Green Rhino input buffer

[Vivek]

Me no understand!

[11-90-an]

I think those are called "stacked" transistors, which are transistors in parallel.

I never saw this with BJT's before, but it was done with tubes in the LA1A compressor that Nathan(@vigilante397) traced, although only 2 in parallel(not like that buffer, which is 4), he said it really had a sonic difference without the extra tube stage.

I could be wrong, though... 

[Fancy Lime]

Apparently, this sort of thing is often done with JFETs as input stages of high end phono preamps, where it is supposed to reduce noise. Each doubling of the number of devices supposedly reduces one kind of noise by half. I don't know what kind of noise though, or how the theory works. Either way probably overkill and/or gimmick in a guitar application with modern transistors and opamps available.

Cheers,
Andy

[antonis]

Each doubling of the number of devices supposedly reduces one kind of noise by half. I don't know what kind of noise though, or how the theory works.

Βy 0.52 (reduce X 1.9), to be more precise.. 
It's current noise (total Collector current is shared among individual transistors) but , IMHO, main target of particular configuration is to "stabilize" h_FE variations, due to Collector current ones, preventing distortion..

I never saw this with BJT's before

You probaly never saw any power amp output stage..
(with or without low value "balast" Emitter resistors..)

[iainpunk]

i dissected a circuit of a bad amp design for school once with the same configuration but with 3055's, after careful current measurements, i found out that only one transistor conducted, the rest didn't in any significant way, it was the one that had the lowest forward BE voltage that did 99.9% of the work. its basically the same with diodes in parallel, but that one i had to find out the hard way. (rectifier in a scooter engine needed a 10A diode, so i put 2 5A's in parallel without current share resistors, they both burned out)

with FET's this is possible for higher current handling and lower noise because they are basically variable resistors, where the BJT's are current amplifiers, and only one will conduct since it has the lowest forward voltage.

as for why they did it in the green rhino? i think its due to marketing or something like that

cheers, Iain

[Vivek]

i dissected a circuit of a bad amp design for school once with the same configuration but with 3055's, after careful current measurements, i found out that only one transistor conducted, the rest didn't in any significant way

cheers, Iain

I think they need "load balancing resistors"

Same situation as using multiple diodes in parallel to increase max current flow

[11-90-an]

i think its due to marketing or something like that

if I may ask... how is there profit in adding extra "unnecessary" trannies? wouldn't that just incite more cost?(unless something similar to government corruption, which I can understand  )

BTW, those transistors are actually individual, independent 3 pin trannies, no 4-trannies-in-a-package here.. (i think) so there would be extra board space needed to accommodate the 3 extra..

[iainpunk]

i dissected a circuit of a bad amp design for school once with the same configuration but with 3055's, after careful current measurements, i found out that only one transistor conducted, the rest didn't in any significant way

cheers, Iain

I think they need "load balancing resistors"

Same situation as using multiple diodes in parallel to increase max current flow

yes, we later added 0.125 ohm resistors to keep all 3 of the trannies from blowing up every time. still it was a bad amp because it was class A inductor loaded, without feedback, it just sounded awful for music but i brought my bass to school and it was awesome for bass, because it was clean with a bunch of harmonic sweetness!! i later build a amp inspired by that design with a DTS410 transistor instead of the 3055's and only 12V instead of 32V, and a bunch of other stuff changed to make it work with a microwave transformer as the inductor.

cheers, Iain

[amptramp]

Apparently, this sort of thing is often done with JFETs as input stages of high end phono preamps, where it is supposed to reduce noise. Each doubling of the number of devices supposedly reduces one kind of noise by half. I don't know what kind of noise though, or how the theory works. Either way probably overkill and/or gimmick in a guitar application with modern transistors and opamps available.

Cheers,
Andy

If you double the number of devices and run the same current through each device, the noise goes to 1/(SQRT)2 or 0.707 of the original noise.  The noise in each device is uncorrelated so the noise would be the square root of the sum of the squares.  Four devices as shown would give you half the noise if the current splits evenly between them.  But this requires emitter resistors to share the current evenly and the noise of these resistors has to be included in square root of the sum of squares form.

I have an Astron RS-35A 13.8 volt 35 amp power supply that uses four 2N3055's for the output and 0.05 ohm current sharing resistors in series with each emitter.  Designers do this all the time.

[PRR]

...make it work with a microwave transformer as the inductor....

https://www.diyaudio.com/forums/pass-labs/313649-build-mofo.html

The microwave transformer may not appear until page 69 or somesuch.

[iainpunk]

...make it work with a microwave transformer as the inductor....

https://www.diyaudio.com/forums/pass-labs/313649-build-mofo.html

The microwave transformer may not appear until page 69 or somesuch.

yes, very interesting thread to read, im positive the core isn't getting saturated, its rated for 240V and im running only 12V.
also, the transformer is not the main transformer bur and autotransformer that converted the 240V to 110V, which then went straight to the main transformer. its a huge hunk of steel, i weighs more that all the other parts of the amp combined.

cheers, Iain

[Rob Strand]

It's seems pointless.  Parallel transistors are often paralleled to decrease noise in low impedance circuits.  It works because transistors have a internal resistance in series with the base which adds noise and paralleling parts effectively reduces that resistance.  I doubt it's going to reduce the noise especially when there's the 1k resistor there, or even worse,  the guitar impedance.