Best cost is no object darilington NPN's?

[Fusion916]

What is considered the best TO-3 and the best of any package?

[wahab]

The original BDX66C/67C if you can find some wich are not fake....

For recent production with semi plastic case there s no doubt
that the Sanken products are the best.

[jcx]

for repair you may be stuck - but for new design or even repair with added mods/tweaks you should know that "the best" monolithic Darlington is a lame joke compared to purpose designed descrete dual ef for audio power applications

[wahab]

Originally Posted by jcx

Darlington is a lame joke compared to purpose designed descrete dual ef for audio power applications

In the 80s surely but the last decade production is largely
on par with discrete double EF

See Sanken s MP1620/MN2488 , 2SB1587/2SD2438 , 2SB1647/48/2SD2560/61.

SANKEN ELECTRIC : Power Transistors : General Purpose : PNP

[Jay]

Remember that darlington is dual transistor in single package. When you design an amplifier, you have a set of objectives, and your transistor must fit to that objective. For a single transistor, speed and linearity at AF is the most important with limitation "only" in Vce. But for dual transistor the situation is different. There is almost nothing to prefer for darlington for audio purpose. Not cost, not space, not power/gain, not flexibility (of course), not absolute sound performance (for sure), nothing that I can think of. Actually I have no idea of the original objective of the purpose of packaged darlington.

[jcx]

they're OK for industrial DC motor speed control, slow switch applications where today you'd just use a MOSFET

[Andrew Eckhardt]

I've seen modern darlington output stages (those sanken #'s look familiar) only in low cost audio equipment. It rids the board and sink of a pair of driver transistors. It also makes you accept having the driver emitter resistors tied at the output transistor emitters.

[Fusion916]

I didn't know audio gurus were so anti darlington.

From a performance standpoint, Why are darlington pairs looked down upon? You give up headroom, I understand that, but what else makes them a crappy output stage?

[wahab]

They are OK on the perfs front , at least the Sankens above , but their
main drawback is that they are more difficult to stabilize thermicaly
than a discrete siblings.

That said , in 30 years or so i had only a single failure with darlingtons
for the numerous amps i built or even repaired using theses devices...

[Jay]

Originally Posted by Fusion916

From a performance standpoint, Why are darlington pairs looked down upon? You give up headroom, I understand that, but what else makes them a crappy output stage?

Originally Posted by Andrew Eckhardt

It also makes you accept having the driver emitter resistors tied at the output transistor emitters.

Driver emitter resistor tied at the output transistor emitter. I think this is the limitation for a darlington amp to become one of the best.

You may disagree, but I have been after the best amps for many years. At "that" class level I noticed that such output stage doesn't sound "good". Not even Krell. Not any quasi amps, except a quasi amp with phase splitter like JLH where driver emitter is not tied to output.

If I have to guess the reason, I think because the dynamic/sonic is limited. Best is to tie the emitter to the other half for symmetrical topology. Or at least to the rail. Even when making a standard CCS, I prefer to connect the resistor to rail instead of to ground.

You may find that an amp with such an output to sound wonderful, well, I have found some also, but when we are talking about the best, we want an amp that has good sonic properties because imo it is the main reason why we enjoy the music (not that sweet vocal). The effect is in long time listening where less sonic amp tend to sound boring and not interesting.