New pcb design for S-100 prototype board available - test-drb@ccmp.vtda.org

3 Jun 2007

...

Subject: Re: New pcb design for S-100 prototype board available
   From: Dan <ragooman at comcast.net>
   Date: Sun, 03 Jun 2007 00:33:41 -0400
     To: General Discussion: On-Topic Posts Only <cctech at classiccmp.org>

Roy,

Thanks for the input. I was thinking about including one with vertical 
rows and one with horizontal rows. That way you can have some flexibility.

One of the recent options for the onboard power supply was to use a low 
dropout voltage regulator. This allowed boards to be installed in S-100 
systems which contained either the original S-100 specs or the IEEE-696 
specs. The latter have a lower power supply voltage on the 
backplane(translates to lower power consumption) which an ordinary 7805 
voltage regulator can't accept. These require at least a 10+ volts on 
the backplane just to get a regulated 5V output. The LO voltage 
regulators need only about 7V input to get a regulated 5v output. 
Wrong data.  The 7805/lm309/lm323 reguire only 8V.  All of the older 
3 terminal regulators must have a 2.7V differential to regulate.

The yabut, most of the S100 supplies were so poorly filtered that
at any load the ripple voltage was excessive and didn't meet the minimum
of 8V at the bottom of the ripple trough.  A good example of that
was the early Altair 8800 before MITS upgraded the transformer.
With a 8A load the DC votage sat at 8.3V but the lowest voltage was
7.1V due to the AC ripple. Needless to say that ripple fould it's 
way to the 5V rail as the regualtor didn't have enough head room to 
regulate.  The fix was a higher voltage transformer or much heaftier
 filter caps or both.

...

The original backplanes were only 2 layer and couldn't handle the 
current capacity. The local(distributed) voltage regulators were the 
only option. You need at least a 4 or 6 layer backplane with a separate 
copper layer for each voltage and then a layer for ground to have enough 
copper for all that current. 
Seals made a decent proto card.  Horizontally gridded for power, up to 
four regulator pads one of which was for +12 (7812), plated though holes
both hole per pad and for power gridding (two sides).

Allison

...
 =Dan

[ My Corner of Cyberspace http://ragooman.home.comcast.net/ ]

Roy J. Tellason wrote:
> Hmm,  good question!
>
> I have currently only two S-100 systems,  a Cromemco System 3 with problems in 
> the PersSci drive.  The floppy that was in the drive when I got it had been 
> in there and run so long that you could see clearly through the track zero 
> location.  :-)  I have some data on this stuff somewhere,  and it appears 
> that they're using incandescent bulbs for such stuff as index sensors and 
> such?  And the Imsai,  here.
>
> Got a Vector S-100 backplane too,  that I've started to build,  only I lack a 
> few parts.  Most importantly the S-100 connectors themselves.  :-(
>
> I also have a couple of prototyping cards,  I *think* they're Vector as well,  
> but haven't done anything with 'em yet to come up with a preference.  I guess

> vertical rows makes more sense in terms of air flow for heat dissipation?  
> That for the heatsinks for sure,  though a lot of systems I saw mention of 
> later on in the popular period for S-100 seemed to be inclined to put a 
> regulated switching power supply in place and simply jumper across the 
> regulator positions.  I dunno,  to me the distributed approach always made a 
> lot of sense.
>
>    