CMOS settings lost when unplugged, battery is good (was Re: [vox-tech] CMOS problems, battery dead?)

[ME]

Just pulled this out for now. Will try to visit other when I am better.

> > We still have:
> > power (we ruled out the battery itself, but the socket for the battery, or
> >   the power supply and feed to the mb are possible too.)
> > CMOS/BIOS chip(s)  (failure)
> > connection of chip to board
> > motherboard failure/short (This includes setting a jumper to reset flash
> >  BIOS to factory configs - possible. Look for it. If set, unset.)
> 
> I gotta dig up my manual. if the reset jumper is set I'll scream.
> 
> How would you sugguest checking the others?

Tricky. :-/

People who work with motherboards on a regular basis (in a shop, building
them for their buddies, etc.) would be better consulted, but there are
some thoughts:

> > power (we ruled out the battery itself, but the socket for the battery, or
> >   the power supply and feed to the mb are possible too.)

Rather paired with thisn for what I would check:

> > CMOS/BIOS chip(s)  (failure)

The socket:
See if you can find some Schematics on the board and the CMOS chip if it
is one of those that is put in a socket. If so, some of the pins may get
power from the batter (or should) when unplgged. A voltmeter can test
this, but you need to know what pins should be hot and what their voltage 
should be when power is disconnected..

> > connection of chip to board

Re-seat it, or pop it off and examine the little contacts for cracks or
other (?) damage.

See if it is an EPROM and has a cover over the transparent window that
should only be opened/exposed when erasing with UV light.

EEPROM, EPROM, and PROM can be tested with EPROM boards. However, this
often requires you to save the content to a file, and then erase the chip,
and perform random write and compares multiples times (better with EEPROM
than EPROM with limited write/erase.) The easiest is to take the intended
image/file out and write to the chip then compare.
If Flash based, ten re-writing various patters (much like a memory
checker) and checking for differences from what was writtern can work, but
is not perfect for chip problem detection. 

> > motherboard failure/short (This includes setting a jumper to reset flash
> >  BIOS to factory configs - possible. Look for it. If set, unset.)

Motherboard shorts? Um. I have only few thoughts on this. Volt meter,
logic probe, oscilliscope, but still need schematics and expected signals
and these kinds of details frequently not provided. In the past, if I rule
out everything else but the MB, I get a new MB. Too much work to detect
otherwise. Assuming you could even find the problem, could you actually
fix it? Multi-layer chips/board.. gonn drill it and.. ? heh. Too much work
and no certainty it will even work. 6 months ago for new MB? (what you
wrote for your replacement) That is not much time for a new MB. Maybe it
was a problem as shipped, but you did not notice until recently since the
uptime has been so high? (blind guess.)

As for a whort, if you knew what the current/voltage drain of the battery
*should* be when keeping the NVRAM happy, you could check that with a
voltmeter. If too high, that suggests some kind of short. You would expect
this to drian the battery quickly and shorten the life of the battery. You
would also expect this to bring your batter to a point close to dead
really fast, but YMMV. An intermittent break is also possible, but even
more difficult to trace. (Very remote, but say a case where when it is
warm, the metal conductors making intermittant contact  actually make
contact when the metal is warm and has expanded, but when cooled
sufficiently, the metal contracts and contact is lost.)

I doubt the last "reason", but toss it out anyway.

Sorry, I still have a headache - me brain is malfunctioning and it is
difficult to think or write. :-(

I'll try to revisit this later. Feedback good. Keep it coming!

-ME

-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCS/CM$/IT$/LS$/S/O$ !d--(++) !s !a+++(-----) C++$(++++) U++++$(+$) P+$>+++ 
L+++$(++) E W+++$(+) N+ o K w+$>++>+++ O-@ M+$ V-$>- !PS !PE Y+ !PGP
t@-(++) 5+@ X@ R- tv- b++ DI+++ D+ G--@ e+>++>++++ h(++)>+ r*>? z?
------END GEEK CODE BLOCK------
decode: http://www.ebb.org/ungeek/ about: http://www.geekcode.com/geek.html
     Systems Department Operating Systems Analyst for the SSU Library

On Sat, 27 Apr 2002, Ryan wrote:
> Date: Sat, 27 Apr 2002 21:39:32 -0700
> From: Ryan <ryan@mother.com>
> Reply-To: vox-tech@lists.lugod.org
> To: vox-tech@lists.lugod.org
> Subject: CMOS settings lost when unplugged,
>      battery is good (was Re: [vox-tech] CMOS problems, battery dead?)
> 
> On Saturday 27 April 2002 08:52 pm, ME wrote:
> > On Sat, 27 Apr 2002, Ryan wrote:
> > > After it's been off AC, I go to turn it on, and it whines about a CMOS
> > > checksum error, and my settings have been reset to thier defaults.
> >
> > Is this a home-built system, or made by a vendor?
> >
> > If built by a local vendor or by a buddy, then I would suspect power
> > supply and/or how it is connected to the motherboard.
> 
> Local vendor
> 
> > How old is the system? (I seem to recall you said 2 or 3 years, but am not
> > sure.)
> 
> About a year and a half, motherboard was replaced 6 months ago
> 
> > On a related question, if on a power strip, what happens if you leave the
> > power strip "on" but halt the computer and reboot "sfot reboot via
> > control-alt-delete?
> 
> Not a problem
> 
> > How about leave plugged in, and AC on, put cycle the
> > power switch? How about power down, unplug, plug in, power on?
> 
> It lost my CMOS settings after an hour when a car hit a power line last week, 
> was fine when off for a few seconds.
> 
> > About how much time off of AC is required for this to happen? (Right away
> > or hours or ?)
> >
> > > Leaving it soft-off overnight didn't cause me any problems.
> >
> > Is soft-off the same as "standby mode" where it is sleeping, or "sleep
> > mode" where it is low power, but in use or suspend mode where an image of
> > memory is written to disk, and then upon powering up the image is read
> > into memory toa llow the machine to pick up where it left off?
> 
> Soft off == PSU on, plugged in, but fully powered down. (it cold boots when i 
> turn it back on)
> 
> > We still have:
> > power (we ruled out the battery itself, but the socket for the battery, or
> >   the power supply and feed to the mb are possible too.)
> > CMOS/BIOS chip(s)  (failure)
> > connection of chip to board
> > motherboard failure/short (This includes setting a jumper to reset flash
> >  BIOS to factory configs - possible. Look for it. If set, unset.)
> 
> I gotta dig up my manual. if the reset jumper is set I'll scream.
> 
> How would you sugguest checking the others?
> 
> > What I can think of right now. Cant remember others. :-(
> >
> > There may be others, but power seems a good place to check.
> > (Other lister reader welcome to make additions. :-)
> _______________________________________________
> vox-tech mailing list
> vox-tech@lists.lugod.org
> http://lists.lugod.org/mailman/listinfo/vox-tech
>