🚀 go-pugleaf

On Sun, 30 Jan 2000 19:08:53 +0000, Mike Tomlinson
<mike@jasper.freeserve.co.uk> wrote:

>In article <j3t89s88tmqegcp4luu8bmh18l3mmbsc1v@4ax.com>, Harry Broomhall
><haeb@pc60.easynet.co.uk> writes
>
>>   I don't agree with this.  My POST card has fairly often pointed me
>>at the problem.  Typical items were 'faulty cache' on older Socket 7
>>mobos,
>
>'faulty cache' could also cover faulty processor, faulty cache
>controller, and faulty or incorrectly fitted memory.

  Actualy no.  The cache test that the POST does at this point really
tests just the cache. Earlier tests for minimum memory have the cache
switched off.  While a faulty cache controller is possible, my
experience is that a fail at this point on such boards means either
the BIOS has been set to enable cache when there wasn't any, or one or
more chips neads pushing home properly.
 
>
>> and that hardy perenial, VGA card not working.
>
>wouldn't you in that case hear the 'VGA failure' beep sequence from the
>BIOS's POST?

  Again no.  The beep sequence has been much abused by the BIOS
makers. I find that these days there is an all-purpose 'it's f***ed
beep' without any effort to distinguish the actual error.

>
>>   The other main use is to see if the machine starts at all.  Without
>>a POST card it is very hard to know if the machine has failed to boot
>>because of a faulty VGA card or cache or what ever, or just hasn't
>>started.
>
>Agreed, but it's possible to make an intelligent guess from other clues.
>For instance, you can feel the temperature of the CPU with the back of a
>finger.  An excessively hot or cold CPU indicates a problem.  I've had
>several boards with secondary power failures (the on-board circuitry
>used to produce the 2.2 -to- 3.2 volt core CPU voltage)

   There are a lot of possibles here of course.  Again, my experience
is to have a 'dead' machine brought to me.  In such cases, where it
has been transported, the temperature is unreliable; it just hasn't
been on long enough.


[SNIP]
>
>It's worth noting that the meaning of the number sent to port 80
>differed from BIOS maker to BIOS maker, and so a look-up list is needed
>to see what a certain figure means.

  Indeed.

>
>>   In the early stages of POST, before the VGA card is woken up, a
>>failure of a test will usualy put the processor into either a HALT or
>>a tight loop.  At this point the value displayed can be looked up in a
>>table to get a guess as to what is wrong.
>
>I will admit, POST cards can be useful.  My argument was that knowing
>that, for example, the DMA controller has failed is of little use when
>the easiest and quickest route to curing that failure is motherboard
>replacement.

   This is only partly true.  When I find such a failure (actualy
suprisingly rare, modern electronics is *amazingly* robust) then I
know not to waste time with extensive replacements.

   In the end any diagnostic tool merely is a time saver.  Yes, you
can do it by other means, but the time saved is, IMHO, worth it.

   Regards,
        Harry.