Hi,
Last week I did a little research for currently available power transistors. I used Silicon Expert to look for available power transistors in TO-247, TO-3P and TO-264. I looked for complimentary BJTs, darlingtons, DMOS and to make it complete Lateral MOSFETs.
I started with (lateral) MOSFETs and noticed first, that Exicon is not a listed manufacturer in Silicon Expert. Then I continued with DMOS and I noticed, that there are only very few PMOS available anymore. Even with BJTs (incl. Darlingtons), there are comparatively few types still available, although more then complimentary MOSFETs and in contrast to MOSFETs BJTs are also specified as complimentaries.
Is class AB/B dying out? 😕
Last week I did a little research for currently available power transistors. I used Silicon Expert to look for available power transistors in TO-247, TO-3P and TO-264. I looked for complimentary BJTs, darlingtons, DMOS and to make it complete Lateral MOSFETs.
I started with (lateral) MOSFETs and noticed first, that Exicon is not a listed manufacturer in Silicon Expert. Then I continued with DMOS and I noticed, that there are only very few PMOS available anymore. Even with BJTs (incl. Darlingtons), there are comparatively few types still available, although more then complimentary MOSFETs and in contrast to MOSFETs BJTs are also specified as complimentaries.
Is class AB/B dying out? 😕
That certainly seems to be the case in the car audio sector.
As more and more electronic equipment is crammed in behind the dashboard, class D offers distinct advantages in terms of small size and cool running.
The Texas Instruments 2.1MHz TPA6304-Q1 Class D amplifier described in the link below is designed for automotive use and provides four channels in a 17x15mm package.
Concerned about transitioning from Class-AB to a Class-D? Well, you shouldn’t be. - Analog - Technical articles - TI E2E support forums
Attachments
Class B amps are essential for things like ultrasound transducers, sonar, etc, they won't go away but devices may become more expensive and exotic as volume sales go away. There's still use of class G and H in commercial PA systems too.
I don’t see vertical mosfets going away anytime soon, since large inverters will have a solid future in the power conversion sector.
Power MOSFETs are getting less suitable for class A, B and AB as the on resistance is reduced due to Spirito instability (basically thermal second breakdown). High voltage MOSFETs are usually still OK, no idea if the same will hold for future generations of high-voltage MOSFETs.
Of course vertical MOSFETs will not die out, but there are almost no complimentary types available. Most of the switching applications (e.g. motor drivers) use charge pumps to be able to use full NMOS bridges, which offer lower on resistances due to higher electron mobility.
Many of the newer high voltage mosfets are NOT ok. It’s getting harder and harder to find decent, current manufacture mosfets for use in tube circuits where you might be dropping 200 or 300 volts on them dissipating 20 or more watts (even short term, with a lower average). Anyone who has ever had a screen regulator or source follower go bang for “no good reason” can certainly relate. It’s bad enough to make you consider going back to bipolars and consider the statement that “mosfets don’t suffer from second breakdown” a myth.
I meant that high voltage MOSFETs are often still OK at the relatively low voltages used in class AB transistor amplifiers. No idea if that still holds for the very latest types, though.
You can make class-AB amplifiers with only NMOS transistors in their output stage. I've been using one with BUZ10 power MOSFETs since 1994, the good old days when there was no Spirito instability yet.
You can make class-AB amplifiers with only NMOS transistors in their output stage. I've been using one with BUZ10 power MOSFETs since 1994, the good old days when there was no Spirito instability yet.
Many of the latest types handle virtually zero power above 10V VDS. That’s not even useable with 50 volt rails. Got to stick with IRF240/9240 and their derivatives - unless you do run into some old stock BUZ10’s or Hitachi laterals.
A later Spirito paper https://core.ac.uk/download/pdf/162660523.pdf predicts that the thermal stability future belongs to SiC and GaN mosfets. Not really unexpected, these devices will slowly push silicon mosfets out of the market; its only a matter of time, until the SiC and GaN devices prices become competitive. Which will, make no mistake, happen, sooner or later.
I'm not sure what you mean; browsing through the paper, I get the impression that SiC MOSFETs have similar problems as Si MOSFETs with a too low on resistance.
The SOAR of this one looks pretty good, though: https://nl.mouser.com/datasheet/2/308/1/NVH4L020N120SC1_D-2319659.pdf
The SOAR of this one looks pretty good, though: https://nl.mouser.com/datasheet/2/308/1/NVH4L020N120SC1_D-2319659.pdf
Maybe this one is still usable:
https://nl.mouser.com/datasheet/2/308/1/NTHL095N65S3HF_D-2318679.pdf
Read this
https://ntrs.nasa.gov/api/citations/20100014777/downloads/20100014777.pdf
The vendors SOA curves need a very big pinch of salt, cheating by only going up to 10ms
It is secondary breakdown, just like in BJTs
Even familiar "old" part numbers have had die shrinks with finer cell structures and become more prone to the failure.
I guess this is the real reason many commercial amplifier brands switched back from mosfet outputs to BJT. It certainly wasn't to save money
https://ntrs.nasa.gov/api/citations/20100014777/downloads/20100014777.pdf
The vendors SOA curves need a very big pinch of salt, cheating by only going up to 10ms
It is secondary breakdown, just like in BJTs
Even familiar "old" part numbers have had die shrinks with finer cell structures and become more prone to the failure.
I guess this is the real reason many commercial amplifier brands switched back from mosfet outputs to BJT. It certainly wasn't to save money
Last edited:
It seems too many lateral fets and bipolars have been fried due to missing current limiting and a simple output short.
So stock has vanished.
So stock has vanished.
Liar! Liaaaaaar! (Excuse me while Princess Bride swims about in the brain).
The trouble with that is you don’t know how high in VDS you can really go, whether it is 10 volts or 100. Really big dies can suffer the worst from this. I would try to find parts that actually publish the real SOA in the data sheet.
I have used parts that originally advertised “purely thermally limited SOA” many years ago when they were first introduced, but later rev’s of the data sheet show Spirito-limited SOA.