Cmos battery test.

[cyberwasp]

I hope I'm in the right place. I haven't been building anything for a long time and my scope has been lonely and the old brain isn't recalling things to well. Recently I was having difficulty with my PC retaining its Cmos settings. Since being handicapped I waited for a friend to visit to assist me with changing it.

Using a DVM the old 2032 shows 2.97v.  I'd like to make a simple circuit that would simulate the cmos drain that I could monitor with the dvm or scope;) I know it sounds crazy and that's ok. Would just like to see what happens. I have a bunch of electronic parts including several arduinos if needed. the simpler the better.

I was thinking just an led/resistor in series with the dvm or scope unless there's something better!!  TIA

[bob91343]

There should be no drain from the battery.  It's just a polarizing voltage and only miniscule leakage current is taken from the battery, which essentially should last for its shelf life.

[cyberwasp]

There should be no drain from the battery.  It's just a polarizing voltage and only miniscule leakage current is taken from the battery, which essentially should last for its shelf life.

The old one was roughly 5 years, not including how long the motherboard was sitting in the warehouse. I was just thinking it would be something to get me back into experimenting!

[m3vuv]

seems like a waste of time regarding how much they cost,im suprised the cmos lost its marbles at that voltage,ive had them under 2v and still keeping the cmos data!

[Vovk_Z]

The capacity of 2032 battery is about 200 mA*h. It may work up to 5-10 years, let's take 10 years (with a good battery with small self-discharge). That is 87600 hours. So CMOS current is about 200 mA*h / 87600 h = 2 uA.

[BrokenYugo]

seems like a waste of time regarding how much they cost,im suprised the cmos lost its marbles at that voltage,ive had them under 2v and still keeping the cmos data!

I actually have one in a Super Nintendo cartridge that wouldn't hold at something like 2.9 or 3.0. A new battery fixed it and hasn't died yet, so I assume the SRAM is fine otherwise.

[Shock]

With infrequent use and no other parasitic load problem 5 years is about how long they last disconnected from the mains.

[Per Hansson]

One thing not mentioned: when measuring battery voltage you should apply some load.
Example: I have three CR2032 batteries that have been used for CMOS and RTC backup in mainboards till they failed.
With a 10M ohm input resistance DMM they measure 2,9v, 2.7v and 2.2v respectively.
With a 3K ohm input resistance meter (resulting in 1mA load) they measure 0.0v, 0.0v and 1.4v respectively.

[jmelson]

There should be no drain from the battery.  It's just a polarizing voltage and only miniscule leakage current is taken from the battery, which essentially should last for its shelf life.

No, on almost any PC, the CMOS battery also runs the real time clock.  That has a 32768 Hz crystal oscillator and divider chain, running a time counter.
Not a lot of drain, but not really zero, either.
Jon

[viperidae]

Intel suggests a current of 5uA for the RTC of their PCH chipset. Which works out to be nearly 4 years for a 170mA cell.
Power is also only drawn while the standby motherboard power is disconnected.