What are the parts in this schematic

MOSFETs

Plus those diodes are almost always built right into the MOSFETs.

--

Michael Karas
Carousel Design Solutions
http://www.carousel-design.com

Maybe you can help me understand. MOSFETS have a parasitic diode. How do you tell whether a FET has a built-in diode that's designed to take significant forward current in normal operation? If I stick it on the curve tracer, is that good enough or are there considerations other than the V-I curve?

...

e

..

Isn't that the normal feature of a MOSFET? They take very heavy current with very low resistance.

Only when on. Reversed and off, a common MOSFET looks like a junction diode.

This behavior isn't necessary, it's just typical of three-terminal devices where the substrate and source are tied.

Some examples without are: monolithic circuits, four terminal MOSFETs (usually old metal-can devices), JFETs and some newer RF or GaN devices.

Generally, CMOS circuitry diffuses NMOS into the bulk substrate, which is connected to Vss. PMOS is not, and its substrate can be tied anywhere (though Vdd is of course typical). (NMOS can be fabricated with an isolated substrate as well.)

FETs without substrate connection (MOS- and J-) still act like a diode in reverse, but the characteristic is the Vgs=Vds curve, so it has a different voltage drop, curve, tempco and such. For small Vgs(th), this isn't so bad, and it comes without recovery charge which is nice (when Vds is negative, the gate is biased on, in the drain direction -- when Vds rises again, it turns back off just as by varying gate voltage).

These things are GaN MOSFETs:

but because they don't have a GaN substrate (the GaN is a layer grown on top of an insulating AlN layer, which in turn is grown on an actual silicon die), they look like very well insulated, enhancement mode JFETs. Reverse bias behavior (when Vgs=0) is not unreasonable, because Vgs(th) ~ 1V, and transconductance is very high (so the "resistance" is small).

SiC JFETs also exist, which as far as I know, are constructed same as a regular Si JFET, but just as SiC schottky diodes have much higher breakdown voltages, so do SiC JFETs. I haven't seen these available through most distributors, but there are a few manufacturers. SiC MOSFETs are available, such as:

which are true three-terminal (with substrate) MOSFETs, in the conventional sense. The reverse recovery behavior is very familiar, because SiC isn't really a magical material: its main advantage is the high breakdown voltage; electron mobility (which determines conductivity of the channel) and recombination (which determines minority carrier lifetime -- recovery time of a junction diode) are fairly similar to silicon.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

Read the datasheet. In some cases these diodes can also act as zener diodes (reverse avalanche rated) in other cases external schottky diodes are better.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

There are two diode junctions (per MOSFET) shown in your schematic. The one commonly referred to as parasitic is normally reverse biased between the body (substrate) and the conduction channel. This is internally connected to the source terminal and is the little arrow pointing to the left.

The other diode is intentionally added (not parasitic), is normally reverse biased, but conducts when the voltage across the D-S goes negative. This can occur when the MOSFET shuts off current through an inductor and keeps the reverse D-S voltage from breaking down the device. This is shown as the diode pointing up.

Not forward. Reverse.

Look it up on the datasheet.

The presence of this reverse current diode can be fooled by the existence of the parasitic junction diode. If you don't know what you are looking at, one can look like the other.

--
Paul Hovnanian     mailto:Paul@Hovnanian.com
------------------------------------------------------------------
This isn't right.  This isn't even wrong. -- Wolfgang Pauli

Can you make a decent diode (low Vf at high current, low C, low Trr) with the same MOSFET process on the same die, or are the diodes added chips?

I used the word "forward" diode current because that's what I meant. Most of the buck converter app notes I read suggested that the internal diode was crap and some efficiency can be gained by adding an external diode. It's the FORWARD characteristics (of the diode) that matter in many applications where a diode is required.