[SunHELP] sparc 10 crash - help needed

[sunhelp at sunhelp.org]

You can't assume it's a bad boot prom. Also, it likely is "not" a bad prom. Do
this:

pull out the frame buffer and tip to the system again without the keyboard
connected, see if you get anything. If you have another sparc 10, take an
nvram that has the diag-switch? set to true and pop it in there. In the older
systems (ss1, ss2, ipc, ipx, etc..) you could just  remove the NVRAM and it
would go thru POST but not the sun4m machines.

If you can't get POST, the first thing you should swap out are the CPU modules
and memory. Some might think power supply first and if you have a VOM by all
means check that first but because the CPU and memory are pluggables, you can
check them very quickly.

If you don't have other memory, reduce the memory sticks to one and try
swapping with another if the system is still not working. For CPU modules,
hopefully you have extras or two in the same chassis, in which case just
follow the same type of procedure as I describe above for memory.

If the system is completely dead all you can do at that point is try to assess
how dead it is. You can put a scope probe on the boot prom to see if it's
getting a clock signal, also you will want to check the 5 and 12 volt lines
from the power supply and make sure they aren't sagging ( both should be
really close to their respective voltage).

I have repaired many ss10 main boards that are completely dead. They don't
contain much glue logic (small miscellaneous chips like bi-directional
transcievers, PALS, etc..) most of the circuitry is contained in a small
number of monolithic ASICS.

The two chips that are usually culprits are the SEC ( SBUS to EBUS Controller
) and the DMA2 chip. They are the two black surface mount chips in the middle
of the main board. The problem is if one of these things goes bad and it
usually ties one or more of the data or address lines low causing the system
to be completely dead. Your lucking if you even see anything from the boot
prom.

So unless you have a PACE soldering station or some chipquick and some spare
chips laying around, you better just stick with divide and conquer among the
pluggables. Swap the MBUS modules and memory and if it's still broken, throw
it away or send it to be fixed.

If you really think it's worth repair, you can send it to Base 2 Technologies,
Kent, WA ( I used to work there, they have been repairing ss10s for years now
)

1-800-756-4766.

Regards,

--Buddy






-----Original Message-----
From: Arthur Wouk [mailto:awouk at ra.nilenet.com]
Sent: Tuesday, July 15, 2003 3:00 PM
To: sunhelp at sunhelp.org
Subject: Re: [SunHELP] sparc 10 crash - help needed


here is my first post. the current state is described below.
:
:my sparc 10 crashed today. the main symptom is a lack of symptoms.
:
:i removed the keyboard and monitor, attached them to another sparcsattion
:and it boots just fine.
:
:if i tip to the dead sparc 10, i connect but get no response when i
:power reset.
:
:when i reattach the monitor and keyboard, and power on with stop-d
:keys held down, the leds on the keyboard go on and off two times, then
:everything halts with only the caps lock key illuminated. thus the
:POST does not run. the monitor's signal light keeps flashing - it is a
:GM 20D10. there is no signal to the monitor apparently.
:
:is this a power supply failure, a boot prom failure or an nvram
:failure? any clues as to how to test?
:

i reported the crash a few days ago. i now have the maintenance
manual, and it confirms that the power supply is fine: both test
criteria are satisfied: the led on front of the case  lights up, and
the keyboard emits its desired aqueak.

however, there is no output either to the monitor or to serial port A
when i power on while holding down stop-d. the heyboard LEDs go on and
off twice, and then i am left with the cap locks led on permanently,
and no other signs of life.

this leads me to think that the boot PROM is fried. would a fried boot
prom cause a crash in mid-operation? a fried boot PROM would fail in a
manner similar to  what i have described.

i do not believe this is a problem with the cpu - it is not involved
in the POST, only the boot PROM is. it cannot be a problem with the
drive, which starts up normally (which i can feel with my hand on
power reset) and is not involved in the POST.  it can't be the frame
buffer since i get the same result coming in through serial port A.

so i am left with the motherboard. how do you test a motherboard?
i am guessing that replacing the boot PROM is the only feasible test.
if that doesn't fix it, i gues the MN is shot.

does anyone have any other analysis or insight?
_______________________________________________
SunHELP maillist  -  SunHELP at sunhelp.org
http://www.sunhelp.org/mailman/listinfo/sunhelp