[ale] computers sometimes befuddle me

Fri Jan 18 20:13:54 EST 2013

Various things are possible.  The case fan ran on the 5V bus that powers 
the hard drives.  It was still running (just loudly) when I unplugged it.  
Best guess is the shaft bearing was failing.  Motor noise on the bus is a 
reasonable possibility, but I don't understand why that would kill the 
video. The POST beep pattern I was getting was one Long, two Short, which 
for an ASUS board means there is a significant video problem. For the past 
several months, the behavior had been to rattle for a few minutes on 
startup and then quiet down -- but remain running -- for weeks of uptime.

But, HELLO! It had just died!  The other computers taking up space in my 
office are too old to handle more than a gig of ram, nor can they deal with 
SATA drives. I blew all my computer budget on a new monitor last month, so 
a new box is out of the question for now.

So when this one died and then magically came back to life again, I was 
not about to start fiddling again in an attempt to solve the (interesting) 
question of "why?"

The worthwhile adage is, "If it aint broke, don't fix it." :)

Sean

-------------------------------------------------------------------------------------------

On Friday, January 18, 2013 05:50:37 pm Michael H. Warfield wrote:
> On Fri, 2013-01-18 at 16:42 -0500, Calvin Harrigan wrote:
> > Could be that the fan was shorted?  Killing the 5V supply in that
> > area of the board?  Where you getting a POST beep?
> 
> Those fans usually run on 12V (yes, there are exceptions).  Especially
> the larger chassis fans and even the CPU and GPU fans (some laptops
> have 5V fans).  You wouldn't typically want them on your more
> sensitive 5V rails where they would draw more than double (2.4X) the
> current and potentially introduce motor noise on that bus.
> 
> When I run into an anomaly like that, I would do something like
> plugging the fan back in and confirming it was the cause rather than
> assuming. It could be equally possible that it was thermal and the
> temperature fluctuations of powering it down and back up again could
> have triggered it.  I've had more than one computer that had been
> running hot for months (or more) without batting an eyelash but, when
> turn off, would not power up from a "cold" state.  Let it warm up with
> power on those rails (dead as it were) and then reset it and up it
> comes.  Those sorts of problems are USUALLY in socketed chips but not
> always.
> 
> My favorite story in that regard relates back to Hayes Microcomputer
> Products days back when I was married to my first wife, Julie, and she
> was one of their lead techs.  They were having chronic problems with
> their new 9600 baud modems (damn, we've come a long way baby) where
> they would eventually start to fail after months of service.
> 
> She had given me an early production model for my birthday one year.
> Mine eventually failed but I figured out why.  She had told me about
> the problems they were being plagued with.
> 
> It had two connected boards, one upside down.  The upside down one had
> a big square socketed, edge contact (leadless), chip AND it had cut
> corners on all four corners of the socket (meaning each side was not
> connected to the other sides other than through the base and could
> flex) so you could use a chip-lifter to pull the chip.  When I took my
> modem apart and pressed firmly on that chip and heard a nice
> satisfying SNAP! Modem worked after that.  I told her that it was bad
> socket design and that thermal cycling would cause the upside down
> chip to walk its way out of the socket.  They needed solid corner
> sockets with positive retention.  She relayed the word up to the
> engineers and they fixed the problem and she got a bonus.  Case
> closed.
> 
> DIP (Dual In-line Package) chips use to (still have) have similar
> problems in sockets.  From the supplier, the legs bend outwardst
> slightly.  That's for the machines that insert them.  The legs act like
> little springs that hold them in place in the machines until inserted.
> But...  That same spring action causes them to walk out of DIP sockets
> over time (years).  Old service man magic was, when a machine started
> acting flaky, to take a board and press all the socketed chips back
> into their sockets or even take the chips out, square the leads, and
> reinsert.  9 times out of 10 that resurrected the balking machine.
> 
> Same thing happens with boards and board edge connectors.
> 
> Regards,
> Mike
> 
> > On 1/18/2013 4:35 PM, Sean Kilpatrick wrote:
> > > I was doing routine maintenance -- cleaning out dust bunnies.
> > > 
> > > Also wanted to isolate which fan was rattling on startup. It turns
> > > out the extra case fan is making the noise -- noise goes away
> > > after 2-3 minutes of up time.
> > > 
> > > Finish the cleaning, close the case, and turn on the power: get
> > > beep code indicating no video. WTF
> > > 
> > > Open up the case and try reseating the video card (mobo video ports
> > > have died.) No help.
> > > 
> > > Changed video cables. No help.
> > > 
> > > Plugged my netbook into the big monitor. Monitor works. (little ARM
> > > netbook is working as hard as it can to drive the larger monitor at
> > > full resolution, so browser is slowed to a crawl, but it does
> > > work.)
> > > 
> > > Finally, while I am sitting on the floor scratching my whatever, I
> > > pull the internal power plug to the noisy fan. On powerup video
> > > works again.
> > > 
> > > WTF!
> > > 
> > > Sean
> > > 
> > > 
> > > 
> > > _______________________________________________
> > > Ale mailing list
> > > Ale at ale.org
> > > http://mail.ale.org/mailman/listinfo/ale
> > > See JOBS, ANNOUNCE and SCHOOLS lists at
> > > http://mail.ale.org/mailman/listinfo
> > 
> > _______________________________________________
> > Ale mailing list
> > Ale at ale.org
> > http://mail.ale.org/mailman/listinfo/ale
> > See JOBS, ANNOUNCE and SCHOOLS lists at
> > http://mail.ale.org/mailman/listinfo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ale.org/pipermail/ale/attachments/20130118/45638649/attachment-0001.html>