The "Sort-of" H8 (WAS:Re: [sebhc] My H8 Systems)

[Christian Fandt]

Upon the date 04:20 AM 5/29/04 +0000, Steven Parker said something like:
    -- snip --
>On my own status:
>
>Thanks to Walt, all the power swtiches work now.  :-)  I also replaced the 
>blown cap on the 3rd H8's CPU board and now it comes up .. sortof.
>Sometimes it comes up and looks normal, other times it seems to come up 
>running in user mode (the pc is displayed, and changing).  Even when it 
>comes up and looks normal something else is odd.  If I try the memory test 
>it crashes with lights out and continuous tone.  I'll have to do some more 
>testing on that one.

Since there was a blown cap on the CPU board, there may be another which is 
exhibiting a too high ESR and as a result, is not doing its job. That is to 
say, electrical noise filtering or ripple suppression on the particular +5V 
line going to whatever chip or set of chips the suspected high ESR cap 
serves to filter is not being done as well as designed. That noise probably 
makes the logic in that area unstable. Figure as much as a quarter volt 
ripple riding on the DC supply could cause logic timing to be affected. The 
DC supply should be rather clean. Check the voltage levels with a meter.

A high ESR (Effective Series Resistance) can be caused by the electrolyte 
literally drying out of an electrolytic capacitor. The cap's dielectric 
breaks down under voltage, things heat up, and pow! You should see what 
that is like with a large main filter cap when that happens! Something you 
will NOT forget if you were there to witness it.

If you have an oscilloscope you should check the +5V supply all the way 
from the main filter caps in the power supply and all the way through the 
CPU board. Check other boards too if the main supply caps are okay. You 
will be looking for too high ripple or other trash noise on the supply 
line. Should be, I figure, ten to twenty millivolts or so maximum ripple 
for a nice clean DC supply on something like this. Preferably less.

Just a strong suspicion based upon past experiences . . .

Good luck,

Chris F.

NNNN

Christian Fandt,    Electronic/Electrical Historian
Jamestown, NY  USA      cfandt at netsync.net
         Member of Antique Wireless Association
         URL: http://www.antiquewireless.org/ 

--
Delivered by the SEBHC Mailing List