1x2SC5200 or 3x2SC4793

[Bigun]

I'm just wondering ....

For a Class AB or Class A amp with push-pull outputs with requirement for say 25W output.

I could easily use single pair 2SC5200 / 2SA1943 which I've done in the past.

OR

I could use a triple of the 2SC4793/2SA1837 driver devices to get adequate power handling. I'm thinking the advantage would be lower distortion because I have less current variation = spread over 3 devices instead of 1. Attraction is lower price fact that these 'driver' devices can be mounted to heatsink without insulators since they are plastic only.... and they are Really nice performing devices.

[CBS240]

25W at what Z? Even for 8R peak current is 2.5A, or 833mA per device. Beta drops quickly after 300mA as does Ft for 2SA1837. 10W @ 8R would be more reasonable for 3 pair of these devices as final OPS, I would think. Of course you could parallel 1000+ small signal J-fets....Oh wait, hasn't that been done before? Mr Pass.


I ran into a similar situation with my HEC vertical fet amp. The TO-3P P-type device I aimed to use is now discontinued. So I had to use two TO-220 instead. The con is more complexity but on the good side TO-220s cost half as much as the TO-3P. Also the two of them together have more conductance and can handle more power than the single TO-3P. In fact, in this case they are the same die, just a different package.



BTW that avatar is pretty gnarly

[Bigun]

10W is fine but a little boring, so perhaps this is a daft plan in the first place.

And I ain't going to solder by hand 1,000 transistors - that's just begging for a pick-n-place machine at a pcb assembly house

[Ian Finch]

I think you might need to be careful keeping PCB tracks tidy to maintain comparable low distortion arrangements of the Emitter/Collector traces. OK, so some distortion can be good, but I haven't yet heard of anyone getting great sonics from EMI distortion.

Electronically at least, it should give 1/3 the HF distortion of a single device at the same time as tripling power, so why not give it a try?

There is also the T0247 style 2SA1491/C5198 if the price is right and. of course, Ostripper's NJW On-Semi types as a single pair to consider. ~$3,00 for all the OP devices and only 2 screws and pads is hard to beat!

[AndrewT]

25W into 8r0 requires ~ 5 to 6 times that as total device dissipation. Say ~150W.
A single pair of 1943/5200 comes to 300W. OK.
A 3pr stage would need 6 devices each rated @ >=25W.

There are not many insulated To220 that can do that.

[Jay]

Originally Posted by CBS240

BTW that avatar is pretty gnarly

Yes, too much distortion Should be changed with lower THD avatar...

[Bigun]

Originally Posted by Jay

Yes, too much distortion Should be changed with lower THD avatar...

How's that ?

[Ian Finch]

Originally Posted by Bigun

How's that ?

Seriously? ... looks gnarled, now.

[Ian Finch]

Originally Posted by Ian Finch

the T0247 style 2SA1491/C5198 if the price is right

I should straighten that out to 2SA1941, sorry - dyslexic as well as slow.

Seeing the 2SC4793/A1837 are a bit light as drivers even, there are other "fullpack" style TO220s and On-Semi MJF15030/31
look like they could kick a bit of current around. Rated at 36W dissipation, 2 pairs could push 25W and they're readily available.
Alternatively, Fairchild seem to put 5200/1943 chips into anything and you can have them in "fullpack" types FJPF5200/1943
rated at 50W which would just make 20W output alone.

Maybe these are a bit slow for what you were thinking of but there are other suitable types with plenty of dissipation in numbers.
As I recall, most "budget" 10-25 watt stereo amps way back in the early 70s used just a pair of TO220s, so doubled up "fullpacks"
seem a fair bet if the dissipation rating is there. It's still a good idea for small amps

[Bigun]

it turns out that I have a dozen of the TO220 version of 5200/1943 in my 'box of bits' so maybe this is a place to start. But at lower currents, important for first watt of Class AB, smaller devices may have better linearity than larger power equivalents ...