CMOS battery circuit diagram?

[stf92]

Hi: in a PC motherboard the CMOS RAM keeping the BIOS parameters is fed by a 3.0 battery. What is this circuit exactly? I have searched in Google to no avail.

 

[Slobodan-888]

EEPROM chip is powered directly from 3V battery, without any voltage converters. If that is what you are asking.

 

[stf92]

Errr... EEPROM? What is the battery then for? I read the article in Wikipedia and found the old CMOS RAM is now an EEPROM, alright, but my motherboard is from 2003, so maybe its still the traditional RAM. The battery was needed to power the RAM.

The first question, on second thought, I think you already answered it. But the E2 is supposed to be written when the machine is on, and here a battery is not needed. And the E2 I can put it in my pocket and still will have the data there, with no battery!

EDIT: Well, I guess it needs it to keep the time, of course! Are there any capacitors involved in the circuit, besides decoupling ones?

 

[Slobodan-888]

Well, I don't really know what type of memory is used for storing BIOS. But I assume that is EEPROM, since it is non-volatile. BIOS itself is stored in it, but the BIOS settings are stored in some volatile memory (probably CMOS RAM).

So, the EEPROM is not actually powered from the battery (since there is no need for it), but CMOS RAM is.
I have never seen schematics for it, but I assume that is is directly powered from the battery (without any DC/DC converters). There could be a supercapacitor, maybe. But since there is a battery, I don't see a need for it. RTC (real-time clock) is powered from the same battery.

RTC schematics looks something like this: http://openenergymonitor.org/emon/sites/default/files/RTC.png

 

[stf92]

Thanks for the circuit. See this from wikipedia, "Nonvolatile BIOS memory":

The CMOS RAM and the real-time clock have been integrated as a part of the southbridge chipset and it may not be a standalone chip on modern motherboards.

The problem is I don't know if a 2003 motherboard can be considered modern. But I think your remark about capacitors is what I needed. I'll assume there are no capacitors there and be left in the dark with this scenario: the battery reads 3.05 volts, when plugged into its socket but after having the machine turned off for an hour the BIOS settings are gone. I thought capacitors could be guilty and thats why my post #1. In fact, after only one minute being off, the BIOS won't beep. I have to push reset one or two times to make it come alive (beep). Once operating correctly, if I remove the power for a second and reapply it, the machine will boot correctly. So I began replacing capacitors. The main suspects
, because of bad reputation of the brand, were six 1500/6.3 caps. of the OST brand. I replaced them by identical ones, another manufacturer, to no effect.

The BIOS settings problem seemed to be an independent one, so I wanted to attack it.

 

[norsestar]

Are you measuring battery voltage from the pads on the board (or better, pins on the southbridge)? The battery may not be making good contact in the socket, or a solder joint or trace could be cracked.

Have you tried replacing the battery? 3.05V should be okay but it's cheap to try a fresh one just in case. Is the 3.05V after it's been off for the hour needed to lose the settings?

 

[Slobodan-888]

I see your problem, now. It could be faulty/dead battery. Batteries can read good voltages without any load on them, but as soon as you load them (the dead batteries, I mean) their voltage will drop, and they will discharge fast. But, I am not really sure about this, because in this case current draw (load) will be very small (in microamps). But then again, if battery is dead, its internal resistance have become high enough to cause significant voltage drop, even with microamps load.

There is also a possibility that there is some extra load (that shouldn't be there) across the battery (besides CMOS RAM, or whatever it is used).

Try with a new battery.

Those electrolytic capacitors you have replaced have nothing to do with that. When they go bad, motherboard will not work at all (due to increased ripple). Make sure that those new caps you have installed are low ESR ones (low equivalent series resistance), otherwise they will cause more ripple, heat up more, and fail in relatively short time.

 

[oliver89]

stf92, you’re actually right, once the data has been written, it’ll keep the data safe until re-written. That’s precisely because this EEPROM is a non-volatile RAM and doesn’t need lose the data written inside once the RAM loses power. However, the battery is required in-order to keep the time and also, in some cases a bit of additional information such as whether or not to turn the CPU on when PSU gets powered up without the user having to press the power button and some other custom programmed functions.

pcba manufacturing