That internal diode exists with every MOSFET. In most cases it can not replace an external diode over the inductive load.
Stefan, you apparently do not understand what a "freewheeling-diode" does.
Perhaps you need to research it a bit - this thread and this may I hope, help.
In a single MOSFET circuit the body diode (as its called) is useless for protection since the free-wheel diode needs to go across the load.
In an H-bridge circuit the diodes are useful as each diode protects another MOSFET.
Sometimes for extra efficiency external Schottky diodes are added to an H-bridge (Schottky diodes have a lower forward voltage, reducing bridge losses).
It depends on what "efficiency" means to you. Diodes allow the current to persist, for some time, for smooth use with PWM. For fast turn-off time instead RC snubbers perform better.
Well for instance the engineers designing large power converters really like to see 98.5% efficiency over 97.5% efficiency(*) because that's a huge 40% reduction in heat to get rid off - thermal management is a large proportion of total BoM cost.
Also paring away at switching losses allows higher frequency switching, thus reducing the size and cost of inductors and capacitors.
(*) naively you might not worry about 1% change in efficiency, but at high efficeincies its the losses that become important.
That "efficiency" is not applicable to PWM applications. Please read and understand what I wrote in #5.
Just to underline what the others have said: that is NOT a freewheeling diode. It cannot perform that function.
In what application has a snubber better efficiency than a H-bridge?
Whenever fast turn-off is required.
And that is?
Assuming you are driving inductive load H-bridge can return the energy to the supply. RC Snubber dissipates it in the resistor.
Hi everybody,
thank you very much for your answers. I think I understand now. Yes external diode is nescessary.
best regards Stefan
That's a feature of the H-bridge. With MOSFETs it also is more efficient than Schottky diodes in that circuit with regards to energy return.
A diode dissipates the energy in itself - what's the difference with regards to efficiency?
The snubber dissipates all energy stored in the inductance. The diode dissipates only a small amount, most is returned to the supply.
Please think a bit more before answering. A freewheeling diode over an inductive load can not return anything to the supply.
Now I see the problem: MarkT was speaking about adding Schottky diodes to an H-bridge. In a H-bridge a snubber makes little sense. For a single transistor switch a snubber may be sometimes useful. Except for a relay coil where do you want fast current decline?
E.g. in braking a motor - the opposite of freewheeling.
That's wrong, isn't it? The fastest way to brake a motor electrically is to short-circuit it, such that the current is free to flow as much as it wants. The energy is dissipated by the motor windings.
Using a resistor to dissipate the energy causes less of a braking effect (the higher the resistance, the less the braking).
Or are we expressing the same thing in different ways?
As the induced current is (initially) constant the dissipated energy increases with the load. With no resistance (superconducting coil) the current would flow forever. That's radically different from capacitive storage that drains slower with increasing resistance. I had that discussion already a year ago in an electronic newsgroup.
To stop a motor dissipating (electrical) energy stored in the winding is not enough. It is only the first step. For "braking" you need to reverse current in the winding somehow to also revive the mechanical energy. The easiest way is to short the motor and let the back-EMF do this for you. Even a superconductive motor would be quickly stopped this way - the mechanical energy would be converted to electrical energy in form of massive current. Of course this current wouldn't "go away" as the motor stops and will cause the motor to start rotating to an opposite direction. A shorted superconductive frictionless motor would oscillate forever. Adding some resistance would damp the oscillation but also decrease the braking speed (a frictionless motor with infinite resistance between terminals never stops).
In other words braking is not about electrical energy in the winding but about mechanical energy of the motor and its load. To stop a motor you need to convert the energy to electrical and get rid of it (either in a resistor or store it somewhere).
