The 2N2/256-BSCP [was: Homebrew Drum Computer]

[Allison]

>
>Subject: Re: The 2N2/256-BSCP [was: Homebrew Drum Computer]
>   From: ard at p850ug1.demon.co.uk (Tony Duell)
>   Date: Fri, 21 Dec 2007 23:26:06 +0000 (GMT)
>     To: cctalk at classiccmp.org
>
>> > Do all transsitors have to be of the same flavour, or could I use TUNs
>> > and TUPs?
>> 
>> I'd never encountered the TUN & TUP TLAs before, so I looked them up:  
>> Transistor, Universal NPN/PNP.   Neat!
>
>I suspect the acronyms were invented by Elektor magazine over 30 eyars 
>ago. Although they're not still sued in said magazine, theu're handy :-)
>
>There's also DUG and DUS (Diode, Universal, Germanium/Silcon). 
>
>
>> 
>> And any TUN or TUP will do, provided a device of comparable spec was  
>> reasonably available to hobbyists in the mid sixties.  For my first  
>> efforts, at least, I'll be using the PN2222 in modern TO-92 plastic 

Old TUNs and TUPs are in the 2n400 and 2n130x series of germainum computer
transistors.  Next generation was the silicon 2n70x series and their 
plastic cased counterparts.

PN2222 is a fair devices to use for that.  Far better than early 60s devices 
and being silicon far more stable too. Those should be easy to find for 
under 4cents at the 100 peice level.

>Sure. I tened to use 2N3904/2N3906 to excess round here. I don't think 
>they're quite the same spec as a TUN.TUP, but they're pretty universal in 
>their own right :-). 

Good for both switching and RF too.  Also the metal can TO18 parts the 
2n2222A and 2n2905 when you need a higher Ic and of course the 2n2219 
and 2n2907 in TO5 for the next power level up.  All been around a long 
time and dirt cheap.  

> >> packages.  I'm not sure when the TO-92 package was introduced, but  
>> there were definitely plastic package transistors available then.   
>> SOTs are right out, though :)

The TI silicon TIS9x transistor series was popular.  

>Given that there's a limit on the number of transistors, I don't see what 
>advantage using SOT23s would give you. OK, you could make it smaller, 
>maybe clock it faster. But the change in package wouldn't allow a totally 
>different architecture. So I don't see any particular reason to ban them, 
>since the same design could be built with compoents in 1960's packages.

The 2n3904/6 is about 10X or more faster than the 1960s germainium
parts and still better than late 60s parts. Most of the 1960s wer the 
MADT germainium in it's odd metal (to39?) can or TO5 sized parts so 
most devices like TO18 or TO92 will be smaller.


Allison

>-tony