imsai 2

From: Richard Erlacher <edick_at_idcomm.com>
Date: Tue Aug 17 18:16:31 1999

What you say is true, but the cases you bring up are not the same.

The teensy regulator providing power for a processor is designed with pretty
solid knowledge of the behavior of the load. The + 8 Volt supply in an
S-100 box is a completely different thing, with loads which vary widely from
one application to another, and, for that matter from one day to another.
If you reversed the bias on the little regulator on a processor, the current
stored in the processor would be negligible. If you have a memory circuit
on an S-100 card, you probably have forty small capacitors and half a dozen
larger ones (e.g. 33 uF) which would be sourcing current to the 7805
regulator(s) while the input now sources current to the disk drives or
whatever other low-impedance loads run from that supply.

I'm with you about the use of switchers which save on energy and reduce
waste from dissipation. However, the proposed "booster" hasn't been thought
through yet. They're not very cost-effective, nor do they offer much power
per buck.

The problem with linear supplies is not their intolerance of high voltage
inputs, but rather, my intolerance of their high temperature when operating
at a high input/output ratio, as the regulator dissipates it all as heat.
That's not likely to happen here, however.

Dick



-----Original Message-----
From: jpero_at_cgocable.net <jpero_at_cgocable.net>
To: Discussion re-collecting of classic computers
<classiccmp_at_u.washington.edu>
Date: Tuesday, August 17, 1999 4:50 PM
Subject: Re: imsai 2


>> Date: Tue, 17 Aug 1999 16:35:15 -0600
>> Reply-to: classiccmp_at_u.washington.edu
>> From: "Richard Erlacher" <edick_at_idcomm.com>
>> To: "Discussion re-collecting of classic computers"
<classiccmp_at_u.washington.edu>
>> Subject: Re: imsai 2
>> X-To: <classiccmp_at_u.washington.edu>
>
>> I'm not sure that "jumpering across" the 7805 will be healthy for the
7805.
>> Of course, if the input doesn't drop below the output during shutdown
this
>> may not be a concern. However, if there's a significant off-board load
on
>> the input switching supply, and significant capacitance (bypass, for
>> example) on the board, it's likely to cause problems.
>>
>> Dick
>
>Hi,
>
>Actually that is not harmful thing. In fact, nearly all peecees of
>any brands and generic of any quality short out the 3.45V regulator
>via input and output to get 5V by either a jumper or a low
>resistance/high current capable FET in (automatic detection type like
>GX4, SP3 and such by Asus for 486). I'm sure that is same story for
>the split or single voltage supply modes for Pentium class boards.
>
>One thing about double PSUing from one to another voltages is bit too
>much to me from reliablity and cost viewpoints. Feeding 5V and
>12V direct to bus is assumed that designer upsize the conductors
>enough to keep voltage drops low.
>
>THERE is decent PSUs that can be had in any voltages in multi outputs
>from one box and still be switcher. I'm all for switchers for
>efficieny and less heat output of quality built.
>
>Most linear regulators are happy on input voltages up to 30V BTW.
>
>Wizard
Received on Tue Aug 17 1999 - 18:16:31 BST