Very old PC keeps losing Data and Time and BIOS settings. I took the
battery out and it's completely flat. As the machine is several years
old I'm waiting to get a new battery to fit. Just out of interest,
with the battery out of the mobo (pwr on of course) I measured the
voltage at the battery housing (on the board itrself). It was only 0.3
V. Is this normal? The lithium battery is 3V. I have no idea what the
circuit looks like to charge the battery so maybe as the battery is
not in circuit that might explain this very low voltage?
thanks

On Mon, 02 Nov 2009 18:07:51 +0000
dave  wrote:

> Very old PC keeps losing Data and Time and BIOS settings. I took the
> battery out and it's completely flat. As the machine is several years
> old I'm waiting to get a new battery to fit. Just out of interest,
> with the battery out of the mobo (pwr on of course) I measured the
> voltage at the battery housing (on the board itrself). It was only 0.3
> V. Is this normal? The lithium battery is 3V. I have no idea what the
> circuit looks like to charge the battery so maybe as the battery is
> not in circuit that might explain this very low voltage?
> 

It's not rechargeable, the lithium cell lasts for several years
because the mobo current requirement is very low, then you replace it.

On Mon, 2 Nov 2009 18:13:07 +0000, Rob Morley 
wrote:

>On Mon, 02 Nov 2009 18:07:51 +0000
>dave  wrote:
>
>> Very old PC keeps losing Data and Time and BIOS settings. I took the
>> battery out and it's completely flat. As the machine is several years
>> old I'm waiting to get a new battery to fit. Just out of interest,
>> with the battery out of the mobo (pwr on of course) I measured the
>> voltage at the battery housing (on the board itrself). It was only 0.3
>> V. Is this normal? The lithium battery is 3V. I have no idea what the
>> circuit looks like to charge the battery so maybe as the battery is
>> not in circuit that might explain this very low voltage?
>> 
>
>It's not rechargeable, the lithium cell lasts for several years
>because the mobo current requirement is very low, then you replace it.
>
AH I see - thanks Rob.

The message 
from dave  contains these words:

> Very old PC keeps losing Data and Time and BIOS settings. I took the
> battery out and it's completely flat. As the machine is several years
> old I'm waiting to get a new battery to fit. Just out of interest,
> with the battery out of the mobo (pwr on of course) I measured the
> voltage at the battery housing (on the board itrself). It was only 0.3
> V. Is this normal? The lithium battery is 3V. I have no idea what the
> circuit looks like to charge the battery so maybe as the battery is
> not in circuit that might explain this very low voltage?
> thanks

 Lithium coin cells (it's not a "battery", just a single cell) aren't
(meant to be) rechargable. They simply have an extremely long shelf life
(10 years at 20 deg C 'on the shelf' and still have 90% of their
original capacity - or 50 years if 50% capacity is good enough, some
RTCs can have a current draw comparable with the cells own leakage
current).

 At a mere 0.3v that cell has been "flat" for quite a few months. New,
they're typically between 3.15 and 3.30 volts, dropping to about 2.9v by
the end of their service life (with perhaps a month or three left before
dropping below 2.5v where the RTC oscillator will stop or else run slow
and maybe another 3 months before dropping below 2 volts and causing the
cmos settings to corrupt.

 I don't know about the ones used in laptops but the standard one used
in desktops is the CR2032 cell (much bigger than the ones usually found
in wristwatches which typically last 7 or more years). You should be
able to identify the CR number and buy a replacement for a quid or
three.

HTH

-- 
Regards, John.

 Please remove the "ohggcyht" before replying.
The address has been munged to reject Spam-bots.

On Mon, 2 Nov 2009 19:34:14 GMT, Johnny B Good
 wrote:

>The message 
>from dave  contains these words:
>
>> Very old PC keeps losing Data and Time and BIOS settings. I took the
>> battery out and it's completely flat. As the machine is several years
>> old I'm waiting to get a new battery to fit. Just out of interest,
>> with the battery out of the mobo (pwr on of course) I measured the
>> voltage at the battery housing (on the board itrself). It was only 0.3
>> V. Is this normal? The lithium battery is 3V. I have no idea what the
>> circuit looks like to charge the battery so maybe as the battery is
>> not in circuit that might explain this very low voltage?
>> thanks
>
> Lithium coin cells (it's not a "battery", just a single cell) aren't
>(meant to be) rechargable. They simply have an extremely long shelf life
>(10 years at 20 deg C 'on the shelf' and still have 90% of their
>original capacity - or 50 years if 50% capacity is good enough, some
>RTCs can have a current draw comparable with the cells own leakage
>current).
>
> At a mere 0.3v that cell has been "flat" for quite a few months. New,
>they're typically between 3.15 and 3.30 volts, dropping to about 2.9v by
>the end of their service life (with perhaps a month or three left before
>dropping below 2.5v where the RTC oscillator will stop or else run slow
>and maybe another 3 months before dropping below 2 volts and causing the
>cmos settings to corrupt.
>
> I don't know about the ones used in laptops but the standard one used
>in desktops is the CR2032 cell (much bigger than the ones usually found
>in wristwatches which typically last 7 or more years). You should be
>able to identify the CR number and buy a replacement for a quid or
>three.
>
>HTH
Yes very interesting thanks. It is indeed a CR2032 and I just "won" a
couple of them (duracells) - er cells, on ebay £2.50.

Johnny B Good wrote:
> 
>  I don't know about the ones used in laptops but the standard one used
> in desktops is the CR2032 cell (much bigger than the ones usually found
> in wristwatches which typically last 7 or more years).

Laptop cells are often the thinner CR2025, but it's not consistent.


-- 
John Jordan

On Mon, 2 Nov 2009 19:34:14 GMT
Johnny B Good  wrote:

>  Lithium coin cells (it's not a "battery", just a single cell)

We know that, but how much easier is it to use the common and incorrect
term than to explain the difference every time?  :-\

In article , dave
 writes

>Just out of interest,
>with the battery out of the mobo (pwr on of course) I measured the
>voltage at the batt

Yes.  There's a diode in the circuit to prevent the board attempting to
charge the battery.  You're measuring the voltage across the diode.

Lithium coin cells are not chargeable - you're supposed to replace them
when they go flat.

-- 
(\__/)   
(='.'=)  Bunny says Windows 7 is Vi$ta reloaded.
(")_(")  http://imgs.xkcd.com/comics/windows_7.png

In article , Johnny B Good
 writes

> At a mere 0.3v that cell has been "flat" for quite a few months. New,
>they're typically between 3.15 and 3.30 volts, dropping to about 2.9v by
>the end of their service life

Agreed.

> (with perhaps a month or three left before
>dropping below 2.5v where the RTC oscillator will stop or else run slow
>and maybe another 3 months before dropping below 2 volts and causing the
>cmos settings to corrupt.

Not my experience.  I find that as soon as the voltage drops below 3.0
the CMOS settings become lost.

> (much bigger than the ones usually found
>in wristwatches which typically last 7 or more years)

7 years for a watch battery?  You must tell me what you're using,
miniature nuclear reactors!?  I usually get 2 yrs max out of a watch.

I have an odd situation with CMOS batteries at work - perhaps you can
suggest something?  We have an installation with single-board computers
which control the various parts of a large moving instrument.  It's
located in a remote, unmanned place where the temperature can range from
-30C at night to +30C during the day.

The computers originally had 3v CR2032 batteries to maintain the CMOS
but they were lasting perhaps 18 months max.  Replacing those regularly
became tiresome (and expensive!) so I made up several plastic holders
with 2 x AA Duracell Ultras, thinking that would be the last of the
problem and we'd never need to touch the batteries again.  They're
connected to the SBC coin holder with spring clips, as there is no
header to attach external batteries to.

However, the Duracells also run low after about 18 months, leaving me
scratching my head.  We've pretty much excluded temperature swings as
the cause.

Any ideas?

-- 
(\__/)   
(='.'=)  Bunny says Windows 7 is Vi$ta reloaded.
(")_(")  http://imgs.xkcd.com/comics/windows_7.png

On Tue, 3 Nov 2009 04:57:25 +0000
Mike Tomlinson  wrote:

> In article , Johnny B
> Good  writes

> > (much bigger than the ones usually found
> >in wristwatches which typically last 7 or more years)
> 
> 7 years for a watch battery?  You must tell me what you're using,
> miniature nuclear reactors!?  I usually get 2 yrs max out of a watch.
> 
My Casio claims 10 years battery life (it says so on the face, so it
must be true) although I doubt that allows for much use of the light.

"Mike Tomlinson"  wrote in message 
news:Vtcmf4D1g77KFw6S@jasper.org.uk...
> In article , Johnny B Good
>  writes
>
>> At a mere 0.3v that cell has been "flat" for quite a few months. New,
>>they're typically between 3.15 and 3.30 volts, dropping to about 2.9v by
>>the end of their service life
>
> Agreed.
>
>> (with perhaps a month or three left before
>>dropping below 2.5v where the RTC oscillator will stop or else run slow
>>and maybe another 3 months before dropping below 2 volts and causing the
>>cmos settings to corrupt.
>
> Not my experience.  I find that as soon as the voltage drops below 3.0
> the CMOS settings become lost.
>
>> (much bigger than the ones usually found
>>in wristwatches which typically last 7 or more years)
>
> 7 years for a watch battery?  You must tell me what you're using,
> miniature nuclear reactors!?  I usually get 2 yrs max out of a watch.
>
> I have an odd situation with CMOS batteries at work - perhaps you can
> suggest something?  We have an installation with single-board computers
> which control the various parts of a large moving instrument.  It's
> located in a remote, unmanned place where the temperature can range from
> -30C at night to +30C during the day.
>
> The computers originally had 3v CR2032 batteries to maintain the CMOS
> but they were lasting perhaps 18 months max.  Replacing those regularly
> became tiresome (and expensive!) so I made up several plastic holders
> with 2 x AA Duracell Ultras, thinking that would be the last of the
> problem and we'd never need to touch the batteries again.  They're
> connected to the SBC coin holder with spring clips, as there is no
> header to attach external batteries to.
>
> However, the Duracells also run low after about 18 months, leaving me
> scratching my head.  We've pretty much excluded temperature swings as
> the cause.
>
> Any ideas?
>
> -- 
> (\__/)
> (='.'=)  Bunny says Windows 7 is Vi$ta reloaded.
> (")_(")  http://imgs.xkcd.com/comics/windows_7.png
>
>

Some relevant facts and opinions on batteries and life can be
found here http://www.bushwalking.org.au/FAQ/FAQ_Batteries.htm

The message <20091103014054.1fd217b5@bluemoon>
from Rob Morley  contains these words:

> On Mon, 2 Nov 2009 19:34:14 GMT
> Johnny B Good  wrote:

> >  Lithium coin cells (it's not a "battery", just a single cell)

> We know that, but how much easier is it to use the common and incorrect
> term than to explain the difference every time?  :-\

 I do see your point, but the word "battery" happens to be longer than
the word "cell" (although, I'll admit, it is shorter than the phrase
"coin cell" ;-)

 Now that "battery technology" (more accurately, "cell technology") has
made available cells which can provide enough voltage on their own to
replace a battery of old technolgy cells (i.e. a Lithium based cells
producing about the same voltage as a two cell battery based on Zinc
Carbon technolgy), I feel obliged to point out the distinction.

 It doesn't matter so much for disposable primary cell batteries, but it
makes a great deal of difference when using a 3.6 volt single cell
rechargable 'battery' in place of a 3 cell NiMh battery pack. The
difference being that the single cell solution is entirely free of the
problem of reverse charging the weakest cell in the case of the 3 cell
battery pack which damages the cell in question, necessitating premature
replacement of the whole pack when this is allowed to happen.

 Equipment designed to be powered from a rechargable battery pack is
designed to shutdown before such a condition is likely to occur
(usually, in the case of NiCd and NiMH, at the 1 volt per cell mark).
However, the more cells required to make up the required battery
voltage, the more difficult it becomes to avoid this risk (the main
cause for premature battery failure).

HTH

-- 
Regards, John.

 Please remove the "ohggcyht" before replying.
The address has been munged to reject Spam-bots.

The message 
from Mike Tomlinson  contains these words:

> In article , Johnny B Good
>  writes

> > At a mere 0.3v that cell has been "flat" for quite a few months. New,
> >they're typically between 3.15 and 3.30 volts, dropping to about 2.9v by
> >the end of their service life

> Agreed.

> > (with perhaps a month or three left before
> >dropping below 2.5v where the RTC oscillator will stop or else run slow
> >and maybe another 3 months before dropping below 2 volts and causing the
> >cmos settings to corrupt.

> Not my experience.  I find that as soon as the voltage drops below 3.0
> the CMOS settings become lost.

> > (much bigger than the ones usually found
> >in wristwatches which typically last 7 or more years)

> 7 years for a watch battery?  You must tell me what you're using,
> miniature nuclear reactors!?  I usually get 2 yrs max out of a watch.

> I have an odd situation with CMOS batteries at work - perhaps you can
> suggest something?  We have an installation with single-board computers
> which control the various parts of a large moving instrument.  It's
> located in a remote, unmanned place where the temperature can range from
> -30C at night to +30C during the day.

> The computers originally had 3v CR2032 batteries to maintain the CMOS
> but they were lasting perhaps 18 months max.  Replacing those regularly
> became tiresome (and expensive!) so I made up several plastic holders
> with 2 x AA Duracell Ultras, thinking that would be the last of the
> problem and we'd never need to touch the batteries again.  They're
> connected to the SBC coin holder with spring clips, as there is no
> header to attach external batteries to.

> However, the Duracells also run low after about 18 months, leaving me
> scratching my head.  We've pretty much excluded temperature swings as
> the cause.

> Any ideas?

 Have you ever tried logging the actual temperature swings in the
vicinity of the coin cell holders? Are these SBCs being powered down
overnight? Or, do they run 24/7? If it's the latter situation and
they're ATX powered, there shouldn't be any problem (the ATX 5vSB line
takes over from the coin cell as an RTC chip power source - overkill, I
know, but standard practice in an ATX PC).

 I have a feeling it's more likely a problem due to the -30 deg C
temperature. Most battery chemistry starts to decline below 0 deg C
requiring them to be warmed a few degrees above zero to work
sufficiently well enough to allow the kit to they're powering to
function. Admittedly we're talking of supplying watt's worth of power as
opposed to a microwatt or three for an RTC chip.

 I suggest you google for Lithium cell technology voltage versus
temperature data sources before you reconsider a better solution than
using duracells (which will also, likewise, suffer from the same
effect).

HTH

-- 
Regards, John.

 Please remove the "ohggcyht" before replying.
The address has been munged to reject Spam-bots.

In article , Johnny B Good 
wrote:
> The message <20091103014054.1fd217b5@bluemoon>
> from Rob Morley  contains these words:
> 
> > On Mon, 2 Nov 2009 19:34:14 GMT
> > Johnny B Good  wrote:
> 
> > >  Lithium coin cells (it's not a "battery", just a single cell)
> 
> > We know that, but how much easier is it to use the common and
> > incorrect term than to explain the difference every time?  :-\
> 
>  I do see your point, but the word "battery" happens to be longer than
> the word "cell" (although, I'll admit, it is shorter than the phrase
> "coin cell" ;-)

Furthermore, say "cell" to most Americans and they will probably think 
you are using it as an abbreviation for "cellphone". I think it may be 
time to accept that the word "battery" as a synonym for a single cell 
has come to stay.

The New Shorter Oxford English Dictionary defines "battery" as (among 
other things):

| 7 A number of Leyden jars connected together so as to act
| simultaneously. M18–M19.

and

| 8 A device, consisting of one or more cells, in which chemical
| energy is converted into electricity. E19.

Note "one or more", from the early 19th Century. Apparently the 
requirement for there to be more than one applies if the things in 
question are Leyden jars (which are, of course, not cells at all).

I put down my valise.

Cheers,
 Daniel.

In article <EnWHm.83085$R52.64663@newsfe24.ams2>, treblesix <mike.666NOS
PAM@ABUSEntlworld.com> writes

>Some relevant facts and opinions on batteries and life can be
>found here http://www.bushwalking.org.au/FAQ/FAQ_Batteries.htm

Thanks. Bookmarked.

-- 
(\__/)   
(='.'=)  Bunny says Windows 7 is Vi$ta reloaded.
(")_(")  http://imgs.xkcd.com/comics/windows_7.png

On Tue, 03 Nov 2009 22:34:48 +0000, Daniel James wrote:

[...]

> I put down my valise.

LOL!

Chris

-- 
Remove prejudice to reply.

In article , Johnny B Good
 writes

> Have you ever tried logging the actual temperature swings in the
>vicinity of the coin cell holders?

No, but it's do-able, when we have someone on site (it's an unmanned
facility.)

> Are these SBCs being powered down
>overnight?

No, they actually power up overnight :)

> Or, do they run 24/7? 

No, only as needed.  I have argued they would be better left on 24/7 but
there are sound technical reasons why we can't do this.

> I have a feeling it's more likely a problem due to the -30 deg C
>temperature. Most battery chemistry starts to decline below 0 deg C
>requiring them to be warmed a few degrees above zero to work
>sufficiently well enough to allow the kit to they're powering to
>function.

Yes, the link Mike posted was interesting and makes this very point.

We took the low temperature thing up with Duracell but they weren't
particularly helpful.  Ad-hoc experiments with putting batteries in the
freezer overnight and monitoring the voltage were also inconclusive.

> I suggest you google for Lithium cell technology voltage versus
>temperature data sources before you reconsider a better solution than
>using duracells (which will also, likewise, suffer from the same
>effect).

I'll do that, thanks.  Found some in the RS catalogue which were
expensive but looked the trick.  Problem is, they're 3.6V and I wouldn't
feel comfortable with a straight swap for a 3v coin cell.  I'd probably
want to insert a dropper resistor or a diode to bring the terminal
voltage down towards 3.2V.

Another option would be to bring the double-AA holders out of the case
and make them more accessible for routine changes.  If done with the
power on, the CMOS doesn't lose its contents.

Thanks for your thoughts - much appreciated!

-- 
(\__/)   
(='.'=)  Bunny says Windows 7 is Vi$ta reloaded.
(")_(")  http://imgs.xkcd.com/comics/windows_7.png

"Mike Tomlinson"  wrote in message 
news:QEI3VlF5CT8KFwA9@jasper.org.uk...
> In article <EnWHm.83085$R52.64663@newsfe24.ams2>, treblesix <mike.666NOS
> PAM@ABUSEntlworld.com> writes
>
>>Some relevant facts and opinions on batteries and life can be
>>found here http://www.bushwalking.org.au/FAQ/FAQ_Batteries.htm
>
> Thanks. Bookmarked.
>
> -- 
> (\__/)
> (='.'=)  Bunny says Windows 7 is Vi$ta reloaded.
> (")_(")  http://imgs.xkcd.com/comics/windows_7.png
>
>

Another thought in view of thee very low RTC poweer consumption
would be to replace the battery with a low leakage reservoir capacitor
of suitable capacity to run the RTC during down time. If used regularly
say daily/weekly ...then this could be re-charged suitably via system power
while the system is powered on. Alternatively, judging from the temp
variations, a solar powered charging alternative should prove adequate..

At the simplest, only one "charger + cap" would be needed to power
all the RTCs. For redundancy reasons, maybe there could be one or
more "chargers" plus a cap for each RTC onboard. Depends on
cost, context and relevant engineering philosophy ????

Alternatively, with all the usual caveats, it may be possible to simply
devise a sequence whereby the RTCs are re-programmed with the
correct date/time during bootup ... possibly via suitable telemetry
and/or other kit such as GPS.

Could be time to think outside the (battery) box for a solution :0)

In article <PDeIm.26151$ep1.12410@newsfe30.ams2>, treblesix <mike.666NOS
PAM@ABUSEntlworld.com> writes

>Another thought in view of thee very low RTC poweer consumption
>would be to replace the battery with a low leakage reservoir capacitor
>of suitable capacity to run the RTC during down time.

A good idea, but not feasible in our case.  We're dependent on the
weather, which means a cap could discharge in a few days of bad weather
(during which the computers would be powered off.)

> a solar powered charging alternative should prove adequate..

:) An imaginative thought... however the machines are cased and buried
behind access panels (one reason why changing the batteries is such a
pain), and are only powered at night.  During the day, the installation
is protected by a dome.

>At the simplest, only one "charger + cap" would be needed to power
>all the RTCs. For redundancy reasons, maybe there could be one or
>more "chargers" plus a cap for each RTC onboard. Depends on
>cost, context and relevant engineering philosophy ????

More time and money than I want to spend on the job.  Though maybe
supplying 3.2V ish from a permanently-powered supply, distributed to
each node, would be do-able.  Everything is on UPS.

>Alternatively, with all the usual caveats, it may be possible to simply
>devise a sequence whereby the RTCs are re-programmed with the
>correct date/time during bootup ... possibly via suitable telemetry
>and/or other kit such as GPS.

They do sync to a stratum 1 NTP server (using GPS as time reference)
during boot.  The problem is that lost CMOS causes them to stop with an
error at POST.

They are also headless - no keyboard or display is attached.

>Could be time to think outside the (battery) box for a solution :0) 

Pretty much just resigned myself to replacing all the batteries (ten
AAs) during routine maintenance once a year.  Best idea so far has been
to bring the batteries outside the case (perhaps stuck with velcro to
the case front), where they are easier to access and change.

Thanks for your thoughts, much appreciated.  I really thought replacing
the CR2032s with twin AAs was going to be the last we'd hear of the
problem for years!

-- 
Mike Tomlinson

"Mike Tomlinson" <mike@none.invalid> wrote in message 
news:vNizk6AyxZ8KFwGT@none.invalid...
> In article <PDeIm.26151$ep1.12410@newsfe30.ams2>, treblesix <mike.666NOS
> PAM@ABUSEntlworld.com> writes
>
>>Another thought in view of thee very low RTC poweer consumption
>>would be to replace the battery with a low leakage reservoir capacitor
>>of suitable capacity to run the RTC during down time.
>
> A good idea, but not feasible in our case.  We're dependent on the
> weather, which means a cap could discharge in a few days of bad weather
> (during which the computers would be powered off.)
>
>> a solar powered charging alternative should prove adequate..
>
> :) An imaginative thought... however the machines are cased and buried
> behind access panels (one reason why changing the batteries is such a
> pain), and are only powered at night.  During the day, the installation
> is protected by a dome.
>
>>At the simplest, only one "charger + cap" would be needed to power
>>all the RTCs. For redundancy reasons, maybe there could be one or
>>more "chargers" plus a cap for each RTC onboard. Depends on
>>cost, context and relevant engineering philosophy ????
>
> More time and money than I want to spend on the job.  Though maybe
> supplying 3.2V ish from a permanently-powered supply, distributed to
> each node, would be do-able.  Everything is on UPS.
>
>>Alternatively, with all the usual caveats, it may be possible to simply
>>devise a sequence whereby the RTCs are re-programmed with the
>>correct date/time during bootup ... possibly via suitable telemetry
>>and/or other kit such as GPS.
>
> They do sync to a stratum 1 NTP server (using GPS as time reference)
> during boot.  The problem is that lost CMOS causes them to stop with an
> error at POST.
>
> They are also headless - no keyboard or display is attached.
>
>>Could be time to think outside the (battery) box for a solution :0)
>
> Pretty much just resigned myself to replacing all the batteries (ten
> AAs) during routine maintenance once a year.  Best idea so far has been
> to bring the batteries outside the case (perhaps stuck with velcro to
> the case front), where they are easier to access and change.
>
> Thanks for your thoughts, much appreciated.  I really thought replacing
> the CR2032s with twin AAs was going to be the last we'd hear of the
> problem for years!
>
> -- 
> Mike Tomlinson

The UPS supply seems an interesting option Mike.
In adverse contexts these capacitors always proved
superior where others failed often catastrophically
..but I am unable to comment further ;0)
http://www.wima.de/EN/products.htm

You may also be able to source suitable components
etc in the Mil, Aerospace and related component
manufacturers catalogues. Can be problematic tho'
without the "reach" and budget (cough)  nuff said.

Good luck Mike!

In article <vNizk6AyxZ8KFwGT@none.invalid>, Mike Tomlinson wrote:
> ... the machines are cased and buried behind access panels (one
> reason why changing the batteries is such a pain), and are only
> powered at night.

When you say "only powered at night" do you mean that there is no power 
supplied to the boards during the day? Not even 5V standby?

That could have some bearing on your problem. The CMOS batteries (OK, 
'cells') of most PCs provide almost no power at all at any time during 
their lives because the 5V standby line from the PSU powers the CMOS 
chip (as well as the various other things it has to do) so no current 
need be drawn from the cell. If you aren't providing even standby power 
from your PSU then the drain on the CMOS cell is unusually large, and 
the cells' lifetimes will be commensurately short.

Maybe the answer is to replace the CMOS cells with rechargeable ones 
(and provide charging power while the PC is powered on)?

Cheers,
 Daniel.

On Wed, 04 Nov 2009 23:14:55 +0000, Daniel James wrote:

[...]

> When you say "only powered at night" do you mean that there is no power
> supplied to the boards during the day? Not even 5V standby?
> 
> That could have some bearing on your problem. The CMOS batteries (OK,
> 'cells') of most PCs provide almost no power at all at any time during
> their lives because the 5V standby line from the PSU powers the CMOS
> chip (as well as the various other things it has to do) so no current
> need be drawn from the cell. If you aren't providing even standby power
> from your PSU then the drain on the CMOS cell is unusually large, and
> the cells' lifetimes will be commensurately short.

I have a number of machines here at home. I always turn them off at the 
wall socket every night. Some of them only get powered up once or twice a 
week. (The Windows ones!) I have never needed to replace a MB cell. The 
oldest machine is now around five years old. I've just checked in the 
BIOS and the voltage is showing as 3.

I don't think this is the solution to the OP's problem.

Chris

-- 
Remove prejudice to reply.

In article <VA.0000002e.00dafb05@me.invalid>, Daniel James
<daniel@me.invalid> writes

>When you say "only powered at night" do you mean that there is no power 
>supplied to the boards during the day? Not even 5V standby?

That's correct.  Elfin Safety is the reason.

>That could have some bearing on your problem. The CMOS batteries (OK, 
>'cells') of most PCs provide almost no power at all at any time during 
>their lives because the 5V standby line from the PSU powers the CMOS 
>chip (as well as the various other things it has to do) so no current 
>need be drawn from the cell. If you aren't providing even standby power 
>from your PSU then the drain on the CMOS cell is unusually large, and 
>the cells' lifetimes will be commensurately short.

That's interesting (obvious when you think about it!) but why do
Duracell AAs in pairs last only about the same time as the CR2032s - 18
months or so?  Would have thought their capacity was much higher.

>Maybe the answer is to replace the CMOS cells with rechargeable ones 
>(and provide charging power while the PC is powered on)?

Hmmm.  I was hoping for a simple solution.  Mike's (indirect) suggestion
of supplying 3.2V from a permanent power backed by UPS looks a good one
in terms of effort and cost.  I suspect though, we'll just add replacing
the batteries annually to the regular maintenance checks, applying a
modification to make them more accessible.

Thanks for your thoughts.

-- 
Mike Tomlinson

In article <LqwIm.4932$uf7.3071@newsfe12.ams2>, Chris Whelan wrote:
> I have a number of machines here at home. I always turn them off at
> the wall socket every night. Some of them only get powered up once
> or twice a week. (The Windows ones!) I have never needed to replace
> a MB cell.

I've had MB cells fail ... but I've also had them last for ages in 
machines that were boxed and in the attic. It seems to depend to some 
extent on the actual hardware.

Of course, I doubt that either of us is subjecting our PCs to the same 
extremes of temperature as the OP, and that surely takes its toll.

Cheers,
 Daniel.

In article <0eRFQ+B87q8KFw1m@none.invalid>, Mike Tomlinson wrote:
> That's interesting (obvious when you think about it!) but why do
> Duracell AAs in pairs last only about the same time as the CR2032s
> ...

No idea ... but they are a different battery technology; completely 
different chemistry.

Cheers,
 Daniel.