Why are some MOSFET body diodes drawn as zeners and others as diodes?

[AlfBaz]

Some data sheets show the body diode with as just that, a diode symbol and others use a zener symbol
Is this an actual thing, lack of convention or some contentious topic?

[Signal32]

> The zener symbol is often used for MOSFETs where the body diode is "Repetitive Avalanche Rated
> The zener in the symbol represents the avalanche rating of the MOSFET and is of use when switching an unclamped inductive load. You won't see it conduct unless you exceed the voltage rating VDS and get to V(BR)DSS.

https://electronics.stackexchange.com/questions/111485/intrinsic-zener-diode-in-mosfet

[AlfBaz]

Thanks for that

After some reading, lets see if I understand this correctly
There are 2 types of "zener" effects. The actual zener effect caused by quantum tunneling and controlled avalanche breakdown with the latter used for higher voltage zeners.

If the body diode is not "Repetitive Avalanche Rated" and dies if it avalanches, then it's typically drawn as a diode as diodes are generally not repetitive avalanche rated.
If the body diode is "Repetitive Avalanche Rated" then it's drawn as a zener simply because like an avalanche type zener, it can handle reverse current without self destructing with Vz equaling BVdss

[T3sl4co1l]

That seems customary, yes.

Note that the body diode is already drawn in the symbol, that's the arrow pointing into/out of the center.  (At least, it is for old school lateral MOS, where the substrate and source and distinct elements.  Using it for modern VDMOS is kind of weird.)  So the antiparallel diode is redundant.

And yeah, there is a continuum from zener (tunneling) action to avalanche (impact ionization) breakdown, the terms are used fairly interchangeably as a result.

Note also that, although a transistor might be rated for avalanche, that doesn't mean it's a good idea to use it that way.  It takes time for the avalanche current to dissipate, greatly increasing device dissipation when used at high supply voltages.  Abusing avalanche ratings in a switching converter, for example (where stray inductance causes flyback spikes), will likely lead to quick failure, even if the device is apparently well within ratings.

There's one more "gotcha" about avalanche ratings: it is exclusive of good MOSFET performance.  If you need that one tiny bit of extra performance (particularly in body diode recovery time), you'll have to keep shopping.  (Probably not something a beginner will worry about, at least!)

Tim

[EEVblog]

Thanks for that

After some reading, lets see if I understand this correctly
There are 2 types of "zener" effects. The actual zener effect caused by quantum tunneling and controlled avalanche breakdown with the latter used for higher voltage zeners.

[AlfBaz]

Thanks Tim and Dave.

The reason I was looking into this was to decide what to use for a standard power MOSFET symbol
Diode, Zener or none of the above.

It would appear that the completely anal way to do it would be to create all 6 permutations and verify the type of mosfet when assigning a symbol.

Some datasheets such as this one show a diode and further down in the avalanche specs show a zener, so no consistent help there

[Cerebus]

Perhaps a picture might be helpful:-

[AlfBaz]

Thanks for that Cerebus.

I know the various FET symbols, what I'm in a quandary about is whether or not I should convey information about the type or rating of the intrinsic body diode which is usually shown in parallel with the drain and source.

From the info posted here it looks like mosfets that can handle repetitive avalanche should have a zener drawn here otherwise it should be a diode
It's most certainly a contrived issue but I cant help myself

[mzzj]

If the body diode is not "Repetitive Avalanche Rated" and dies if it avalanches, then it's typically drawn as a diode as diodes are generally not repetitive avalanche rated.
If the body diode is "Repetitive Avalanche Rated" then it's drawn as a zener simply because like an avalanche type zener, it can handle reverse current without self destructing with Vz equaling BVdss

AFAIK most power mosfets can handle some avalanche energy but only some of them are guaranteed for it. Same applies partly to power diodes.
Very old or exotic types might actually fail with very low levels of avalanche energy. 

[Cerebus]

Thanks for that Cerebus.

I know the various FET symbols, what I'm in a quandary about is whether or not I should convey information about the type or rating of the intrinsic body diode which is usually shown in parallel with the drain and source.

Yeah, sorry I was a in a bit of hurry at the time. Everybody knows the symbols but not everybody knows why they're drawn like that, what the parts of the symbol represent, and I just wanted to effectively underline Tim's point that the diode that a lot of people redundantly draw on the right is already there in the symbol as the substrate connection. Of course, finding a MOSFET with the substrate actually brought out to a pin is nowadays as likely as finding a set of chicken teeth but once upon a time they were quite common.

[exe]

I wouldn't rely on schematic to be accurate in such "tiny" details. Intrinsic diode often now drawn at all (and I don't like this). I'd say a proper part number and a short note on the schematic is a better way to indicate requirement for avalanche-rate fets.

[rfeecs]

"an intrinsic diode forms between the channel and the substrate"  No, there is no diode there.

The arrow just shows the polarity of the body material (p or n).

There are two intrinsic diodes.  One between source and body, one between drain and body.  There is not a diode between the channel and body.

The two back to back diodes can form a parasitic bipolar transistor.  So the body is usually connected to the source to short out one of the diodes.  This leaves a diode connected from drain to source, the body diode.

The schematic symbol is an abstract simplified representation of the transistor.  I wouldn't worry about the details of whether the body diode is drawn as a Zener or not.

[Sceadwian]

Yeah, I'm not sure I would trust it to intrinsically mean anything, though it may suggest that the FET should be rated for repetitive avalanche that depends on the specs of the FET not the symbol used for it. I wouldn't trust anything but the specs.

[David Hess]

"an intrinsic diode forms between the channel and the substrate"  No, there is no diode there.

The arrow just shows the polarity of the body material (p or n).

There are two intrinsic diodes.  One between source and body, one between drain and body.  There is not a diode between the channel and body.

The two back to back diodes can form a parasitic bipolar transistor.  So the body is usually connected to the source to short out one of the diodes.  This leaves a diode connected from drain to source, the body diode.

I have always assumed the customary diode shows the shorted base-emitter junction of the parasitic transistor.  There are *two* parasitic diodes forming a parasitic bipolar transistor but only one diode is shown on the MOSFET with a separate substrate connection.

On my schematics, I explicitly show the source-drain diode if only to remind me of which direction it faces.  The difference with avalanche rated diodes which in this case is the collector-base junction of the parasitic transistor is that they are specially processed to produce a uniform junction preventing hot spots and current crowding.