asus p2b-s adaptec 7890 problem?

  • Thread starter Thread starter `sNaKe
  • Start date Start date
S

`sNaKe

Hi all,

I have a problem with my p2b-s rev. 1.03 mainboard. I updated the bios
to 1014.003 because I installed a 40GB IDE harddisk. but right after
the flash the onboard scsi controller is somehow deactivated. there is
no "press ctrl+a to..." message at startup anymore. the device itself
and its attached drives isn't present under windows, too. I saved the
old bios (1009) before the flash and I flashed it back, but the
controller is still not available. the scsi enable/disable jumper on
the board is enabled and scsi bios is enabled within "pci & pnp
config".

used hardware:
p3-500
256mb sdram
geforce ddr
350w power supply

please help!
 
Hi all,

I have a problem with my p2b-s rev. 1.03 mainboard. I updated the bios
to 1014.003 because I installed a 40GB IDE harddisk. but right after
the flash the onboard scsi controller is somehow deactivated. there is
no "press ctrl+a to..." message at startup anymore. the device itself
and its attached drives isn't present under windows, too. I saved the
old bios (1009) before the flash and I flashed it back, but the
controller is still not available. the scsi enable/disable jumper on
the board is enabled and scsi bios is enabled within "pci & pnp
config".

used hardware:
p3-500
256mb sdram
geforce ddr
350w power supply

please help!
Click to expand...

After a BIOS flash, try "Load Setup Defaults" in the BIOS.

If still no positive results, try the "Clear CMOS" procedure,
_unplugging_ the computer before doing the procedure. It is
in the jumper section of the manual - see "Clear real time clock
(RTC)" for details. Apparently the board uses two solder points
and you short them with the blade of a slot-type screwdriver tip.

HTH,
Paul
 
`sNaKe said:
Hi all,

I have a problem with my p2b-s rev. 1.03 mainboard. I updated the bios
to 1014.003 because I installed a 40GB IDE harddisk. but right after
the flash the onboard scsi controller is somehow deactivated. there is
no "press ctrl+a to..." message at startup anymore. the device itself
and its attached drives isn't present under windows, too. I saved the
old bios (1009) before the flash and I flashed it back, but the
controller is still not available. the scsi enable/disable jumper on
the board is enabled and scsi bios is enabled within "pci & pnp
config".

used hardware:
p3-500
256mb sdram
geforce ddr
350w power supply

please help!
Click to expand...
I'm afraid you have probably destroyed the BIOS on the Adaptec card. There
used to be a warning on the Adaptec site, to remove some of their cards
from particular motherboards, before attempting a motherboard BIOS flash,
because on some motherboards, the addresses used, are the same as those
used by the cards...

Best Wishes
 
`sNaKe said:
Hi all,

I have a problem with my p2b-s rev. 1.03 mainboard. I updated the bios
to 1014.003 because I installed a 40GB IDE harddisk. but right after
the flash the onboard scsi controller is somehow deactivated. there is
no "press ctrl+a to..." message at startup anymore. the device itself
and its attached drives isn't present under windows, too. I saved the
old bios (1009) before the flash and I flashed it back, but the
controller is still not available. the scsi enable/disable jumper on
the board is enabled and scsi bios is enabled within "pci & pnp
config".

used hardware:
p3-500
256mb sdram
geforce ddr
350w power supply

please help!
Click to expand...

Have you tried pressing ctrl+a while it is booting up? Even if the
message is turned off the controller should still respond to the key
presses.
 
Roger said:
I'm afraid you have probably destroyed the BIOS on the Adaptec card. There
used to be a warning on the Adaptec site, to remove some of their cards
from particular motherboards, before attempting a motherboard BIOS flash,
because on some motherboards, the addresses used, are the same as those
used by the cards...

Best Wishes
Click to expand...
He can not destroy the bios for the controller. It is built into the
board so it can not be removed. If he flashed the wrong bios, i.e. one
for a P2B rather than a P2B-S the controller wouldn't work but would
work again once the correct bios was flashed.
 
Paul said:
After a BIOS flash, try "Load Setup Defaults" in the BIOS.

If still no positive results, try the "Clear CMOS" procedure,
_unplugging_ the computer before doing the procedure. It is
in the jumper section of the manual - see "Clear real time clock
(RTC)" for details. Apparently the board uses two solder points
and you short them with the blade of a slot-type screwdriver tip.

HTH,
Paul
Click to expand...

Second that - I've flashed a gazillion of these boards to 1014b3 and
never seen this problem. The OP *should* see a CMOS checksum error on
the first reboot after flashing (because 1014b3 updates the SCSI BIOS as
well), but that should clear after a second reboot - without loading
setup defaults or clearing CMOS. This has to be a CMOS issue.

I find the solder points difficult to use and unreliable - a better
procedure is to disconnect the power cord, remove the battery, and short
the battery contacts with a screwdriver for 10 seconds. Check that the
BIOS clock has been reset to 1980 to verify it worked.

P2B
 
thanks for the responses, but meanwhile i tried all mentioned above.

- clear cmos
- load bios or setup defaults
- hitting a 100 times ctrl+a while startup with no result
- verified that it is really a p2b-s bios
- no errors were found else

but there is one thing i found out: the chip on the board labeled "aic
38600" has right after power on a very(!) high temperature... i think
the board is defective somehow :(
 
`sNaKe said:
but there is one thing i found out: the chip on the board labeled "aic
38600" has right after power on a very(!) high temperature... i think
the board is defective somehow :(
Click to expand...

Indeed the SCSI chip burning out is among the more frequent failure
causes for the P2B-S. (You're lucky the rest of the board is still
working, this can block the whole PCI bus.) I have no idea why, maybe
the power supply design is messed up or the chips themselves are
unreliable. Hmm, the AIC3860 you say? That's the bridge chip for SE
SCSI. You /might/ be able to regain access to the LVD part by
unsoldering the Vcc leg of the IC, but that's really something for the
skilled technician.
BTW, haven't you saved the old BIOS? However, I assume that the SCSI
chip stopping working after the flash was just more or less a
coincidence, or maybe the contained SCSI BIOS is more sensitive to a
failed SE part (so that the bridge chip had been defective for a while
already which only now surfaced due to the different BIOS).

Stephan
 
Stephan Grossklass said:
Indeed the SCSI chip burning out is among the more frequent failure
causes for the P2B-S. (You're lucky the rest of the board is still
working, this can block the whole PCI bus.) I have no idea why, maybe
the power supply design is messed up or the chips themselves are
unreliable. Hmm, the AIC3860 you say? That's the bridge chip for SE
SCSI. You /might/ be able to regain access to the LVD part by
unsoldering the Vcc leg of the IC, but that's really something for the
skilled technician.
BTW, haven't you saved the old BIOS? However, I assume that the SCSI
chip stopping working after the flash was just more or less a
coincidence, or maybe the contained SCSI BIOS is more sensitive to a
failed SE part (so that the bridge chip had been defective for a while
already which only now surfaced due to the different BIOS).

Stephan
Click to expand...

Some chips on a motherboard have multiple power rails on them,
and the chips may also have a sequence requirement for those rails.
I generally don't see a lot of sequencing stuff (MOSFETs) used on
Asus boards, to limit which rail comes up first. (But if you look
at an Intel reference schematic, you'll see more logic on there
than on an Asus board, to control sequencing or back-feed).

All it takes is drawing a lot of current through a parasitic
diode path, in excess of the latchup current spec, and a chip can be
turned into a molten mess. There are also micro-latchup cases, where
smaller portions of circuitry are involved. Some of these are
survivable. (Latchup - PNPN SCR junction formed where none was
intended. SCR conducts current until power is shut off, and then
it has a chance to recover.)

I had a chip once, that appeared to be dead to the world. Couldn't
get it to budge at reset. Left the PCB overnight, came in the next
morning, and the chip ran as if nothing had happened. Saw the same
thing on a couple of other boards. I eventually figured out what was
wrong, and the symptoms I was seeing were another manifestation of
latchup (the minor kind).

The meltdown kind are generally fatal - because you usually cannot
turn off the power before the damage is done. Since a PC power supply
has so much reserve power, whatever the chip wants, it gets.

If the Adaptec chip is a ball grid array, you'd have to blow torch
it off the board, to disconnect VCC :-) A place that builds
electronic boards, will have the necessary hot air gun and fittings,
to remove such a bad IC and allow the rest of the board to
continue to be used. Unless you live in Taiwan, the work won't be
done cheaply.

Paul
 
Paul said:
Some chips on a motherboard have multiple power rails on them,
and the chips may also have a sequence requirement for those rails.
I generally don't see a lot of sequencing stuff (MOSFETs) used on
Asus boards, to limit which rail comes up first. (But if you look
at an Intel reference schematic, you'll see more logic on there
than on an Asus board, to control sequencing or back-feed).

All it takes is drawing a lot of current through a parasitic
diode path, in excess of the latchup current spec, and a chip can be
turned into a molten mess. There are also micro-latchup cases, where
smaller portions of circuitry are involved. Some of these are
survivable. (Latchup - PNPN SCR junction formed where none was
intended. SCR conducts current until power is shut off, and then
it has a chance to recover.)

I had a chip once, that appeared to be dead to the world. Couldn't
get it to budge at reset. Left the PCB overnight, came in the next
morning, and the chip ran as if nothing had happened. Saw the same
thing on a couple of other boards. I eventually figured out what was
wrong, and the symptoms I was seeing were another manifestation of
latchup (the minor kind).

The meltdown kind are generally fatal - because you usually cannot
turn off the power before the damage is done. Since a PC power supply
has so much reserve power, whatever the chip wants, it gets.

If the Adaptec chip is a ball grid array, you'd have to blow torch
it off the board, to disconnect VCC :-) A place that builds
electronic boards, will have the necessary hot air gun and fittings,
to remove such a bad IC and allow the rest of the board to
continue to be used. Unless you live in Taiwan, the work won't be
done cheaply.

Paul
Click to expand...

There are two Adaptec chips on the P2B-S/DS - the AIC7890 is a BGA
device, but the AIC3860 (which appears to be the OP's problem) is a QFP.
It's nearly impossible to replace on the P2B-DS due to a cramped
location, but much more accessible on the P2B-S. Probably not worth the
effort to repair (I've done it once, not sure I want to repeat the
feat), however it would be interesting to test Stephan's theory that the
LVD segment may be still be usable once the SE bridge is out of the picture.

P2B
 
P2B said:
There are two Adaptec chips on the P2B-S/DS - the AIC7890 is a BGA
device, but the AIC3860 (which appears to be the OP's problem) is a QFP.
It's nearly impossible to replace on the P2B-DS due to a cramped
location, but much more accessible on the P2B-S. Probably not worth the
effort to repair (I've done it once, not sure I want to repeat the
feat), however it would be interesting to test Stephan's theory that the
LVD segment may be still be usable once the SE bridge is out of the picture.

P2B
Click to expand...

Do you have a theory on why they fail ? Inadequate ESD protection ?

I doubt there is a datasheet for Adaptec chips like that, but you
never know. It would be interesting to see if they are multiple rail
chips or not.

I never bothered to use the SCSI bus on my P2B-S - I had
seven SCSI drives at the time, but was too lazy to move them
over. So I never got any operational experience with those chips.

Paul
 
Paul said:
Do you have a theory on why they fail ? Inadequate ESD protection ?
Click to expand...

No - in fact I find Stephan's comment about SCSI failure being common on
these boards somewhat surprising, since I've been running and repairing
them for many years and have only seen a dead Adaptec chip *once*. IME
the most common failure mode is a dead BX Northbridge, which I suspect
may be due to inserting/removing PCI cards with standby power applied.

The one board I replaced an AIC3860 on was damaged during a thunderstorm
- took out the HIP6012 regulator. The board worked fine after it was
replaced (a minor miracle given the 6012 was visibly burnt), but SCSI
bus scans would hang if anything was connected to the SE segment. All
passive components around the 3860 tested OK and it appeared to have
Vcc, so I swapped it on spec and got lucky :-)
I doubt there is a datasheet for Adaptec chips like that, but you
never know. It would be interesting to see if they are multiple rail
chips or not.
Click to expand...

AFAIK Adaptec datasheets are not publicly available.
I never bothered to use the SCSI bus on my P2B-S - I had
seven SCSI drives at the time, but was too lazy to move them
over. So I never got any operational experience with those chips.
Click to expand...

Perhaps we can tag-team - I have no operational experience with the
onboard IDE controllers, apart from noticing the connectors collect dust
when you don't use them :-)

P2B
 
Back
Top