MOSFET H Bridge

[kinarfi]

I've been working on an high amperage H bridge and thought I had it figured out so I could use all N channels, the ones hooked source to positive 12 v would have a gate voltage of 24 and it looks good on paper until you look at the zener in the fet and if you turn the fet on that has the source on negative, the zener in the positve leg is forward biased.
any ideas around this?
kinarfi

 

[MrAl]

Hi,


It's not a zener, it's a Schottky diode. Look that up.

 

[kinarfi]

Thank you MrAl, I was wondering what the Zener voltage was, guess it doesn't exist, I was just going off the symbol. The fact that it is a Schottky, makes the problem worse, lower Vf.
Still have the problem as shown in the drawing. Still need help though, maybe a high power diode in series or do they make a mosfet without the diode?
Thanks
kinarfi

 

[Ubergeek63]

Thank you MrAl, I was wondering what the Zener voltage was, guess it doesn't exist, I was just going off the symbol. The fact that it is a Schottky, makes the problem worse, lower Vf.
Still have the problem as shown in the drawing. Still need help though, maybe a high power diode in series or do the make a mosfet without the diode?
Thanks
kinarfi

oy! check the specs... they are not schottkys unless it is a copack with a schottky. it is the optimization of the parasitic substrate diode in the FET. avalanche rated FETs are generally shown with zeners for the substrate diode... shottkys are wildly different in structure from the vanilla PN junctions that make up the parasytic substrate diode of the FETs.

the new FETs also often show the full characterization for the substrate diode so designers can tell if they need an external one or not.

dan

 

[kinarfi]

https://www.electro-tech-online.com/custompdfs/2010/02/AN-9010-1.pdf Start with page 30 (3) Diode Recovery dv/dt.
This is the circuit I have, EXACTLY!!!
And I believe it explains my problem in the thread FET Killer.
Kinarfi

 

[smanches]

Never rely on the intrinsic body diode of the mosfet for reverse protection. Put your own set of protection diodes in across each mosfet.

 

[kinarfi]

Never rely on the intrinsic body diode of the mosfet for reverse protection. Put your own set of protection diodes in across each mosfet.

Relying on it is not my problem, the fact that it's there is my problem and how to defeat it. Looking into IGBTs, do they mind being reverse biased E to C?

 

[kinarfi]

oy! check the specs... they are not schottkys unless it is a copack with a schottky. it is the optimization of the parasitic substrate diode in the FET. avalanche rated FETs are generally shown with zeners for the substrate diode... shottkys are wildly different in structure from the vanilla PN junctions that make up the parasytic substrate diode of the FETs.

the new FETs also often show the full characterization for the substrate diode so designers can tell if they need an external one or not.

dan

After googling parasitic substrate diode, I take it that the "zener" in the diagram doesn't really exist, it just there as sort of a warning about the parasitic pn properties of the FET. At least it's not there on purpose.
Kinarfi

 

[mneary]

After googling parasitic substrate diode, I take it that the "zener" in the diagram doesn't really exist, it just there as sort of a warning about the parasitic pn properties of the FET. At least it's not there on purpose.
Kinarfi

It may be zener, or it may not be. Although the diode is an unavoidable part of the structure, manufacturers can and often do manage their characteristics so that they are useful. Modern MOSFETs have full specs on these diodes and can be used accordingly. The data sheet is your friend.

It's old information to never rely on those diodes, which goes back to when the diode characteristics were not controlled or specified.

 

[MrAl]

oy! check the specs... they are not schottkys unless it is a copack with a schottky. it is the optimization of the parasitic substrate diode in the FET. avalanche rated FETs are generally shown with zeners for the substrate diode... shottkys are wildly different in structure from the vanilla PN junctions that make up the parasytic substrate diode of the FETs.

the new FETs also often show the full characterization for the substrate diode so designers can tell if they need an external one or not.

dan

Hello,


Yes they are not 'exactly' Schottky's, but it's easier to explain it that way then going into
all the details of how these diodes are faster than most diodes yet slower than most
real Schottky's.

 

[MrAl]

It may be zener, or it may not be. Although the diode is an unavoidable part of the structure, manufacturers can and often do manage their characteristics so that they are useful. Modern MOSFETs have full specs on these diodes and can be used accordingly. The data sheet is your friend.

It's old information to never rely on those diodes, which goes back to when the diode characteristics were not controlled or specified.

Hello,


Since when do manufacturers make a Mosfet with a zener on the output side?
Zeners on Mosfets are usually there to protect the gate.
I guess it's possible though, but i've never seen one with a zener across the drain and source (yet).
Perhaps post a link or two.

 

[MrAl]

After googling parasitic substrate diode, I take it that the "zener" in the diagram doesn't really exist, it just there as sort of a warning about the parasitic pn properties of the FET. At least it's not there on purpose.
Kinarfi

Hello again,


Yes that is right, it is not a zener.
It's not exactly a Schottky either, but it's somewhere in between a Schottky and a regular
diode. That is, it is faster than most but slower than the best.

What this means is that for a properly controlled MOSFET the integral body diode can be
used just as if it were a fast recovery diode. If the rate of change of voltage is too fast
however the reverse current will be too high and could blow out the diode and thus the
transistor too.

The data sheets give a dv/dt rating so you could look that up if you like.
The defining parameter to tell if you can use the body diode or not then is the dv/dt spec.
If you have any doubts, simply use an external fast recovery diode. Watch out for some of the
zero recovery time diodes though as they sometimes have a fairly high forward voltage spec.

 

[smanches]

They are drawn as schottkey's because they have a low forward voltage, although they are way slower than schottkeys. If you put four more actual schottkeys in parallel, you wouldn't have the problem you are having, as they would switch much faster than the intrinsic diodes, and take the back current out of the MOSFET itself.

 

[Ubergeek63]

After googling parasitic substrate diode, I take it that the "zener" in the diagram doesn't really exist, it just there as sort of a warning about the parasitic pn properties of the FET. At least it's not there on purpose.
Kinarfi

it is very real, as a side effect of the FET structure. EVERY PN junction is a zener, but not every one is optimized and characterized for that function.

Transistor base emitter junctions typically breakdown at around 6V and will operate as zeners at that point. it is also known as the avalanche region, though in normal operating mode a transistor collector base junction breaks down aand the transistor self destructs.

dan

 

[kinarfi]

Thank you all, I've learned quite a bit here and remembered a lot of stuff I used to know and been prompted to look in the right places for information, I love this Forum. Experimented and realized / noticed that Vce is much less than Vbe, I think I knew that but didn't remember.
Kinarfi