[PLUG] Dead machine

[Galen Seitz]

On 06/01/14 08:07, Denis Heidtmann wrote:
...
> My focus now is on the intrusion detection system.  The pins are
> jumpered and always have been.  Suppose the jumper contacts are noisy
> or the circuit that is designed to detect and record an intrusion is
> faulty.  I do not know how the system is designed to react to an
> intrusion event.  My only observation is that occasionally (sometimes
> every few days, other times it will go a month or more between events)
> I get a message at boot and a halt.  It always is corrected by a
> shutdown and power up sequence. (I do not understand how this behavior
> would protect anybody who wants to implement the feature.)  So I
> wonder if when the machine refused to start the intrusion system was
> involved in some way.  Ideas?  (I still want to replace the battery,
> but am a little gun-shy.)

If you haven't already, I would replace the jumper with a different one. 
  I doubt it will help, but it's easy to try.  (BTW, long ago I spent a 
couple of hours one day at Tek chasing an odd problem.  I had a board 
plugged into a backplane, and it wasn't getting the interrupts from 
another board.  The backplane had jumpers to pass along the interrupt 
signals for slots which were not occupied.  After much head scratching, 
I discovered that one of the jumpers was missing the metal contact inside.)

If you really want to debug this, I'd put an ammeter across the two 
intrusion pins and note the current.  I'm just guessing regarding the 
circuit, but I'd expect to see at least 70 uA (3.3V / 47K).  You could 
leave it connected for a while and see how the current changes, 
particularly with temperature.  Again a guess, but I suspect you have a 
marginal solder joint that might improve with increased temperature.

galen
-- 
Galen Seitz
galens at seitzassoc.com