P4C800E-Deluxe USB Ports have stopped working

[Paul]

MMmm...
Interesting stuff..

So Paul, am I to understand that if there was an incident of small static
discharge trying to ground itself through the outer wall of the USB
connector - through the ground wire and then to pin 10 of the board header -
all would be well?

If pin 10 is NC and it receives a micro jolt or a few milliamps wouldn't
that be enough to cause damage?

JG

The only solid evidence we have right now, is the Gigabyte
web page. The advice I've given above is basically repeating
what Gigabyte is saying, only I consider the problem to be
more severe than they do. I don't feel that any grounding
scheme is sufficient to protect the motherboard (based on
reports that plugging into the motherboard USB connector
stack on the back of the computer is also killing boards).

http://tw.giga-byte.com/Motherboard/Support/FAQ/FAQ_456.htm

"The ICH4/5 SouthBridge chipset on my Intel motherboard
appears damaged or "burnt". How did this happen?

The GIGABYTE Global RMA Service Center has recently become aware
of the randomly occurring risk to ICH4/5 SouthBridge chipsets on
Intel motherboards to become damaged or "burnt" when a USB device
is connected. Please read the following paragraphs to learn more
about this issue.

Why did this happen?

There are three likely causes:

Accumulated static electricity from the human body may have been
transmitted to the ICH4/5 SouthBridge chipset when connecting a
USB device if it has not been "grounded". A connected USB device
does not have any static electricity protection. Accumulated
static electricity in a front-panel USB port that is not properly
"grounded" with the PC casing.

Usage Tips

- Avoid touching the USB connector.
- Release static electricity within your USB device by touching
the USB connector to any metal surface before plugging it into
your computer.
- Use rear-panel USB ports on your PC whenever possible.
- Use a three-pronged electrical cord to connect a power supply
to your PC.
- Use only approved or certified power supplies."

The Asus 2x5 connector uses only eight pins. There are two ground
pins. A typical USB installation needs three ground signals.
Two grounds for the ground pins in the two USB connectors. Plus
a shield connection is needed for the outside of the USB connector.
A good cable assembly will take one of the USB signal grounds
and splice the shield wire into it too (good here meaning not
good design practice, because that is not good electrical
design practice, but good as in working within the compromises
that Asus has created by not having a third ground pin).

VCC
D+
D-
GND -------------------------> To one Asus GND pin

VCC
D+
D-
GND ------------+------------> To the other Asus GND pin
|
GND_shield -----+

If the GND_shield is connected to NC, and static were to be applied
to the wire, the question would be, how much static voltage, and
how much clearance is there from the unconnected pin to any adjacent
conductors. The USB data signals might not really get very close
to that pin, so whether this is an issue or not really depends on
the layout of the board.

Since we don't know how sensitive the USB ports really are,
it could be that just dumping static discharge into the metal
of the computer case is enough to cause latchup in a USB port.
I cannot discount that as a possibility, but there is no way
to protect against that. The advice I've given assumes the
motherboard is resilient enough not to be affected by a static
discharge to the chassis, but only time will tell whether this
is true or not (we know Asus or Intel won't be telling us). I'm
hoping that by removing any "antennas" from the USB headers, and
keeping away from the rear motherboard USB ports, that most users
will enjoy use of their motherboard after the warranty has expired.

Paul

 

[Paul]

Well I just completed the Final test that Paul suggested. Reset the
Cmos. Result USB still not working. I even took out every add in card
iht the PC out of desperation. Still no go. The USB ports are only
useful now to use 1 of those Nice little USB Lights. Like you said Jim,
it finds the controller and everything but the ports just don't
recognize any USB that gets plugged into it.

Another possibility is your case is actually "micro latchup",
meaning a portion of circuitry in the USB area is having
a seizure. I have experienced this in the lab, and the
symptoms are strange. Turning off power for five minutes
did not restore the operation of the IC. I turned off the
power to my project, and went home for the night, and when
I returned the next day and applied power, all was well again.

It could be, that what you need to do, is remove all power
from the motherboard overnight. That means unplugging the
computer, removing the CMOS battery, and letting it sit
overnight. The idea is, that by the next morning, perhaps
all capacitors will have dropped to zero volts, and any
PNPN phantom junctions will have had a chance to shut off.
Then, try reinstalling the battery, plug in and give it a
go.

It just doesn't make sense to me, that other parts of your
chip are not damaged. Just like my jaw dropped when I went
to my job in the morning many years ago, to find what I
thought I'd killed the day before, was alive again. My
micro latchup case was in fact reproducible, and I had
another chip that did the same thing - it lost all
functions, I turned off the power for a period of hours,
and turned on again to find it still worked.

It could be, that +5VSB (the one that lights the green
LED) or the 3 volts from the CMOS battery, is enough to
hold the circuit in latchup.

Paul

 

[Jim Gibson]

You obviously have looked into this much more than I have - so who is to
'blame'.
Intel or the board designers/makers?
A quick look on line at a Toronto supplier's stock of Asus boards shows over
2000 P4P, P4V and P4S boards in stock.
Asus must still be churning socket 478 boards out......and there's gonna be
some disappointed buyers down the road.

To re-visit the case USB connector:
If one were to solder a wire from the outer shield of the USB connector then
to the chassis - how would that work?

Thanks for all you info..

JG

 

[Jim Gibson]

Ha Huh........

I still have my USB-dead P4P800-SE - too busy to package it up and send it
off.
I will remove the CMOS battery - wait 24hrs and fire it up.

I will get back to you.

JG

 

[Ken]

To re-visit the case USB connector:
If one were to solder a wire from the outer shield of the
USB connector then to the chassis - how would that work?

Better. The outer shield must be connected to the chassis.
On my Lian-Li case the front outer USB metallic shield is
bolted directly to the chassis that make very good ESD
protection (but I have not connected them yet).

 

[cold_canuck]

Hey Jim,

I am in T.O. as well. How do I go about getting an RMA from ASUS ? and
do I need to send it Stateside ?

I will try Paul's overnight suggestion as well. Funny I have always
used ASUS boards without a Hiccup, first time I spend for the top O the
Line and Zap doesn't want to work.

Frustrating.

 

[Paul]

Hey Jim,

I am in T.O. as well. How do I go about getting an RMA from ASUS ? and
do I need to send it Stateside ?

I will try Paul's overnight suggestion as well. Funny I have always
used ASUS boards without a Hiccup, first time I spend for the top O the
Line and Zap doesn't want to work.

Frustrating.

Remember that this is an experiment and not a proven cure.
I'm only suggesting removing all power, because I'm having
trouble believing that just the USB logic is dead.

Paul

 

[Paul]

You obviously have looked into this much more than I have - so who is to
'blame'.
Intel or the board designers/makers?
A quick look on line at a Toronto supplier's stock of Asus boards shows over
2000 P4P, P4V and P4S boards in stock.
Asus must still be churning socket 478 boards out......and there's gonna be
some disappointed buyers down the road.

To re-visit the case USB connector:
If one were to solder a wire from the outer shield of the USB connector then
to the chassis - how would that work?

Thanks for all you info..

JG

First of all, we don't understand the failure mechanism. So,
assigning blame is difficult. With regard to warranty claims,
that would be handled behind the scenes, so we don't know if
either party is on the hook financially.

In your list above, the boards with non-Intel chipsets would
not be involved. At least, there are no similar reports for
them (yet...).

The industry does have some awareness of the need for taking
ESD into account, when designing USB devices. But, in this
article, the emphasis is on preventing the computer from
crashing when the ESD happens, rather than preventing latchup
and destruction of devices.

http://www.reed-electronics.com/tmworld/index.asp?layout=article&articleId=CA187442

Intel does work on ESD resistance in its chips. At the end of
page 3, you can see their work on input protection. The 4-8KV
human body model performance targets are pretty impressive.

http://www.intel.com/technology/itj/q31998/pdf/qa.pdf

ESD is tricky stuff. Using your intuition seldom results in
the correct answer. Your wire, for example, becomes a
broadcast transmitter when ESD goes down the wire. A better
solution, would be if the shield touched the metal chassis.

Part of the reason I'm being so cautious, is because failure
has occurred by plugging into the rear ports. The rear ports
have a virtually ideal grounding situation, if the objective is
to shunt ESD to the chassis and on to the safety ground of
the PSU. That is why I'm advocating simply not using the
motherboard USB ports. If the USB connector stacks on the
motherboard cannot protect the ports, your piece of wire is
not going to do it either. Just get a PCI USB card and let
it absorb the abuse.

HTH,
Paul

 

[Jim Gibson]

I replied to your hotmail address...
It may be siitting in your junk folder....

JG

 

[Jim Gibson]

RESULT

Result of the "leave the CMOS battery out for 36 hours and see if the USB
ports come back to life" test.

I removed the battery for at least 36 hours.
I put back the motherboard - this time with my spare Celeron. Everything
else remained the same
It fired up with no problems except BIOS complained about the date not
being set.
Set BIOS to default values.
Fired up XP.
I was sceptical as soon as I felt the Intel ICH5R chip on the board. It was
much too hot for being powered for only 2 minutes.
Plugged in my USB pen drive
Plugged in my USB printer.
I am sad to report that all USB ports were as dead as they were 36 hours
before.
So - - -it appears that whatever happened - static charge or not - the
damage was permanent.
I just don't understand how it can isolate only the USB controller.

JG

 

[Paul]

RESULT

Result of the "leave the CMOS battery out for 36 hours and see if the USB
ports come back to life" test.

I removed the battery for at least 36 hours.
I put back the motherboard - this time with my spare Celeron. Everything
else remained the same
It fired up with no problems except BIOS complained about the date not
being set.
Set BIOS to default values.
Fired up XP.
I was sceptical as soon as I felt the Intel ICH5R chip on the board. It was
much too hot for being powered for only 2 minutes.
Plugged in my USB pen drive
Plugged in my USB printer.
I am sad to report that all USB ports were as dead as they were 36 hours
before.
So - - -it appears that whatever happened - static charge or not - the
damage was permanent.
I just don't understand how it can isolate only the USB controller.

JG

Here is an interesting thread on the issue. Complete with a couple
of pictures of burned Southbridges.

http://www.abxzone.com/forums/showthread.php?t=84122&highlight=usb+port

There is one claim of a failure right after rebooting. Very
puzzling.

Paul

 

[cold_canuck]

I wanted to let you know as well that I left the battery out overnight,
then rehooked everything the next day. Same result everything works
fine except all USB ports are dead.

I guess the days of plugging and unplugging USB connectors is over....
At least with this board.

If I have read correctly however as long as I do not use the sytem USB
ports but rather A PCI usb board I am safe to plug and unplug my MP3,
camera etc...

 

[Jim Gibson]

If I have read correctly however as long as I do not use the sytem USB

ports but rather A PCI usb board I am safe to plug and unplug my MP3,
camera etc...

Only, I have seen a board where the USB ports die first, then the on-board
video died - and then the entire board.
The ICH5R will get so hot that eventually it will fry.
Could be tomorrow, next week or next month, but I'm positive it won't be
long before things get worse.
If I were you, I'd bite the bullet and send it back ASAP. Mine was shipped
out yesterday.

JG.

I wanted to let you know as well that I left the battery out overnight,
then rehooked everything the next day. Same result everything works
fine except all USB ports are dead.

I guess the days of plugging and unplugging USB connectors is over....
At least with this board.

 

[Paul]

I wanted to let you know as well that I left the battery out overnight,
then rehooked everything the next day. Same result everything works
fine except all USB ports are dead.

I guess the days of plugging and unplugging USB connectors is over....
At least with this board.

If I have read correctly however as long as I do not use the sytem USB
ports but rather A PCI usb board I am safe to plug and unplug my MP3,
camera etc...

Thanks

My recommendations are based on whatever symptom descriptions I
am reading posted here or in any of the private forums that
happen to have noticed this phenomenon. My comment about one
board blowing after a reboot (i.e. no USB plugging was noted by
the poster) has to be about the worst report so far.

For most people, the trigger event is plugging a USB device into
the computer while it is running. The USB port in question, is
connected to the Southbridge. Soon after, the Southbridge gets
too hot to touch, and very soon, the board will be dead.
That is how the majority of descriptions run.

Based on that, using a separate PCI USB card, for all your USB
devices, is a workaround. At least until someone gets to the
bottom of this. That means, do not use the motherboard USB
connectors, and do not use the USB 2x5 headers (disconnect
any computer case front panel USB ports from the headers). That
is the most conservative practice I can recommend, short of
outright replacing the motherboard with one based on Via,
SIS, Nvidia, or ATI chipset.

As I own one of these boards too, I'm not happy about this
development, as this is the best computer I've got in the house
right now. I'm not using any USB with the thing at the moment,
and we'll see how long it lasts.

The reason a separate PCI card is safer to use, is if there
really is an issue with static electricity, it will ruin
your $10-$20 PCI USB card and not the rest of the motherboard.
As reports of USB trouble on other devices are rare to
non-existent, a separate PCI card should be pretty safe.

Paul

 

[Chris9999]

My recommendations are based on whatever symptom descriptions
I
am reading posted here or in any of the private forums that
happen to have noticed this phenomenon. My comment about one
board blowing after a reboot (i.e. no USB plugging was noted
by
the poster) has to be about the worst report so far.

For most people, the trigger event is plugging a USB device
into
the computer while it is running. The USB port in question, is
connected to the Southbridge. Soon after, the Southbridge gets
too hot to touch, and very soon, the board will be dead.
That is how the majority of descriptions run.

Based on that, using a separate PCI USB card, for all your USB
devices, is a workaround. At least until someone gets to the
bottom of this. That means, do not use the motherboard USB
connectors, and do not use the USB 2x5 headers (disconnect
any computer case front panel USB ports from the headers).
That
is the most conservative practice I can recommend, short of
outright replacing the motherboard with one based on Via,
SIS, Nvidia, or ATI chipset.

As I own one of these boards too, I'm not happy about this
development, as this is the best computer I've got in the
house
right now. I'm not using any USB with the thing at the moment,
and we'll see how long it lasts.

The reason a separate PCI card is safer to use, is if there
really is an issue with static electricity, it will ruin
your $10-$20 PCI USB card and not the rest of the motherboard.
As reports of USB trouble on other devices are rare to
non-existent, a separate PCI card should be pretty safe.

Paul

Paul,
I have been reading your posts re failure of ICH5 chips with much
interest.
We have 11 PC’s using these chips on the Asus P4P800-E board. So far 3
have failed. The last failure happened last week. They have all gone
up in smoke leaving a severe burn mark on the chip. After speaking to
the student who was working on this PC it appears he touched the PC
case and felt a static discharge, the PC re-booted and then smoke
appeared. I have since looked at the CCTV covering this room and this
looks to be what happened. No USB devices were connected.
The 2 previous failures, I believe battery powered USB devices were
used via the front panel connector.
I am planning on disconnecting the USB front panel and rear plate
connectors ASAP as I believe static must be being picked up via the
USB internal wiring.

 

[Paul]

Paul,
I have been reading your posts re failure of ICH5 chips with much
interest.
We have 11 PC’s using these chips on the Asus P4P800-E board. So far 3
have failed. The last failure happened last week. They have all gone
up in smoke leaving a severe burn mark on the chip. After speaking to
the student who was working on this PC it appears he touched the PC
case and felt a static discharge, the PC re-booted and then smoke
appeared. I have since looked at the CCTV covering this room and this
looks to be what happened. No USB devices were connected.
The 2 previous failures, I believe battery powered USB devices were
used via the front panel connector.
I am planning on disconnecting the USB front panel and rear plate
connectors ASAP as I believe static must be being picked up via the
USB internal wiring.

Keep us posted on your failure rate. I want to see how
effective a strategy it is, to disconnect at least the
front panel USB wiring. I don''t know how you can do that
for the rear USB, unless you can unsolder the USB stacks
and remove them.

Depending on exactly how sensitive the USB ports are,
it might still be possible for a static discharge to the
computer case, to induce enough signal on the remaining
USB wires of the motherboard, to trigger another one off.

It could be that there is an "aging" effect here, and
the more power_on_hours a machine has, the more sensitive
to static it gets. It is not possible to draw any conclusions
from postings of failures here, as to whether there is
something like that at work or not. The reason I postulate
that, is static problems weren't apparent when these products
were introduced, but now suddenly they are "dropping like flies".

In any case, someone who has a large population of boards will
be able to formulate better theories than I can.

Paul

 

[Okoidogo]

Is it possible for all owners (or at least, willing owners) of ASUS
board using the ICH4 and 5 chipsets send a joint email to ASUS to
demand an answer ???

 

[Jim Gibson]

My theory still revolves around the ground wire going to the NC pin on the
USB board headers.
This happens on some PC cases with USB front panel connectors that are not
exactly designed to match Asus' pinouts on the board.
It might be a worthwhile exercise to note the kind of case , USB front panel
configuration and whether the front USB connectors are grounded to the case
and/or are wired directly to pin 10 on the board header, where the ICH5 chip
gets fried.
This of course does not account for rear USB port problems (but static
charge is probably involved)
As for a joint email to ASUS on behalf of everyone affected with the
'dreaded ICH5 burnout' -
ASUS are not going to admit a thing.
Another look online of the inventory of Asus boards at a Toronto area
supplier still reports 1,945 P4P* Intel boards still in stock - so they are
not being withdrawn, even if Asus knows of a problem.

JG

 

[Paul]

Is it possible for all owners (or at least, willing owners) of ASUS
board using the ICH4 and 5 chipsets send a joint email to ASUS to
demand an answer ???

When you send your email, don't forget to include "class
action lawsuit" in the title

I think you may have better luck using the phone. Perhaps
some incautious tech support person will give some details.
If their phone conversations are taped, you can forget
honesty. I've never tried their tech support myself.

And lawsuits do work - here is an example. This one was
for bad caps on Abit boards.

http://www.abitsettlement.com/

Paul