DC Motor Driver Regenerative Braking

The concept is that you use one of the appropriate switches and one clamp diode to short the motor, briefly, using its self generated EMF to drive the current up (ramp rate set by the EMF and motor inductance) to the magnitude that produces the desired braking torque, then open the switch and let the motor inductance pump that current into the supply through the clamp diodes, like a flyback power supply would. This works most reliably if you can monitor the current during the process. Letting the current wind all the way to zero reduces the switching losses for the next shorting part of the cycle, but lowers the average braking torque relative to the peak current, so the peaks have to be higher to get the same average breaking.

Repeat as necessary.

getting

Thanks, John. It's beginning to make sense now. I couldn't see how the back EMF, which is usually lower than the supply voltage, could push current back into the battery. So, in effect, for example, if the BL MOSFET is on and I'm PWMing the AH MOSFET to control speed, then I need to turn on the AL MOSFET for part of the time that the AH MOSFET is turned off. Is that right?

When you mention monitoring the current during the process, do you mean the current through the motor while the AL MOSFET is turned on, (in the example above)?

It is important that braking occurs automatically when the PWM duty-cycle is decreased. I assume that the above process should be carried out the whole time the motor is running, with the low MOSFET turning on for a short period whenever the high-side MOSFET is turned off. Is this correct?

... Humphrey

Before you get happy about the concept think about where the energy is going and how you might switch the switches and what might give you the best efficiency..... and why.

AB CD

You might have A+D on for getting to full forward and then you might have B+C for going full reverse. In between you might have a bit of C+D or A+D to freewheel things a bit and avoid parasitic diode conduction and it will all have an effect on the loop you stick around things to do it.....

But, you're biggest bugger is you will shunt the energy from the motor, and its load, back into the supply. You might spend a lot of time being dead clev about how you switch things on and off but if the supply can't cope with the motor/load dump then you'll be bolloxed.

DNA And then think about where the energy is going again

Hey, if you look at it from a conservation of energy standpoint...

Crank up the h-bridge duty cycle to some goodly amount and let the motor rev up to speed, and acquire a bunch of angular momentum sort of stored energy. Now reduce the bridge-drive duty cycle. The h-bridge average output voltage drops, so the motor back emf must be dumping energy *into* the h-bridge, and the motor is being braked. Since (as we all know) an h-bridge is 100% efficient, the energy being extracted from the motor has to go somewhere, so it goes into the power supply, where it came from in the first place.

So dynamic braking is sort of automagically free... just scrunch down the drive duty cycle. Of course, the switches in the bridge must be able to conduct in all directions, so they should be fets.

Did I get that right?

John

going

and

The supply is a 12V, 24Ah lead-acid battery, so the energy will just replace some of its charge. I'll probably use a PIC16F876 to run the show. Previously, I used a HIP4080A as the bridge-driver. It doesn't allow separate switching of the low-side MOSFETs, so I'll probably use the HIP4081A this time. Still need to double-check the datasheet to see if I can accomplish what I want. I'd like to avoid the extra complexity of current monitoring, so timing is the next major issue.

... Humphrey

Assuming your power supply can sink as well as source energy, or store the excess somewhere, then yes.

Thank you for the clarification.

DNA

I think so. Note that this produces only dynamic braking, not reverse torque at a standstill. You can also drive the motor current up in the braking direction by applying the supply voltage in series with the generated EMF, to be able to push up the braking current, even down to zero speed. This is essentially 4 quadrant operation, where the torque is independent of the rotation direction. This is quite a bit more versatile then just dynamic braking, where you rely on motor speed (generated EMF) only, to drive the braking current.

Having a current monitor that tells the control circuit the magnitude and direction of the motor current, regardless of which switches and or diodes are conducting, is very handy, when you need to control torque as well as speed, since torque is proportional to current (for a fixed field excitation, or permanent magnet field), to a first approximation.

I think you would need a current feedback scheme to control the pulse widths to make the braking smooth and variable, as desired.

...

As long as we're talking about fantasy regenerative braking systems, how about feel feedback? i.e., the harder you press the pedal, the harder you're braking, and the harder the pedal pushes back. With ordinary hydraulic brakes, you can feel with your foot how hard you're pressing the pad against the rotor, or is the deceleration enough feel feedback? I'd hate to have a pedal with no feel feedback if I were in traffic and saw brake lights up ahead - you have to brake hard enough not to rear- end the guy in front of you, but not so hard that you get rear-ended by the guy behind you.

Thanks! Rich

Since (from my understanding) regenerative braking does not wear out like abrasive pads, would a block of rubber[1] under the brake pedal not have the same effect? Just thinking about all that energy wasted in pushing the pedal back.

Mark

It is noteworthy that F1 brake "pedals" are just immobile (or nearly) force sensors. The deceleration is already *too much* feedback :-)

What makes you think that this is a fantasy braking system? I've got it working well. The brake can be applied manually, but also it is applied automatically at any time that the (scaled) motor feedback voltage is higher than the control voltage. The brake duty-cycle is varied by the uC depending on the difference between the two voltages.

... Humphrey

message

Well it really takes a more, the rest of the current paths need to be in place. Transistors usually do not work so well inverted (D-S or C-E swapped), and have to be driven for that mode.

Gosh, who was it that recently posted...

"Of course, the switches in the bridge must be able to conduct in all directions, so they should be fets."

John

In a way, the gate is acting like an amp, but that's just a side-effect of it doing its primary function - the logic operation - It sees your 1 from the osc on A, and the result is that there will be selection processes being imposed upon us In the arrogance of some humans they think that the standards you mentioned are the only ones. As you change temperature or drive or just wait a few us for it to gain an credibility. * Most of what people chat about is just that - chat. It's like when Cheney lobbed Publicly for torturing innocent people as a bushite enemy's pleasure.

A crime that Reagan DID MAKE LAW that STANDS STILL in Johnny's America, to now have demanded life term prison sentences against ALL bushites who would fight as such, and death to those who are outgunned are in a critical balance They also typically top out at 1000V. Do elaborate.