19in rack professional amps

[Nick Sheldon]

Hi Everyone

Has anyone contemplated making a DIY pro-amp? There are some extraordinary specs available now, eg. 9kw continuous RMS in a 3U case (QSC Powerlight 9.0PFC).

Although I'm into valve-amps, an about-to-crawl baby suggests to me that solid-state might be better for the next five years or so. I have a 3U 19in rack, and 250 BD246 PNP bipolar power transistors - what should I build?

Something I'm tempted to try is a "beanstalk" circuit.This permits low voltage transistors to be used in a high voltage circuit, and was once used with TV circuitry, before the production of high voltage devices.

If this works (voltage sharing is not perfect) I'll post the circuit and PCB and make them copyright-free. I'm hoping to get about 1500 watts out of the thing, using more transistors than you'd expect (50-100) owing to the voltage sharing problem.

Nick Sheldon

 

[JoeM]

If you are serious about building a pro-audio amp it may be worth your while looking into class "T" technology. These are cheaper and lighter than even the QSC's! Tripath -who are pioneering the technology are offering an evaluation board for $200USD- 1500Watts @4 Ohms, 0.1%THD.

http://www.tripath.com/0104_eboard.htm

 

[Nick Sheldon]

Astonishing technology! Unfortunately I can't really use this sort of approach with slow, out-of-spec, low voltage bipolar power transistors. My real goal here is producing something quite powerful at virtually nil cost - the BD246s cost 5p each!

Nick Sheldon

 

[thylantyr]

cool

What are you design goals? Voltage, transistors, etc.

 

[Petter]

If you have n and p type bipolars

Perhaps you should consider a Stasis 1. You should be able to get the schematic from Passlabs. I have a Stasis 3 (one third or one fourth or whatever). It is not quite so powerful as you would like, probably more like a few hundred watts, but I really like it!

The good thing about the Stasis 1 for low voltage devices is that one uses complementary pairs everywhere (both top and bottom) and so share voltages. You should also definitely bridge the outpout to double the voltage and quadruple the power (if it does not blow up first)

With lots of transistors, you need a powerful driver stage.

 

[Nick Sheldon]

power for fun

What I'm trying to do is have a recreational alternative to my 3kw and 100-250kw valve amps which form part of my professional work. I love valve stuff, but now that I do it "for real" I need to have something which is totally frivolous as a hobby.

This means the amp has to be quick, cheap and fun to build. Quality is not a consideration! One idea I had was to make a *very* low frequency class D amp using a 555, using it as a driver for genuine sub-woofer (many subs are really just woofers). This improves efficiency, given my slow bipolar transistors, but the device count is still high as the BD246s can't take more than about 50v, which limits the output per paralleled push-pull pair.

I'm tempted to build the thing in a 7in black metal cube for historical reasons!

Nick Sheldon

 

[Nick Sheldon]

pro 19in rack amps

I've been looking at some QSC amps today in a pro audio shop in Norwich, UK. The 2U ones are very small for 3000W RMS, but Powersoft do one (the Digam 7000) which is 7000W RMS in 1U!

The Digam 7000 is 95% efficient, which is the highest figure I've seen for any amp. If you allow an internal case height of 1.5in, then its volume is about 420 cu.in., or in other words very approximately equivalent to a 7 1/2 in. cube. The Carver M400 was a 7in cube and could only manage 350W/channel into 4 ohms!

I've been seduced by all this, and have designed and built a class D circuit which I hope will appear in print in a few months time.

Nick Sheldon

 

[Rolloatm76]

I jsut checked that tripath link, and those pre-assembled boards look pretty good. Does anyone know anything about them? Would it be worth my time to put something like that together or is the tripath not sonically worth it? Thanks for the help!!!!

 

[km]

tripath etc.

hi.

please note that the power specs for the digam(s) are with a crest factor of 4:1 , meaning that the true rms output is like 25% of the stated output power (a fact that are often "forgotten"), if you like check the specs from outline.it , the makers of these amps ...

feel free to mail me for more info on the tripath modules, we made a version with the 0103 on a relatively small board and they seem to work pretty ok.

bye k madsen

ps. class-d with bipolars seem like a contradiction of some sort...