output device technology as of 2012

[DJNUBZ]

Now that I have built a few pieces of my own I have started to realize that a lot of designs available to the DIYer use output devices that seem quite old considering how fast technology moves in other circles. I wonder why this is? Has transistor technology moved so little in the past 10 years that there isn't anything that could be considered an upgrade? Is it because topology is the limiting factor, not device design? I actually find both hard to believe but I don't know enough about EE to find the answer on my own. What do you guys think?

[Enzo]

Depends upon the design. More recent circuits often use power amp ICs rather than discrete transistors. If nothing else it is simpler and cheaper. In this day of battery operation for many things, efficiancy is very impoortant. So digital amplifiers, and class D are popular.

But we continue to use older devices because they work. If I need a chip amp, and I have LM3886 in my drawers, unless I just wanted to experiment for the sake of it, why would I go buy something else? And if I need a robust bipolar output transistor and have been using MJ15024 for years, and they are still stocked, then what is my incentive to use some other transistor I haven't seen before? Some folks might suggest the MJ21194 as newer. Yes it is, but it pretty much drops in place of the older part, so no advancement.

Newer digital circuits are more demanding. A DIY amp can be pretty simple to make, but if you go digital, then the high freqs involved will put greater demand upon layout and wiring. Some parts may now onoly be available in surface mount, the boards for wwhich may be beyond many DIY guys.

Look at it the other way, what would your dream device be like? Don't worry about HOW they accomplish it, just what features would it have? And of those, which are not available now? What would you consider advancement to be?

[fpitas]

To some extent you're right, because the manufacturers go where the money is. Better transistors for switching power supplies will sell in very large quantities, but specialized audio devices, not so much. Sometimes there's overlap, like the SemiSouth switching JFETs that Nelson Pass uses in the J2, and various switching MOSFETs that are used as linear output devices.

[sofaspud]

These are good answers. The highest value is placed on components and systems that are smaller, lighter, faster, and lower power. There just isn't a large enough market for big audio output transistors to justify increasing the R&D budget. The chip amp engineering group gets priority. In a way, this implies that the market for those types of transistors is somewhat satisfied by the parts currently available.

[von.ah]

This is an interesting thread!
Forgive my ignorance, but would it be possible to make something along the lines of 'the Beast with a 1000 jfets' using SMD transistors?

[fpitas]

Originally Posted by von.ah

This is an interesting thread!
Forgive my ignorance, but would it be possible to make something along the lines of 'the Beast with a 1000 jfets' using SMD transistors?

It's a lot tougher to get surface mount devices hooked to a heat sink.

[jan.didden]

Some of those new Silicon Carbide Static Induction Transistors are exiting for new audio designs. Nelson Pass pioneered them, and even went as far as to have a custom batch produced, the PASS-SIT-1.
Linear Audio Vol 4, due to come out 1st September, will have a couple of articles on designs with SITs, by NP, Lennart Jarlevang and Alfred Hesener.

jan didden

[fpitas]

Originally Posted by janneman

Some of those new Silicon Carbide Static Induction Transistors are exiting for new audio designs. Nelson Pass pioneered them, and even went as far as to have a custom batch produced, the PASS-SIT-1.
Linear Audio Vol 4, due to come out 1st September, will have a couple of articles on designs with SITs, by NP, Lennart Jarlevang and Alfred Hesener.

jan didden

As the owner of a J2, I can say the SemiSouth JFETs are nice-sounding beasts, too. NP is having SemiSouth make special audio versions for him, but from the specs it looks like the main difference is the Vgs threshold voltage.

[jan.didden]

Originally Posted by fpitas

It's a lot tougher to get surface mount devices hooked to a heat sink.

Not necessary. With a thousand devices, each Watt amplifier dissipation is only 1 mW per device.
The beast also has no heatsinks, the FETs are TO-92 case.

jan

[fpitas]

Originally Posted by janneman

Not necessary. With a thousand devices, each Watt amplifier dissipation is only 1 mW per device.
The beast also has no heatsinks, the FETs are TO-92 case.

jan

Interesting.