How to interpret the SOA line?

Jump to Latest
#21
wg_ski said:
That’s called secondary breakdown. It just happens to be happening at zero bias, as opposed to the usual forward or reverse. It will likely follow the FBSOA curve, because it really IS forward biased, with Icbo. In some big power devices, especially ones designed for switching, the Ic limit at rated Vceo may not be much more than the spec Iceo. Sometimes just a handful of watts.

Electromigration is a function of current density. Jc can be quite large in a hot spot, that’s why it’s hot.
Click to expand...

It also happens in MOSFETs, there it is called the Spirito effect, the guy who first mentioned it. The cause is unequal Vgs(th) on the die, which causes some small parts to conduct more and hog the current. The end result is the same as secondary breakdown in BJTs: release of the build-in black smoke ;-).

Jan
 
#22
Many “switching” MOSFETs have absolutely terrible FBSOA (but good RBSOA). A 300 watt device can be as poor as 30 watts at 10 volts. Many older devices that I’ve used for linear and switching originally claimed full power handling to full VDS rating. I have the old data sheets that show it. But look up the data sheet now, and it shows the SOA kink. Not nearly as restrictive as these hyper fast trench types they make now, but not what they used to claim. Some parts like IRF240 family still show full power at full VDS, and I’m not sure what to believe. The old standby parts originally made by IR do still seem to hold up as linear amps, but what about current production? And they go obsolete all the time and get replaced by “updated” ones which usually show the kink. When I really worry about it is using hexfets in tube amp circuits, like screen regulators or K-followers, where a proven old school type may not even be available for the required B+.
 
#23
I share similar confusion about SOA specs to some degree.

Recently when trying to choose cheap devices for a valve anode load CCS I chose Zetex ZTX558.

For the required voltages I run, and CCS current, I am probably at about 80% the DC SOA at 25C.

I thought that was a bit close to the mark - but (fingers crossed) it has worked fine for the last few days, at 6mA and 100V.

I'm still none the wiser (unless it fails)
 
You must log in or register to reply here.
Top Bottom