Using Many Small Signal Transistors Paralleled in place of one output Transistor

Jump to Latest
#21
ChocoHolic said:
....you mean that you had issues with the storage time
of your BJTs... ?
Ok, storage time has heavy tolerances, but I would not expect
trouble with this when you use a BC327 in linear mode (not switched to saturation) at HiFi frequencies....

any scope pics or description of the effects that you had observed?
Click to expand...
Storage time or whatever you want to call it, i am not pro at semiconductor physics. (Actually hate it, failed in semiconductor physics exams continuously) 😀
Note that I was talking about frequencies up to 10Mhz. lower than 100khz was just fine with or without feedback.
I can supply some scope pics on next week. Base drive current was rather strange looking when monitored over 100r resistor
 
#24
Mzzj and ChocoHolic,

"Storage time or whatever you want to call it, i am not pro at semiconductor physics. (Actually hate it, failed in semiconductor physics exams continuously)"

"semiconductor physics are also not my favourite."

Ha. I used to teach this stuff. Loved it, and viscous fluid dynamics.

😀 😀

My new commercial 150W MOSFET R2R amp uses BC546/556 as drivers and reverse biassing to remove charge and speed switching!
:cheerful: :cheerful:
 
#25
There was at least a commercial amplifier having an output made of dozens of medium power transistors, in TO220 package or similar with no heatsink. It was about fifteen years ago; I don't remeber the name except it was european and had no success at all.

Using a huge amount of devices for an output stage, problems can arrise because of parasitic capacitors and inductors.

~~~~~~~~ Forr

§§§
 
#26
amplifierguru said:
Mzzj and ChockoHolic,


Ha. I used to teach this stuff. Loved it, and viscous fluid dynamics.


My new commercial 150W MOSFET R2R amp uses BC546/556 as drivers and reverse biassing to remove charge and speed switching!
Click to expand...
Maybe you are not as lousy teacher as i had. I think my professor hates semiconductor physics or keeping lectures even more than me 😀

150W R2R amp, can you reveal some more light on this? sounds intresting
 
#29
mzzj,

do you have any schematic?
One problem can also be the Miller capacitance at such high frequencies you are using, also depending on the source impedance.


Greg,

best luck to you and your new products!


Cheers Michael 🙂
 
#30
This approach seems to be very good in view to the sonic quality. Thule's integrated amp IA120 work with a bunch of paralleled TO-126 (unfortunately I haven't the schematic).
Mr. Nelson Pass does basicly the same by this project:
http://www.diyaudio.com/forums/pass-labs/150048-beast-thousand-jfets.html

Who knows more commercial brands with a bunch of small signal transistors in the output power stage instead only one BjT power device in each half?
 
#34
Cyberschelm said:
Perhaps not exactly the same, but a few years ago meier audio (http://www.meier-audio.homepage.t-online.de/) used to sell a power amplifier with like 40 accurate small signal opamps paralelled in the power stage.
Click to expand...
Thank you for this advice - this could be also an approach, which promises good results. I will ask there, which model it was.
Was it the amp showed about
analog_sa said:
This idea is not even crazy enough.
Click to expand...
???

idlehands said:
How about a 10 Toshiba 2SA1930/2SC5171 in stead of a gazillion little ones? these things are fast like smaller transistors but with 10 times the current capacity.
Click to expand...
Also good, but I would prefer non isolating outline like 2SA968/2SC2238 about
http://cdn.head-fi.org/e/ef/ef5e9c88_IMG_1444.JPG
for more rugged applications like Class-A.
The thermal transition resistance by the ISOWATT-220 (SC-67) outline about
http://bg-electronics.de/catalog/images/2SA1930 PMC.jpg
is much more larger than by usual TO-220.
forr said:
There was at least a commercial amplifier having an output made of dozens of medium power transistors, in TO220 package or similar with no heatsink. It was about fifteen years ago; I don't remeber the name except it was european and had no success at all.

Using a huge amount of devices for an output stage, problems can arrise because of parasitic capacitors and inductors.
~~~~~~~~ Forr §§§
Click to expand...
Can you try to find out this with help of a forum in your language like this:
 
Last edited:
#35
forr said:
There was at least a commercial amplifier having an output made of dozens of medium power transistors, in TO220 package or similar with no heatsink. It was about fifteen years ago; I don't remeber the name except it was european and had no success at all.

Using a huge amount of devices for an output stage, problems can arrise because of parasitic capacitors and inductors.

~~~~~~~~ Forr

§§§
Click to expand...

maybe this brand: GTE Audio Trinity Class A
TRINITY Electronic Design GmbH
High End 2009. Hifi Anlage. GTE Trinity. Lansche Audio. | Audiodoo.de Hifi-Blog
http://www.diyaudio.com/forums/pass...audio-power-amplifiers-100-000-600-000-a.html (post #6)
http://www.diyaudio.com/forums/plan...la-information-about-efficiency-wanted-2.html (post #17+#18)

The amp about
http://www.diyaudio.com/forums/soli...ke-pure-classa-se-despite-low-idle-power.html
uses also a lot of power devices in parallel mode.

Here various diy projects:
http://www.elektroda.pl/rtvforum/topic985463-0.html
 
Last edited:
#36
This is the Momentum output stage, there is nothing special with the numbers of transistors. It has local decoupling dought. The Sanken BJT transistors are low cost.
 

Attachments

  • D.jpg
    D.jpg
    208.1 KB · Views: 1,006
#38
Actually every LAPT is basically a collection of small transistors in parallel with their own emitter resistors, on the same silicon die. Which is precisely why they can go up to and over 50-70MHz and have large SOA - but being integrated on the die, have less stray capacitance and inductance.
Also, most MOSFETs are cell-based designs, albeot without source resistors (as efficiency is the name of the game), in this case for reasons of efficient scaling - basically change te die size and hence the number of cells, but the behaviour is still like a collection of many small MOSFETs in parallel.
Before the dvent of LAPTs using smaller devices in parallel was the only way to get high speed (however this also depends on the actual technology used for the small devices). Also, it's a case of 'nneds must' - for instance, what if you wanted a JFET with Idss=500mA? There are no current production units like that, hence, use what you have and parallel it 🙂
 
#39
tommy1000 said:
This is the Momentum output stage, there is nothing special with the numbers of transistors. It has local decoupling dought. The Sanken BJT transistors are low cost.
Click to expand...
Hi, tommy1000 thanks the picture.

Sanken bjt has good reputation at their good sounded amps from the past like luxman amps, sony, etc. And my 1st good sounded amp using sanken bjt both driver and ops, but most the driver blow in ~3years and changing the driver to toshiba make it sounds worse.
 
You must log in or register to reply here.
Top Bottom