Conference noted::ibmpc-95

Instead of "capacitor", substitute "battery".

Need more info to get an idea what's wrong.  There are many other
possibilities if it "breaks" within the first day.

    The 386's and 486's I've worked on have an internal re-chargeable
    battery that keeps the BIOS settings when the PC is in the OFF state.
    If I remember, on most MB the battery is close to the keyboard
    connector in the rear of the MB. The voltage is about 3.6-3.9V Ni-Cad
    battery. If the new MB has been sitting around, its a good chance
    that the battery has grown chemical whiskers (internal shorts) that
    prevent the battery from being charged, by the internal charging
    circuit. You can check the "hobby" notefiles on how to use a very large
    capacitor to repair the Ni-Cad battery or replace it with a NEW one.
    
    //John
    

You might have a jumper set wrong too.  There is sometimes a jumper to select
whether to use the "soldered on" battery or an external one.

Burns

    Thanks guys,
    
    yes, the 'blue "battery"' is what I was looking at. It didn't look like
    a "traditional ;')" battery, so I thus assumed it was a capacitor -
    whatever, I knew it stored electricity for the thing to remember what
    the heck it was ;')
    
    Ok, my first thing will be to power the system up, and leave it running
    for an extended time, just to try and ensure the battery is charged.
    
    What made (makes) me suspect something other than this, was that the
    other m/b (the 386) had a very similar device (similar ratings on the
    side) and the losing the settings was why I invested in the 486 m/b.
    
    For both to fail in a similar way thus made me suspect something else.
    
    Anyway, I'm just starting for the day (Oz), so I'll be doing some
    searching on 'PX486DX33/P50' with AMI BIOS - and see what I can come up
    with.
    
    re: another battery - hmmm - nothing "obvious" on the m/b - but
    hopefully a search will reveal some more info - thanks for the
    direction given
    
    H

    Me again,
    
    I just put the multimeter across the blue battery on the 386 motherboard
    - barely registered in the tens of millivolts.
    
    Pulled the battery from the old 286 motherboard (the pc started as a
    286/12, 'grew' to a 386/16, and is now 486/40), which was removed some
    2 years ago now, and it still has something over 1 volt in it!
    
    Assuming that leaving the 486 running all night does not recharge the
    existing battery, my plan is, tomorrow night, to solder wires from this
    battery to the appropriate terminals on the existing battery on the
    motherboard, and attach the 'new' battery with hook-and-loop tape to
    the case somewhere nearby.
    
    Can anybody see a problem with having this "new" battery basically in
    parallel with the existing (presumably dead) battery? There is, of
    course, the caveat that the 'new' second-hand battery will die sometime
    soon ;'), but casting that aside, are there any known problems with
    this?
    
    H

    re-1
    
    Do not put the batteries in parallel, The stronger battery will
    discharge through the weaker one, (Anyone remember kirkof's(sp) laws on
    current flow).. cut or unsolder the old one out of the circuit.
    
    Most motherboards have a connector for an external battery, check your
    motherboard doc's
    
    jj

    re .6:
    
    �(Anyone remember kirkof's(sp) laws on
    
    Kirchhoff.
    

Re:  Don't do it ...

Especially true if the old battery died by developing internal shorts
(fairly likely).

However you add the new one (best to use a connector if your board has one
and move the jumper to enable it), measure the battery's voltage before and
after connection.  If there's much change, assume there's something still
wrong with the board.

    When you shutdown the system, press the reset and go into your BIOS
    setup page. Are all parameters still there? Now power it off.
    
    Just to make sure that when you power off we have indeed a valid
    setting in the CMOS. If the CMOS setting is gone before powering it
    off maybe a program causes this.
    
    Guenther

    Assuming you take the advice to use the external battery connector and
    jumper (which I agree with), you may not be able to use the old
    rechargable battery there.  I think those external connectors are
    usually for long-life non-rechargables, and thus would not have the
    recharge current there.  Hopefully you have the motherboard docs to
    check this out...  Assuming the battery you have there does not charge,
    do double-check the internal/external battery jumper position.  That
    might explain it not charging also.
    
    Burns

    Hullo,
    
    gee, am I *glad* I asked ;')
    
    No, no doc's - second hand M/b from a dealer who specialises in
    upgrading people's machines with new m/b's and keeps the old one as a
    'trade-in'.
    
    Ok, looks like the best I should do is go and buy a brand new battery,
    desolder the one already there, and put the new one in. Good thing I
    have most of the 'anti-stat' gear I need to do this ;')
    
    Last night I checked the notes in here on "batt"eries (dir/titl=batt)
    and read through them. I luv notes :')
    
    Again, much thanks for the help.
    
    H

The manuals for my motherboards show a jumper to select the battery's
voltage range ... mine talk about two different voltage ranges.

It seems risky to replace the battery when you don't know which voltage
to use.

I don't know ... but I'm less likely to pin it on a dead battery so soon,
on a new motherboard.  It could be just that the voltage jumper is in the
wrong position and the battery isn't charging.  Or maybe another problem
entirely.

    Seems to me that there is an implied warranty from the dealer
    you bought this motherboard from. Take it back to them and
    ask them to verify that the motherboard functions as you bought
    it. Tell them what you observed upon arriving home after leaving their
    store. Also tell them that you don't intend to pay them to diagnose 
    something that should be implied in the sale of that hardware. 
    That is, that for some reasonable amount of time it should function 
    as a functioning motherboard should function, and that you shouldn't 
    be seeing this problm for something so recently bought. If they won't 
    diagnose this motherboard without charging you, then ask them for a 
    replacement, otherwise you will be forced to write a certified letter 
    to your State's Consumer Affairs Office complaining of their business
    practices.
    
    Bob

.0> BIOS forgets its' stuff. My first thought was that the capacitor on the
    
    What does this mean in technical language? The system didn't boot
    because of message blabla? The BIOS page showed no or nonsense entries?
    Or...?
    
.0> wonderfully. Today I get a call, "it's broken". Sounds suspiciously
    
    A call? So you don't have experienced this yourself? Anything more
    specific than "it's broken"?
    
    I doubt that your problem has anything close to do with the CMOS
    battery given close to zero information on the symptom.
    
    Guenther

    ok,
    
    first, some background. This box was a 286, I upgraded a couple of
    years ago to a 386, and last weekend to a 486. In the meantime my wife
    has bought a P90, and I have a DECpc MTE d2/466 (yay!).
    
    This machine (with the problems) is one we have given to my Father in
    Law to repalce his XT clone that has just died (the hard disk has
    irretrievably failed - I grabbed backups a week before it failed ;')
    
    I have observed the problems, and because I work for a computer company,
    therefore I know something about computers ;') Having replaced the
    motherboard in this box twice now, I know a little - very little.
    
    The major symptoms are:
    
    1 - power on system for first time for the day, error message ensues
        "BIOS [or is it CMOS?] checksum error - press <F1> to continue" which
        takes you into the BIOS set up, to enter the disks (both hard and
        floppy), date, time, keyboard installed, etc. - I accept defaults
        for most of the other settings (e.g. memory and such)
    
    1a - sometimes the date and time is remembered, sometimes the time,
        most often nothing at all
    
    2 - exit & save BIOS settings - machine reboots & loads Win95 correctly
    
    3 - work as per normal (wp, games, etc.)
    
    4 - if machine powered down, and then restarted within several hours,
        it works as expected (normal boot)
    
    5 - if the machine is restarted after some 10 to 12 hours (or more),
        i.e. overnight, or next evening, then it loses the BIOS settings, 
        and step 1 is repeated
    
    I have checked the battery (located just behind the keyboard connector
    on the m/b), with a multimeter and it reads 4.1V (rating on the side is
    3.6V but I'm not too worried about that at this time). So, there
    appears to be enough voltage. Without desoldering the battery, I can't
    really determine the current it is drawing.
    
    I had hoped, that by leaving the machine turned on overnight, the
    battery would receive a full charge (it did), and this would solve the
    problem. It now appears that the problem is in fact something else
    (ain't it always? ;').
    
    I have checked all the socketted chips on the m/b and ensured that they
    are properly seated.
    
    I will check the m/b tonight to see if there may be a problem with the
    soldering of the onboard battery. I will also be checking the clearance
    around the m/b and between cards and the caseing, to verivy that there
    is no 'short' happening.
    
    If there is nothing visibly wrong, I will then contact the store, and
    see what we can do from there. Thanks again for your ideas.
    
    H
    

Since you have a multimeter, there are a few other things you can try that
MIGHT give you another clue or two.

What's the battery voltage when it's running, and been running a while?

What's the battery voltage just after turning it off?

What's the battery voltage after being off all night, before turning it on?

I would expect the first to be slightly higher than the other two (due to
charging), but the other two ought to be about the same as one another.

As far as I know, PC power supplies are as good as unplugged when turned
off (i.e., they don't continue to trickle-charge the battery when "off").

Sometimes something other than the battery (like a runaway program) can
cause BIOS corruption, but it does sound like the battery is most likely.

    Here is a shot in the dark.  Could there be a CMOS reset jumper that
    came loose on your MB?  I had a problem with setting up a CMOS password 
    setup and got locked out.  I had to change the CMOS jumper on my 486 MB 
    to reset the CMOS information to reboot.  I lost the disk(s) setup 
    information and the BIOS reset to the default setting after I did the 
    reset. I had to reenter the information.
    
    	Reset=move a jumper (J5 in my case) from pins 1-2 to 2-3, power up
    	and then power down.  Move the jumper back from 2-3 to 1-2 and then
    	reboot.  This somehow removes power from the CMOS memory and resets
    	the contents to the default settings (to no password check, etc).
    
    This might not explain why a second MB does the same thing.

    Well,
    
    last weekend, I took the board (in machine and connected up) back to
    the shop, and the technician had a quick look, and thought "ah! new
    battery" and attached an external battery to the beast. His only caveat
    was that he might have got the polarity wrong (!), so I should swap it
    around it it doesn't hold the settings :'\
    
    It didn't, and I did. And it still didn't {sigh}. He then suggested I
    call the local manufacturer/distributor of the board, which I did.
    Unfortuantely, they no longer have the documentation on these "old
    boards" (copyright 1993!), they "use only Intel boards now". So I struck
    out there as well.
    
    I called the store again, and he agreed to swap the board, which I
    installed yesterday, ran for several hours, probed the battery, etc.,
    checked jumpers, all that stuff, and it seems to be "ok" today.
    
    No idea what was the fault with the other board was - and it seems that
    the technician at the shop doesn't either - so I'll chalk it to "one of
    those things". If I had the (a) time, (b) inclination or (c) the test
    equipment, I'd probably try and track it down - but as I have none of
    the above - and it's no longer my machine anyway (given to my Father in
    Law), I will consider this case closed (for the time being).
    
    Thanks for all the help and suggestions.
    
    H