Incorrect Circuit for Regen run-off - diode

[marcuso]

So I got some advice from an electronics web site about the diode necessary for preventing charge floating around when the battery connection is cut from a switch. The people over there (who seem to be pretty clued up), think that the circuit diagram is wrong for the diode placement. Rather than having it on the +ve between the pins on the switch, it should be between the switch output on the +ve and the -ve battery terminal. So my question is, should the run-off/regen power from running the motors when disconnected flow back to the battery on the +ve terminal or the -ve, or does it matter ?

[Basicmicro Support]

That would be a diode used to clamp an over voltage(If I understand your description). You would usually use a TVS diode for that(and we have them on some models of the Roboclaw just for that purpose).

What we want is a simple one way path back to the battery if the switch is turned off or the fuse blows while the motors are spinning and energy is moving into the battery positive terminal we dont want to prevent that(bad things happen if the regen energy cant go anywhere).

[marcuso]

I've really started a lively debate with this question of regen and diodes on some other forums, and I keep getting a response of

"No amount of regen charge will flow in to the +ve battery unless the amount of regen from the motors > level of battery, and also there will not be any regen at all, unless the motors are turning faster than the stated RPM of the motor (mine are quoted at 330RPM)."

So I've got my diode, and I'm going to put it across the switch pins anyway, but most people are saying it will never be a problem, because the motors will never generate anything substantial anyway, and even if they do the circuit wouldn't work as the battery will not take the charge anyway.

I'm going to try it with a spare motor I have of the same rating as my robots, I dont think I'll ever get it faster than the 330RPM its quoted as, but I will see what charge I get when spinning it at 120 RPM, which would be the equivalent 40ft/min or 20 cm per second or 1 metre every 3 seconds. This is a feasible speed if its traveling down hill on Jupiter

It's a track robot BTW so this really is exaggerating its run away speed.

[Basicmicro Support]

Simply put they are right and wrong at the same time. The battery will not recharge/take regen energy if the voltage from the motor system is less than the voltage of the battery. However when you slow down a motor, energy is converted from kinetic to electromagnetic. That energy has to go somewhere. Some of it will go into the large filter caps on the Roboclaw but thats not a lot. The rest will try to go back into the battery but wont be able to since the voltage wont be as high as the battery, initially.

However, Ohms law sas voltage will increase if the resistance to a current increases. Eventually regen voltage will get increase to a level higher than the batteries voltage and current will start to flow back into the battery.

Note this is why motor control systems run from power supplies usually need a power dump/voltage clamp circuit to burn off this regen energy. Otherwise the voltage would just continue to increas(until the motor stopped) and in some cases could get high enough to damage the controller and/or power supply.

I can garuntee "never being a problem" is a serious mistake on their part. I've had more than my share of RMAs to repair because someone used a power supply that could not take any regen and poped components on the board because the voltage increased to well of 50v or someone disconnected the battery while the motors were running causing the same effect.

[marcuso]

OK, but is the 50v + a result of voltage spike on disconnect ? As you said you had built protection in for that already on the board, so voltage spikes are already handled by your design right ? If so it's stil not necessary for a diode ! I've put one on anyway.

[Basicmicro Support]

The voltage spike protection will only protect for a very short time(its there for inductive spikes not regen caused over voltage, think spark vs generator output). it is not designed for protecting long over voltage spikes(regen spikes can last as long as the motor is still spinning, though they are worse the faster the motor slows down when the h-bridge is active) which will be caused if you disconnect the battery + when the motor is spinning or when you use a power supply that cant handle regen and slow down the motor too quickly.

Also only some boards(60A and larger) have the TVS diodes for inductive spike protection.

And when I say 50+ volts thats because if you hit 50v you WILL(absolutely garuanteed) damage something on the board(34v boards anyway). However I've seen voltage spikes go up past 60v when running a 12v motor on a power supply(luckily I was using an HV board when that happened so I didnt damage anything) that couldnt handle regen. The same can happen when a battery is disconnected on the + side.

[marcuso]

OK thanks for your patience with all this, I've wired up the diode and fuses and e-switch, I will be posting it on my blog, so I'll point people to it to hep them out, when Ive typed it all up.

Marcus,

[Basicmicro Support]

Sounds good. Please reply in this thread as well with the link. I assume you'll be adding it to your project thread.

[scampa123]

Yes, I'd love to see what you are doing, and the specifics on the diode and switch...Thanks!!