-=> George White said to Roy J. Tellason
 -=> about "VEHICLE LED'S" on 10-22-03  19:41.....

 GW> There are some aweful designs out there. A magazine published one
 GW> once and one of my collegues sent them a letter pointing out all the
 GW> problems with the design. They wrote back saying the one they built
 GW> worked OK for them so he (in fact it was we) were in error, there were
 GW> no problems...

Of course. It's a matter of pride that once something is
published, it has to be correct, even if you have to alter the
data to prove it.

 GW> In fact there were very significant problems, but it's too long ago
 GW> for me to remember the details. Things like insufficient gain if the
 GW> transistors were at the low end of their gain range, poor choice of
 GW> zener voltage, and more beside...

NEVER test with a range of components. Never test under a range
of conditions. And NEVER admit you might have made a mistake...

CASE 1: a professional maker of audio mixing desks was asked to
supply some desks to the Australian Broadcasting Corporation
where I was working some years ago. One of the specifications was
that they would operate satisfactorily under conditions of 15%
overvolatge of the mains supply and elevated ambient temperatures
up to (I think) 35 deg C.

When the desk was checked it met these conditions separately but
not when subjected to them at the same time - the power supplies
would go out of regulation with all sorts of horrible effects.

The manufacturer claimed that it was impossible to meet BOTH of
these specs at the same time due to limitations of the integrated
voltage regulator chips he was using, I think they were TO3 5A
adjustable, possibly LM338 or similar.

When I heard about it I showed our engineer the applications
sheet for the LM723 regulator with external pass transistors. He
sent a copy of the sheet to the manufacturers of the desk and
they rebuilt the power supplies that way... then they had the
temerity to try to charge us extra "for modifications after
manufacture".

CASE 2: I had a 386 motherboard given to me several years ago
with a failed onboard NiCad battery. The symptom was that the
clock would stop as soon as power was switched off, although it
didn't lose its settings. And the CMOS would stay up alright.

The battery measured 2.6v with poer off and about 2.8v with power
on. It looked to me as though one cell was shorted.

I removed the battery and powered up the computer, and I found
only about 3 volts at the battery terminals into a moving coil
voltmeter, approximate load about 60uA.

Checking the circuit I found the 5v rail was feeding through (1)
a blocking diode (2) a 1k limiting resistor and (3) another
device that looked like a diode and acted like a 1.5v zener.
Total voltage drop at the 60uA current was 2v!

There was ABSOLUTELY NO WAY that the battery could ever charge up
to 3.6 volts!

Later tests on another board led me to believe that a new NiCad
would hold up the CMOS and the clock for over two years with use
of the computer for about 1-2 hrs a day, and a lot longer with
heavier use.

I suspect a lot of computers went out with this fault, which
wouldn't have been found during the normal "life' of the computer
as most people would upgrade anyway as soon as it started giving
trouble.

My method of "fixing" the fault was to solder a 1/8w 4700 ohm
resistor across the zener, so that the battery received charge to
beyond its terminal voltage. I measured it after about two months
and the battery was holding at about 3.9 volts, a little high but
IMO quite reasonable.

I wrote about it to one of our electronics magazines. They
published my letter but commented that it could not have been a
common problem otherwise there would have been more reports of
the fault.

Well, I wonder how many people would have noticed the effect and
whether anyone would have worked out the cause.

With upgrades of motherboards as common as they were at that
time, and the fact that no-one but an idiot like me would attempt
to trace out the circuit, it's not so surprising that no-one else
had found it.

Nope, even the "experts" can get it wrong.

From Greg Mayman, in beautiful Adelaide, South Australia
   "Queen City of The South"    34:55 S  138:36 E

... I'd love to, but my palm reader advised against it.
___ Blue Wave/386 v2.30

--- FLAME v2.0/b