Regenerative braking, how?

So I've got a decent motor controller tested and operational and I'd like to begin experimenting with regenerative braking.
I found a circuit that is basic but somewhat resembles the mosfet output of the included circuit. What would I need to add to a circuit like this to gain regen to the battery when the flywheel is spinning?


Image from: http://www.dprg.org/tutorials/2005-11a/index.html

Thanks

HellTriX said:

So I've got a decent motor controller tested and operational and I'd like to begin experimenting with regenerative braking.
I found a circuit that is basic but somewhat resembles the mosfet output of the included circuit. What would I need to add to a circuit like this to gain regen to the battery when the flywheel is spinning?

Image from: http://www.dprg.org/tutorials/2005-11a/index.html

Thanks

Click to expand...

Is it a permanent magnet motor?

Are they using half an H-Brigdge?

http://www.4qd.co.uk/fea/regen.html

Diagrams
http://www.4qdtec.com/pwm-01.html

The circuit you have is only for driving the motor. Here is yet another link about DC motors & regeneration - http://homepages.which.net/~paul.hills/SpeedControl/SpeedControllersBody.html.

Yes, it is permanent magnet.

I been trying to determine if I need a high side to form half bridge to begin
dabbling in regen braking, of If I could achieve the same with some diodes.

Someone pointed out I need diodes hooked up with the motor which I do in my circuit. But as I step up to a larger motor I don't think my four 3 amp diodes will suffice. I'd like to look into a P channel mosfets or something for that function, but they are more expensive. I must weigh the cost vs efficiency between those options. I've been tempted to ignore regen braking altogether since its only a small percentage of total efficiency gain.
Then again, when I step up the project to full scale the gain from regen would be significant enough to attempt so might as well try it out on this small controller.

HellTriX said:

Yes, it is permanent magnet.

I been trying to determine if I need a high side to form half bridge to begin
dabbling in regen braking, of If I could achieve the same with some diodes.

Someone pointed out I need diodes hooked up with the motor which I do in my circuit. But as I step up to a larger motor I don't think my four 3 amp diodes will suffice. I'd like to look into a P channel mosfets or something for that function, but they are more expensive. I must weigh the cost vs efficiency between those options. I've been tempted to ignore regen braking altogether since its only a small percentage of total efficiency gain.
Then again, when I step up the project to full scale the gain from regen would be significant enough to attempt so might as well try it out on this small controller.

Click to expand...

As i understood you dont want only to brake the motor but to charge the battery with this energy right?

yes. charge battery with reverse energy.

I'm thinking I could use a unipolar drive with the mosfet on the high side.
Using a boot strap driver. Then use a parallel sense resistor or shunt for the high side to detect positive of negative current flow. Then use that sense with an op-amp to control the PWM for regenerative braking. I could then set a given amount of regen via PWM or have it increase if it senses more reverse current.

HellTriX said:

yes. charge battery with reverse energy.

I'm thinking I could use a unipolar drive with the mosfet on the high side.
Using a boot strap driver. Then use a parallel sense resistor or shunt for the high side to detect positive of negative current flow. Then use that sense with an op-amp to control the PWM for regenerative braking. I could then set a given amount of regen via PWM or have it increase if it senses more reverse current.

Click to expand...

Use a proper circuit for a battery charger and a relay to connect it to the motor terminals. So when you you want to run the motor you apply a voltage to the relay coil and the PWM you have controls the motor. When you want to stop the motor remove the voltage from the relay coil and the relay connected in the right way will connect the battery charger to the battery charger. Also you have to disconnect the battery from the rest of the circuit when charging it. You can use the same relay if it has enough contacts or use another one.

I'm not sure using relays will work.
This current design is 48 volts and 100 amps.
I plan to parallel a few more mosfet for a 48v and 300-500amp peak.

For regen I don't expect to more then 55v at 90-150amps peak for short duration.

The IRFZ46 will not handle 100 A. At the very least you will need several in parallel and heat sinked. Also, note that your drive spec. is 3 to 18V. The gate on the IRFZ should have at least 10 volts, so you should not try the lower voltages. Finally, as I understand it from the DPRG site, your PWM frequency is fairly low. Nevertheless, you are driving the mosfet though a 10K resistor, which leads to long turn-on times and lots of heating. For what you propose, I suggest you consider a more typical oscillator (or keep the configuration you have , if it is working for you) and a specialized mosfet driver chip. The right chip will be able to take the single PWM you have and create the drive for both top and bottom sides of a half-H bridge. John

As I mentioned in the first post, this was just a close resemblance of my circuit, since I didn't feel like making my exact circuit up for show.
I have several HRF3205s and some IRFB3077PbF's in stock that I plan to use.

I currently have two HRF3205s in parallel just for a unipolar controller.
I haven't even etched my board yet since I would like to implement regen braking before doing so.

Found a few sites that discribe half bridge with regenerative but they are not very well written so was looking here for more advice.

I also have 6 amp driver chips capable of suppling 12 of these mosfets in parallel with a 10-25khz switch rate with very low switching losses. The 10k resistor would be the input for that driver chip eliminating the slow turn on/off.

So nearly all bases are covered for the basic circuit, just need advice on a half bridge regen. what other curcuity needs to be added for a semi functional (simple as possible) regen cuircuity.

So far I have come up with OP-amp current sensing, and even possibly added a PWM charging circuit with step up transformer possibilities.

for current sensing you will need to use an op amp.. and a very very small 4 terminal high wattage sense resistor. i suggest 0.0005 Ohm and linear technologies has a paper on using their op amps as current sense amplifiers, you will want a very very good op-amp because no one ever uses this method of current sensing for such high amps.. so the resistor must be very small, so error must be reduced.

ok, this is 2nd half, fire alarm went off lol

i use an allegro hall effect current sensor, for current limiting. its only the size of a to-220... and provides isolation, meaning the sense signal doesnt contain any noise from the power side. if your going to try to use a current sense signal to determine when current is flowing a particular direction, this will prolly not work, because it has a 10us response time.. so depending on freq, thats a decent delay.

so you would need a op amp with a very very good slew rate. go to linear technologies and look for that paper on current sense methods and the op amps they reccomend for high or lowside, and just cross out all the ones that dont come in a dip package.. youll be left with a few, buy the one with the fastest slew rate , that still has the gain you need(because your sense resistor will be a fraction of a milliohm) ... there will be error here because the resistor is so small, and needs to be otherwise will burn off massive watts.

BUT! you have a pwm wave, so that tells you when current is flowing in reverse!

i have been looking into regen.. and couldnt understand why with the simple lowside setup, it wouldnt have regen... when the fets off, negative can flow back through the diode across the fet, and positive is connected to battery.. so why no regen?? .. the only difference with the half bridge setup, is that when the pwm wave is off, the highside shorts out the motor. thats it.

with these setups, they make sure that there is a small delay after the highside turns off, before the lowside turns back on again...

i have come to the conclusion that
regen can only happen by shorting out the motor when pwm wave is low, and then quikly turn off the highside thats doing the shorting, creating a spike in voltage that shoots through the lowside body diode while lowside is off.

so it seems this shorting out of the motor is needed to reverse current, make current very large, and then suddenly open circuit that loop, shooting the reverse current to the battery/lowside diode loop, with a voltage that exceeds the battery voltage.. WHILE LOWSIDE IS OFF, if lowside was on, then the motor would be powered.

i do not think your method will work. think about it, any time the battery is connected to the motor, its driving the motor... so what will you do with this information "that current is flowing in reverse" ?

anyway, if you want to try it, like i said, there is no need to sense when you have reverse current, ... when the square wave is low, you have reverse current... so just get a tc4421 inverting driver(really simple device, only has +in,-in,pwm in, pwm out), and send it the same pwm signal as your sending your mosfet.... and now you have a pwm signal that is 180 degrees off from your original... aka a square wave that is on when your fet is off... normally you would use this to drive the highside... but you can use it as your information of when motor is reversing current... but like i said... i dont know if current flows opposite unless its given a load ... which is what using the highside to short out the motor does...

and that delay i was taking about, where both fets are off, ... that is achieved by running a tiny capacitor to ground from the "pwm in" on the driver for the lowside mosfet... or from the gate of the lowside fet to ground.

you really should be using a driver for lowside, and one for highside if you use it. the lowside(non inverting, just mimics the pwm signal with more power) driver is tc4422.