read about this topology many times, it is usually brought up as a way to make your controller more efficient in terms of heat loss, this is because the FET has a significantly lower voltage drop across it than the diode does. I had no idea that it also functions as a regenerative brake, until now. Here is how it works.

A little elaboration

Keep in mind that in the original version of this controller there is a voltage across the motor that depends on the duty cycle of the PWM, but it is referenced to the positive output of the battery, not the negative side. That helped me to keep it in perspective.

Assume we have a 12 V DC battery and there is 6 V DC across the motor. That means you would see an average of 6 V DC from the bottom of the motor to ground. Now let's say you spin the motor faster than 6V, 7V for example. If you keep the same average voltage at the bottom of the motor, then you will have 1 volt extra to dump into the battery. Now this explanation doesn't entirely jive, but I think it will get you in the right frame of thinking. If you follow it to conclusion, you will think that the version above with the diode should also regenerate, but it does not.

Let me elaborate, with the diode version, there is no braking force generated. That comes into play when the diode is replaced by the FET. When the freewheel FET turns on, the voltage generated by the motor is shorted back into itself. This provides the braking force and a current flow in the opposite direction through the motor. Remember the rule of inductors (since there is a decent sized inductor in the motor). Once a current is flowing it doesn't like to stop. So when the high side opens and the low side closes, current is now pushed into the battery and viola, you have regeneration!

The final say

It turns out that regeneration isn't so tough at all. In fact, it is almost a side benefit of making your controller more efficient if you want to look at it that way. Now if there were just some way of making it more than 100% efficient, hmm...

----------------------------

Footnotes

ⁱ I find it very easy to ignore such thoughts when I am playing Nintendo. In fact back in my college days, I had to redo an entire quarter of school (except for one class) due to a severe Nintendo addiction. But we'll save that story for some other time. :)

ⁱⁱ Can you use the word "reams" when referring to the Internet? After all, it isn't really on paper is it? :)

ⁱⁱⁱ If any of you net junkies out there have such a link, please pass it on. There is an awful lot of data out there, but once I get sick of the pop-ups I just stop looking.

Product Engineering Archive