Announcement

**Phaihn** · 08-16-2009, 11:35 AM

Re: CMOS battery mystery

did you try replacing the battery? i know it says its good but might want to try anyway.

**Newbie2** · 08-16-2009, 11:53 AM

Re: CMOS battery mystery

Originally posted by Phaihn

did you try replacing the battery? i know it says its good but might want to try anyway.

I think he already tried that: "even with a brand new and good CR2032 installed, the BIOS would say the settings were invalid and clear the CMOS"

**gg1978** · 08-16-2009, 03:01 PM

Re: CMOS battery mystery

I'm assuming you have the clear CMOS jumper in the correct position..

Other than that, faulty southbridge/battery circuitry..

Might try pulling power cord , and the CMOS battery at the same time, to get a "hard" reset of CMOS memory, for 15 or 20 minutes, then put CMOS battery back in, and plug back in.. Reset all your settings, and try it again..

**Harvey** · 08-16-2009, 04:12 PM

Re: CMOS battery mystery

If its one of those bent-over-top battery connections, take the battery out, slightly bend the top spring down and try again. The current taken by the backup supply is so low that even the slightest bad connection can cause this fault. I know, I've had it before and its a right bugger to fix...

**pentium** · 08-16-2009, 05:33 PM

Re: CMOS battery mystery

I'm assuming you have the clear CMOS jumper in the correct position..

If that was set the system would not POST at all.

Can we assume bad battery circuitry? Is thre a way to fix something like that?

Originally posted by Harvey

If its one of those bent-over-top battery connections, take the battery out, slightly bend the top spring down and try again. The current taken by the backup supply is so low that even the slightest bad connection can cause this fault. I know, I've had it before and its a right bugger to fix...

You mean one of these?

I don't see why there would be a current draw fault if the system can see its voltage (though voltage and current are completely different).

**370forlife** · 08-16-2009, 05:42 PM

Re: CMOS battery mystery

There are those that have a big clip that look like a bobby pin that go over the battery and are one lead, and at the same time they push down the battery onto the other.

They can get worn out, and you bios may be able to see the battery as good, but the voltage is intermittent.

**weirdlookinguy** · 08-16-2009, 06:40 PM

Re: CMOS battery mystery

You said it measured 3.01V. That's not a good battery. Should be 3.3V.

**gg1978** · 08-16-2009, 07:44 PM

Re: CMOS battery mystery

~3V should be enough to retain the settings.. Most boards have to have less than 2.5V to drop the settings.. Course it's dependant on the design of the motherboard circuitry. Remember that battery voltage either open or in-circuit can change a bit..

Have you tried the hard CMOS reset i described? I have an Abit board that that worked on, it would just drop it's CMOS settings, even with a good battery. Now i can leave it totally unplugged from an ATX psu, and it keeps its settings..

As for the circuitry, the implementations can vary. It'll be close by typically.. Usually you'll see at least a couple of diodes, resistors, etc.. Then again some southbridges may have the battery directly connected. Basically remember that the battery is only in use when the AC power is totally off, as in the PC is unplugged..

So if this computer is dropping the CMOS settings with the computer "off" but with the ATX PSU still supplying +5VSB (soft off), it's probably a dud southbridge.. Rare type of fault, but it happens..

**gg1978** · 08-16-2009, 08:17 PM

Re: CMOS battery mystery

Ok, i just went through my pdf archive, and it so happens i have a BX chipset motherboard schematic reference design, from Intel.. Shows Vbat from the soutbridge as being directly connected to the center pin of the CMOS jumper, with one pin going to ground, for when you set the jumper to "clear", and the other position connects Vbat to a simple circuit using schottkey diodes to have Vbat come from either the 2032, or 5VSB.. They're basically diode OR'd together, just like setting up current sharing for power supplies.. The diode network prevents the +5VSB from damaging the battery, and to prevent the battery discharging into the +5VSB when the ATX PSU is all the way off..

Here's the link to the BX reference schematic if you're interested in how the circuit works.. It's on the part of the schematic for the southdbridge.. Just look for the Vbat pin, and you'll find the rest..

Unfortunatly, Intel doesn't seem to have a reference schematic for an i865 reference design, which most motherboard manufacturer's follow at least somewhat when they design motherboards..

http://www.intel.com/design/archives.../440/index.htm

Luckily i almost never delete files off my hard drives, and that i remembered i had it..

The only difference i can think of between BX and i865 is that on an i865, the +5VSB is dropped into an LDO regulator, that drops it to 3.3V, hence they call it +3.3VSB instead.. Same purpose though, it's just a lower voltage because the later southbridges can't tolerate the direct +5VSB, it'll burn them out..

**Wizard** · 08-16-2009, 08:27 PM

Re: CMOS battery mystery

Yes, 3.0V. Those are what lithium cell coins are.

3.3V doesn't exist, that's for voltage supply of running chips. There are 3.4, 3.6 and 3.7 (Li Ion) but in wrong form factor.

Pentium, your lithium coin is GOOD by measured voltage. And I'll just refrain my comments on where you gotten that board. I'll just wait for you to come to realization some doing of things is fruitless.

I also advise you to trace the circuit on that mainboard and see if you can trace to the southbridge from that battery, you'll find some circuits and one 32KHz crystal (this is time keeping clock). Did time keep correct when unplugged from power cord and BIOS setting intact, say a 2 hours?

BTW, this 32KHz crystal is same type used in quart watches and quartz clocks.

Cheers, Wizard

**gg1978** · 08-16-2009, 08:32 PM

Re: CMOS battery mystery

Just thought of this.. Try a new jumper.. If the jumper isn't gripping the pins of the CMOS header tightly, a loose connection may be causing your issues..

**Wizard** · 08-16-2009, 08:35 PM

Re: CMOS battery mystery

Good point on jumper block. I salavage quality jumpers long way back when these jumpers were really the stuff of quality not like we have now found on everything.

Cheers, Wizard

**ratdude747** · 08-16-2009, 08:47 PM

Re: CMOS battery mystery

i like mobos with DIP switches. no jumpers to lose and if you need to add one when you have none available it really helps.

**pentium** · 08-16-2009, 09:38 PM

Re: CMOS battery mystery

nope, even after ten minutes the clock dies as well :/
so you are saying that if I run a line from the + in the battery clip to the center post on the CMOS reset jumper block, I can use that as an alternate route to keep the southbridge and thus the CMOS alive?
Sorry, I'm a little confused. I'm more used to pictures and the likes describing something. Lots of text just seems to confuse me at times.
also, no, I think the jumper is good. It's a snug fit onto the jumpers at least.
also, as a note, I can't really tear the system apart to look at traces. When my replacement ram comes in (that's another thread which we don't need to talk abou now) I'll tear the system apart and then I can look at whatever you want me to but until then...

**weirdlookinguy** · 08-16-2009, 10:51 PM

Re: CMOS battery mystery

Ah, my bad on the 3.3v battery. I always assumed that "3.3vsb" came from the battery, not from the 5v line on the PSU. Learn something new every day.

**LesCoke** · 08-17-2009, 12:57 AM

Re: CMOS battery mystery

Check the battery with a DMM that has a battery test setting. Most batteries (especially coin cells) will continue to show close to full voltage when tested with an open circuit. The battery test meters place a small load on the battery while showing the voltage.

If you do not have a DMM with battery testing capabilities, one of the cheap analog meters can be used, they tend to have a lower input impedance. You can also use a 330 ohm resister across the cell while measuring the voltage with a DMM.

When I'm working on a system, I usually check the battery. Anything below 3V, I go ahead and replace, otherwise the machine will come back in with a dead battery within a few months.

Les

**ratdude747** · 08-17-2009, 01:00 AM

Re: CMOS battery mystery

did you measure voltages with the machine off and power supply discharged? if not give it a hard reboot and dont flip back the power, then test. that way you are measuring the battery only.

never had this happen but i can only imagine...

now THAT is a s***ty computer!

**Junk Parts** · 08-17-2009, 05:36 AM

Re: CMOS battery mystery

All I can tell you is that my dells go queer when the CMOS battery gets weak and they start doing exactly what your PC is doing. Just for GP I'd change out the CMOS jumper its self and replace the battery and try again. I'd also take a real careful look in the case for anything that could be shorting the mobo to ground.

**pentium** · 08-17-2009, 08:42 AM

Re: CMOS battery mystery

Before this cell saw use in the current board, it saw temporary use in my older K7N2 Delta2 and it worked flawlessly. It held settings and held a charge both with the PSU energized and with the PSU discharged. I ahve also tried SEVERAL other CR2032 cells and unless they are all dead as well, the system is still clearing the CMOS when the PSU is discharged for any amount of time.
Would it really be that painful just to solder in a line directly to the southbridge from the battery and bypass the existing trace?

Announcement

CMOS battery mystery

CMOS battery mystery

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment