Suspect my CMOS battery is dying

Starting yesterday, I noticed my laptop clock is lagging. I'm unsure... but I could guess the loss of a second every 5 hours or so. Nevertheless, I'm almost positive it is slowing down by synchronizing it with my watch and waiting a few hours.

I have no way of measuring power voltage on my laptop CMOS battery. But do you reckon this is a clear sign the battery is dying? This laptop is well over 5 years, on its way to 6. I remember once thinking about this potential problem and thinking to myself it would be ok because probably my CMOS battery was rechargeable...

Is this happening when the laptop is powered or unpowered?

If it's when you have battery/AC power, then it's 99% sure that it's NOT the CMOS battery, as that is only in use when you have "no power".

--
Mats

Powered yes...

I'm not sure if that is good or bad news. *starting to titter*

I'm not sure if it's good or bad either.

What OS is this? How do you know it's the CMOS clock that is changing, rather than the system time?

--
Mats

Well, good point. I'm looking at the system time, mind you.
Guess I need to test this with my laptop shutdown -- which I will tonight; can't afford to have it turned off for 6 hours now.

This is on Windows. I haven't check it on Linux yet. But I don't see how Windows could possibly slow down the clock...

It does seem strange indeed. I'm sorry I have no suggestions - I just happen have have studied enough schematics of PC motherboards to know that there is a transistor that detects "power" that essentially shuts off the CMOS battery from the CMOS clock whenever there is another power-source - it's only there to keep the CMOS clock going when you turn the power OFF.

--
Mats

Thanks matsp. I'll do some more checks tonight and see if I can get more detailed information about what is going on... if anything, probably it's me going crazy.

Modern systems dont use the FET anymore, its all integrated into the RTC chip. You may just need to replace the RTC, it sounds like the oscillator is drifting with age. This can happen with inexpensive RTC's, as they use capacitors rather than quartz. Over time the capacitor value drifts, and so the frequency drifts. 1 second every 6 hours is a pretty big drift though, so its probably more than that. Bottom line, first replace the coin battery, if that doesnt fix it replace the RTC. The battery is cheaper, and you want a new one if you replace the RTC anyway, so try that first.

Replacing the RTC will be difficult if it's integrated into the chipset (which has been the case for many years).

But I agree, replacing the battery won't cost much - I got 2 x CR2032 for £1.30 the other day - that's about €2 or $2.50, give or take a bit.

Of course, if it's a laptop, it can be hard to get to the battery... :(

--
Mats

Replacing the RTC is way out of the question on this computer. I'd rather just enable the time synchronization service and forget about it. Replacing the battery will be a bummer though, but probably that's all this is.

It will be a bummer because Toshiba support sucks in Portugal... badly.

Sync every hour and the 1 second every few hours is really insignificant. Or are you worried that the time dialation could become more serious?

Put it this way, you get to use trial software for longer than everyone else :)

Quote:

---

Originally Posted by zacs7

Sync every hour and the 1 second every few hours is really insignificant. Or are you worried that the time dialation could become more serious?

---

More that something must be wrong because it's happening :)

It's my CMOS battery, no doubt now. Starting yesterday, when I power the laptop on the clock resets. The battery died for certain. I've already sent it to get a new battery with a strong warning not to disturb any wild life they may find inside.

Thanks everyone.

That doesn't confirm it was the battery, it could have been a dry solder for example.

And that makes just as much sense. Ooh well, I hope the spiders inside your laptop will be okay.

Yes. And could be that my wrist watch was running too fast and the time server packet sent to my machine was corrupted. That would make sense too.

But... it's the battery.