Does a MOS-FET with integrated freewheeling-diode need another external freewheeling diode

I think you may be overcomplicating it. When you short the terminals of a running DC motor, the current in the motor immediately reverses, rapidly braking the armature, exactly as you say. However, this current quickly dies away because the windings have resistance. The resistance means the windings warm up, dissipating the energy as heat.

Thus, the mechanical energy in the armature is converted into heat in the motor windings.

Obviously in your theoretical frictionless, super-conducting motor there is no way to dissipate the energy, so oscillation would result. But they don't exist in the real world.

Using a resistor across the terminals reduces the braking effect because it resists the flow of the reverse current (the current that stops the armature). The higher the resistance, the smaller the reverse current so the slower the deceleration. Obviously the energy is then dissipated by both the coils and the resistor.

So, the fastest way to brake a DC motor is to put a dead short across its terminals. And in those circumstances the only place energy can be dissipated is from the windings.

At least that's my understanding. I do wonder if we might be in violent agreement here. :rofl:

Yes, you are right. You can get even faster braking than dead short by applying voltage of opposite direction.

Thermal management potentially applies to every circuit, efficiency affects how much heat you have to dissipate, whether in MOSFET or a separate diode, or an RC snubber. Schottky diodes are used to reduce heat dissipation sometimes in switching circuits, nothing controversal there. In small low power circuits thermal management is usually not a difficult problem, but with higher power switchmode/PWM circuits its one of the major design challenges as it impacts cost directly. There's even been a $1,000,000 prize challenge to advance state of the art for this: https://en.wikipedia.org/wiki/Little_Box_Challenge

Nothing you've said in #5 changes these facts of life for switch-mode power conversion circuitry. Schottky diodes can reduce switching losses in two ways, lower forward voltage and faster reverse-recovery time - however they have disadvantages too, reverse leakage and need to be kept cool to prevent thermal runaway.

The energy stored in a coil equals the heat to dissipate, regardless of how it is dissipated. That's why I asked for your special meaning of efficiency in the case of switching inductive loads. Dissipating the heat only within the coil is not desirable, instead splitting the heat into the (sealed) coil and an external resistor allows to better get rid of the heat in the external resistor.

Freewheeling diodes are not there for heat reduction but for reduction of EMF (voltage spikes allover...) when a coil is turned off.

Not sure that makes any sense at all! :roll_eyes:

The inductor is already heated by the current passing through it. When you switch it off, the current drops to zero; how fast it does so depends on the mechanism you have provided to mitigate a voltage "spike".

But that is all that happens - the current falls to zero and it no longer heats due to its resistance. However that heating effect is ipso facto less than when it was being powered. There is nothing to "dump" by way of heat beyond what it was before. During the fall in current it is in fact much less than before.

There continues to be a perverse mentality that there is some sort of "surge" occurring when you switch off an inductor. This is completely untrue.

1 Like

You are right, my argumentation about heat distribution is unimportant. But I hope to have shown that the kind of external freewheeling circuit is unimportant WRT efficiency.

I'm first going to take the common example of motor drive:

Motor windings (for instance) are made of copper and forcibly air cooled for free. Yes, you definitely want heat dissipated there rather than in your motor driver unit.

The motor windings will typically be dissipating 10% to 30% of the total power anyway, adding a few more percent of dissipation to them is a small change. The motor driver might be dissipating 2% to 5% of the total dissipation, moving 1% of dissipation from the driver to the motor is a big win for the motor driver, and a negligible change to the motor's thermal behaviour.

However the whole premise you have doesn't make sense - if you reduce the energy dissipated in free-wheel diodes, that energy stays in the magnetic circuit - it cannot just turn to heat by itself. For PWM typically you'll see that during free-wheeling the current is dropping, but not all the way to zero, so energy not dissipated stays in the circuit as extra current for the next cycle, reducing the load on the power supply.

That's a special case mentioned long ago. For freewheeling the external loss should be low, for braking or fast turn-off high.