well . . . back when dinosaurs roamed the earth, and IMSAI was a major
computer maker, the 25.2VCT was about as common as its companion, the 6.3V
filament heater (starting to sound familiar?) and there were a few supplies
which got away with that. If you filtered well and used only one diode at a
time, you still had enough headroom to allow for 2.2 forward volts in the
regulator.
I've got switchers in the basement which have adjustable voltages for three
(5-10Vdc, +/-12-18Vdc). The previous statement that switchers like a
constant load is quite correct, though and what is meant there is that the
switchers don't like capacitive loading, nor do they like loads which are
intermittently on.
S-100 boxes sometimes housed floppy and hard drives, though, and if it was
only the 5-1/4" types, and floppies only, the +12 saw its worst nightmare.
It saw nearly fully on, vs nearly fully off, since one floppy took more +12
than all the RS232 ports.
I don't know where they're going to end up, but I believe that the power
supply thing will have toox have quite a bit more attention than it's gotten
so far.
My advice to anyone wishing to sell hardware in the quantities they're
likely to sell, would be to build a kit, no assembled units, find a catalog
standard switcher to sell them, and use the fact that it's all just spare
parts to bypass the FCC reg's. Sell the box top, sell the box bottom, the
front, the rear, but no fully assembled boxes. Sell the front panel as a
replacement part, too. That way there's no enclosure to require approval.
What's more, it will make the phone support burden lighter.
If they want a fully assembled and tested unit, tell them to buy a PC.
Dick
-----Original Message-----
From: Allison J Parent <allisonp_at_world.std.com>
To: Discussion re-collecting of classic computers
<classiccmp_at_u.washington.edu>
Date: Tuesday, August 17, 1999 8:00 PM
Subject: Re: imsai 2
><Anyway, why does it have to be one transformer? Why not a 12-0-12 for the
><16V lines and a 6V one for the 8V line (those should produce DC voltages
><within the range of any normal S100 board regulator). Such transformers
><are trivial to obtain.
>
>Then you'd end up with the piece of crap altair supply...
>
><There is a myth doing the rounds that PSUs -- even simple linear PSUs --
><are impossible to design. No idea where it came from...
>
>Because any error means all your silicon is junk. Also if you make a
>small error the core you using could get quite hot. Or maybe the ripple
>from switching exceeds the reasonable level by say a volt or two. Then
>there are the high current ground loops that cause instability at
>something approaching max load or minimum load. The worst is when you
>forget the chopper side has 300V on it, S*!^^ d_at_**. The number of
>errors that can be made are far higher. Then again having designed a
>few, once you've done it you learn... mostly everything they taught you
>was far from enough. that and fixing all those that were really not so
>well designed.
>
>Allison
>
Received on Tue Aug 17 1999 - 22:14:00 BST