Can dust create a short?

[Gavin]

My MacBook is having intermittent kernel panics and is reporting:

"i2c0::_checkBusStatus SCL is stuck low; last write status 00010108 int shadow 00010100 xfer 00000000 fifo 00000000 for device audio-tas5770L-spkr-woofer-l”

The bits that caught my eye there were the words: i2c, SCL stuck low, and spkr-woofer-l. The task that panics is called bridgeaudiod.

I can’t think that I have installed any new software that might trigger this behaviour, so am thinking that it could be a hardware fault. Because the problem is intermittent, I am hoping that nothing is proper broken, but instead maybe something is perhaps shorting. I don’t think I have spilled anything on it that might cause this, so am wondering if a build up of dust could be the cause? The computer is now 3 years old

A few questions:

1/ could dust cause a short that would hold the SCL low?

2/ looking at the iFixit guides, the speakers are detachable and replaceable but I have not seen any specification for the speaker connectors. I would imagine that the connectors are carrying audio signals rather than i2c - is it even possible that they could be carrying i2c?

3/ or maybe I am looking at this the wrong way, and SCL could be stuck low because of a loose connection somewhere?

All thoughts on how I might go about trying to fix welcome

Gavin

[Nigel Worsley]

> A few questions:
> 1/ could dust cause a short that would hold the SCL low?

Highly unlikely, the I2C specification is for the pullup to between 1K
and 10K so the dust would need to be fairly conductive - metal dust or
very wet ordinary dust.

> 3/ or maybe I am looking at this the wrong way, and SCL could be stuck low because of a loose connection somewhere?

No, the I2C master should have pullup resistors so any open circuit
would have to be a failed resistor or the very short PCB trace between
them and whatever IC acts as the master. Not impossible, but
exceedingly unlikely.

I would put my money on a blown chip shorting the SCL line, no idea
how to go about finding that though!

Nigle

[Gavin]

Thanks Nigle. Sadly the blown chip sounds much more plausible than conductive fluff, and my guess, from the look of the replacement speakers on AliExpress, which are (a) < £15 for a pair, and (b) have six contacts on them, I think that chip is mostly likely on the motherboard rather than in the speaker units.  

I will take the cover off and have a blow around inside, just in case there is a little bit of metal wreaking havoc in there, but these may be the last days of this laptop :/

Gavin

--
You received this message because you are subscribed to the Google Groups "rLab / Reading's Hackspace" group.
To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/reading-hackspace/CAEq1Cr4BxGtOoepxv5oR1GYt5wzUD8_t91yL36zbPR9KvYPViw%40mail.gmail.com.

[Mark Robson]

I can't help with the problem, but I am very impressed at OS-level logging giving such a detailed (if correct) error message about an electrical fault.

We are truly living in an age of wonder, when the OS can tell you which wires are shorted.

Mark

--

You received this message because you are subscribed to the Google Groups "rLab / Reading's Hackspace" group.
To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/reading-hackspace/7F89530D-8222-4058-92FF-5D93AE620D06%40gmail.com.

[Bob Dunlop]

Hi,

On Thu, Nov 24 at 09:06, Mark Robson wrote:
> I can't help with the problem, but I am very impressed at OS-level logging
> giving such a detailed (if correct) error message about an electrical fault.
>
> We are truly living in an age of wonder, when the OS can tell you which
> wires are shorted.

Beware false positives and diagnosis that don't dig deep enough.

The pin is stuck low, but is a pullup failure, a short to ground, a dead
chips, a dodgy connector ? I'll add a combo I've just thought of, what
if the Vcc (power) line to the speaker has died due to a dodgy connector
then the I2C chip in the speaker could be dragging the line low.

--
Bob Dunlop

[Gavin]

Yes, a mile away from the ‘Something went wrong’ screen offered by a competitor 

--
You received this message because you are subscribed to the Google Groups "rLab / Reading's Hackspace" group.
To unsubscribe from this group and stop receiving emails from it, send an email to reading-hacksp...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/reading-hackspace/Y382ZYVRU47u%2BnQk%40xyzzy.org.uk.