Triac vs MOSFETs for AC switching

...and 200A can melt silicon to a nice short.

Commutation in SCRs and Triacs relies on reversing the current flow in a device for a specified length of time.

With a short circuit load and a subsequent high di/dt I can't imagine there being any effective method of turning off one of these devices through commutation!

know if it's better to go with a pair of FETs or with a Triac? My design ha s to be able to withstand a short circuit current of 200A for a few msec, b ut normally it's running around 0-8A. I like that the FETs can open up rea l quick, I have to wait for zero-crossing to open my Triac.

Use a socket to plug the triac into. Supply extra triacs. Then, when the tr iac blows, plug another in. Advertise this as a breakthrough - the fast response solid state fuse. a feature.

I confess to be ignorant of any issues with turning off a relatively small IGBT.

I'm only aware of di/dt limitations for TRIACs and Thyristers but not for IGBT.

I've looked up a couple of IGBTs and as an example STGP8NC60 cites a di/dt turn-on as a feature rather than a limitation. It's description says "short circuit withstand time 10us" and has a helpful Turn-off SOA graph to suggest it can handle 300A at 400V at turn off.

Here's a bit on it (soft shut down):

AFAIUI, the problem is more what happens to the energy in the external inductance than di/dt on the device itself.

Best regards, Spehro Pefhany

Perhaps I was being pedantic, that the IGBT is unlikely to be destroyed by a fast di/dt alone, but from other circuit consideration such as from a turn off HV spike.

Whereas a SCR or TRIAC will themselves have a overriding di/dt specification without any further circuit considerations complicating the issue.

Yeah, AFAIK it's just dV/dt for FET structures. Something about forward-biasing the parasitic BJT structure (same idea as thyristor dV/dt, but much less sensitive).

I don't think I've ever heard it explained why that's a bad thing -- surely it just manifests as excessive Miller effect at high rates. Presumably, the minority carriers also stick around for a while, with an effect similar to body diode recovery (taking on the order of ~200ns in power devices), which will cause excess dissipation, but not outright destruction. Dunno.

Thyristor dI/dt, of course, is bad because of hot spots, and dV/dt because of re-latching.

Tim

In the application i foggily remember the commutation of the SCR was part of normal operation. Like horizontal output drive in some (now older) TVs, but different purpose. IIRC in case of failure fuses were supposed to blow before damage to more valuable equipment occurred.

?-)

Triac?

that

to

off PDQ.

description

SOA

Along the lines of Tim post, high enough dV/dt can cause retrigger and resumed conduction in SCRs. It is often specified (or at least used to be).

?-)

I think the fuses blew after destruction of the GTO.

Sony had this idea of having 2 GTOs in series to share the flyback voltage. Not something I would recommend!

It's no pb in a flyback-like structure. The FU occurs with bridge-like ones when you hard turn on one device of the branch. The other device excess current can't lower the VDS/VCE rise rate and that leads to destruction (can't use Joerg's copyrighted noises here).

just

in a

part

supposed

This was not GTOs. They weren't around back then.

?-)

In part you are correct. In those days they were called a Gate Controlled Switch (GCS).

It's quite easy to find articles of these 70's TVs:

Yes, and they failed very often! It seems that in many cases there were some sort of break down issue that you could use to calculate life expectancy.

Jamie

On a sunny day (Mon, 28 Apr 2014 20:33:55 -0400) it happened "Maynard A. Philbrook Jr." wrote in :

The ones in the Sony TVs failed mainly becaue bad soldered PCBs. The drive would be intermittent, and the swith would stay on. With a normal transistor switch you would just get some irregular scan.