CMOS checksum fails

Not sure what could be causing it.

Seems to happen after power loss... and hard reset... also after weird electrical problem which happens rarely, which is solved by full power down, disconnect of cables etc.

After saving settings and exiting bios, PC works as normal.

Checksum error indicates some bits flipped ?

What exactly is checksummed for CMOS ?

Bye, Skybuck.

Reply to
Skybuck Flying
Loading thread data ...

I guess it's battery:

formatting link

Kinda weird... it seems to keep some settings but maybe that's not true.

The clock/date is reset to 2005 causing weird problems in windows and apps which is easily solved.

So I guess that's a clear symptom that battery might be dead.

I ll try replacing it sometime in future perhaps and see if that solves it.

Bye, Skybuck.

Reply to
Skybuck Flying

Now is a good time to change it.

The CR2032 is available at all sorts of shops. It's not hard to find. Verify the writing on the top, mentions CR2032.

When the voltage of that battery drops below around

2.3V, then there may not be enough voltage to run that circuit. On some motherboards, if the CR2032 drops to zero volts, the motherboard will no longer boot. It is unknown exactly why that happens (haven't seen an explanation). Not every motherboard exhibits those symptoms. I've had completely dead CR2032 here, which had no impact at all on eventually getting the system running (after setting the clock and entering the correct boot order).

The only thing I would recommend, is some care when working with the CR2032 socket on the motherboard. One poster, I told him how "easy" it was to replace... and he ripped the socket right off the motherboard. Um, don't do that... OK?

Paul

Reply to
Paul

A run-down battery is the most common cause of those symptoms. Checking and replacing it is a trivial matter for someone actively involved in electronics.

Not all motherboards behave the same way. Some will work normally when the 3V battery is almost down to 2V and others exhibit symptoms at 2.8V. Some batteries need replacement after 2 years while some go on merrily for 5 yrs.

But the battery's not the only possible cause. Some BIOS chips get corrupted or lose their ability to retain user settings even with a new battery. Accumulated dust and moisture can cause that too.

Reply to
Pimpom

Those are probably default BIOS settings.

The battery might not need replacing. If the motherboard hasn't been used in a long time, the battery may simply need recharging. You can do this by powering the board for a few hours, and then using it daily for a week or so. Then, don't use the motherboard for a month or so. If the settings are lost after a month without use, replace the battery.

You might also want to check that the CMOS reset jumper is in the correct location. One position preserves the settings. The other position resets the CMOS and can cause wierd quirks with some motherboard, or damage the CMOS circuitry on others.

Your first post may indicate the board is failing electrically, e.g., power supply failing or faulty memory or heat damaged processor, or perhaps it's an older board that hasn't been used in a long time. If the latter, it my just need to be used a for a while before it starts functioning normally.

I saw these types of problems many years ago with motherboards in the late 1990's, but I haven't seen any of these issues a long time. Supposedly, those motherboards had issues with cheap capacitors and static shock. They also didn't use thermal grease on the processors. Today, the main problem is BGAs unsoldering due to high heat, e.g., video cards, which you can sometimes hot air reflow using a hair dryer. They'll unsolder again though, if you don't remove the excessive heat.

Rod Pemberton

--
Cars kill more people than guns in the U.S. 
Yet, no one is trying to take away your car.
Reply to
Rod Pemberton

I'm afraid I have to disagree. Motherboards use primary (non-rechargeable) CR2032 batteries. As with most primary batteries, they may partially recover when they are left unloaded. This is what happens when the computer is turned on and the power supply takes over the job of supplying the tiny amount of power needed to retain the BIOS settings. This is why a CMOS battery actually lasts longer if the computer is in regular use.

That said, it is often possible to rejuvenate a primary cell to some extent by charging it, but it's not as good as recharging a secondary cell And the practice is often discouraged because it can produce undesireable consequences, even to the extent of exploding.

Reply to
Pimpom

The CR2032 is not intended for recharging. The spec sheet for one, shows a maximum current flow into the battery ("attempt to recharge") limited to 1 microamp. And it is likely set that high, only to allow diode based current steering on motherboards, not because they actually want people jamming that tiny current into the battery. They'd probably prefer the number to be zero, but then that would make it hard to design motherboards. The 1 microamp limit allows for reverse biased diode leakage at elevated temperature (for the typical Schottky diode used in that circuit). One feature of the Schottky diode in the battery path, is to prevent current flow into the battery.

formatting link

Max Rev Charge: 1 microampere

formatting link

Fig.2 Ta=125C (ridiculously high temp), extrapolate back to less than one volt reverse potential, gives 1000nA or 1uA leakage. The "real" Ta value is lower, and so will the current flow be lower.

So the trick then, is to examine your motherboard first, take note of the part number on the battery, to decide just exactly what you're dealing with. CR2032 and LR2032 are not the same thing. This case is likely to be CR2032, but the OP can figure this out for himself.

CR2032 = not rechargeable = buy a new one when your CMOS clock fails to hold the time properly

For any other battery type you might find on the motherboard, consult your local computer shop for help. For example, I have an old Apple computer, with a relatively large cell (not a CR2032), which is good for around ten years. And mine finally went flat. Replacements for that one are still available. If you see something which is not a CR2032 coin cell, ask for help.

Paul

Reply to
Paul

It's my Dream PC from 2006, repaired and down graded throughout the years (until 2015 soon I may build a new computer since 2016 is coming and I promised myself a new computer for 2016 ;))

It's on basically every day for 10 hours or so at least.

I put the computer into sleep mode.

Perhaps sleep mode keeps the CMOS settings alive ?

Bye, Skybuck.

Reply to
Skybuck Flying

Den mandag den 4. maj 2015 kl. 18.10.11 UTC+2 skrev Skybuck Flying:

pc supplies have since forever had a low current 5V standby voltage that is always on when the supply is powered, I would assume the MB manufacturers are smart enough to use that to power the cmos when it is available

-Lasse

Reply to
Lasse Langwadt Christensen

The battery is only used, if no power is available. The battery is used if you switch off the computer at the back, or you unplug the computer.

On an Asus motherboard with that green LED, as long as the green LED is glowing, the battery is not being used.

A CR2032 has 240maH capacity. The motherboard draws 10uA for the CMOS RAM and RTC. 240maH / 10uA = 24000 hours of battery time.

formatting link

If you unplug the computer for three years, expect the battery to be flat.

By sleeping the computer, the computer stays powered and the CMOS battery is not being drained as a result. The battery has a shelf life of 10 years, and by using sleep, then the battery would be expected to last for 10 years.

If you leave the computer unpowered at night, then that is 8 hours off, 16 hours on, you can do the math to see how many years you get from the 24000 hours of battery runtime.

*******

This is an example of a rechargeable coin cell. This is *not* a drop-in replacement for a CR2032.

formatting link

It has a higher open-circuit voltage, and a lower milliamp hour rating (until the circuit board it is on, recharges it again automatically when power is available).

It lasts for fewer days in a "no-power" situation, as that one is only 40maH.

You might find something like that in a laptop. In which case, take care to replace it with the same type of battery.

Paul

Reply to
Paul

I have seen both used on motherboards. But the rechargeable is normally a small stack of 3 x 1.2v NiCd. After a few years they tend to leak and ruin the main board shortly after that.

Reply to
David Eather

So, with 14 hours of off-time per day and 24,000 hours of CMOS power according to Paul, your battery should've failed 4.7 years after 2006, i.e., mid

2010, instead of 9 years later in early 2015, if it wasn't being recharged ... I also have a computer which operated for about a decade with many more hours of off-time. The CR2032 never failed. Even if we reduce the CMOS amp draw by half, you can easily compute that the battery should've failed at about 7.5 years. I also know first hand that the technique I mentioned earlier on recharging the CMOS battery works. Of course, it used a battery much larger in diameter than the CR2032. Given that the CR2032s are apparently not dying as fast as they should for a supposedly non-rechargeable battery, I think it's valid to conclude something else is going on with Li-ion CR2032s : 1) battery draw is much, much lower than expected for CMOS, or 2) the Li-ion CR2032's are actually being recharged from the motherboard. Wikipedia says ATX motherboards continue to power CMOS as long as the machine is plugged in. So, it wouldn't be hours of off-time where CMOS consumes power, but hours of unplugged-time for ATX machines where CMOS consumes power. That could extend the battery life to many decades. Of course, your motherboard might use something other than CR2032, most of which are actually rechargeable but recquire desoldering.

Rod Pemberton

--
Cars kill more people than guns in the U.S. 
Yet, no one is trying to take away your car.
Reply to
Rod Pemberton

My DreamPC died a couple of times.

Last time it died and was repaired was 30 november 2012.

I fried the motherboard by misplacing a usb connector on a usb header.

The fried motherboard was replaced with a "second hand" motherboard from a special dealer who trades in used electronics.

So this motherboard has had power for 3 years almost non-stop, that's why I am a little bit surprised that the battery might be dead.

When I think about this possibility it seems unlikely to me, hence my surprise.

Perhaps there is something else causing the checksum error.

For now I will assume the battery is the problem.

But I wouldn't be surprised if it's some strange electrical or magnetic interference problem.

For now it's not a problem to leave the "dead" battery in.

I rather leave it in then risking any further damage to my computer.

Maybe when I buy a new computer, I might replace the battery of the old one... just to see if this was indeed the problem ;)

Bye, Skybuck.

Reply to
Skybuck Flying

Let's suppose for a moment that the battery should not have failed and it's indeed "dead"/out of power.

Perhaps there is a short circuit on the motherboard being caused by dusts, hairs or other small little elements...

Perhaps small little metal atoms blown in.

Perhaps this dust could lead to a very small short circuit... drawing away power from the battery.

Just a little theorie/hypothesis to explain the short life of the battery.

Bye, Skybuck.

Reply to
Skybuck Flying

On Tue, 5 May 2015 11:55:03 +0200, "Skybuck Flying" Gave us:

You are a true idiot. Go piss and moan in a children's group.

Reply to
DecadentLinuxUserNumeroUno

On Tue, 5 May 2015 11:58:17 +0200, "Skybuck Flying" Gave us:

I would suggest taking an electronics course, but that too is miles above your brain grade.

Reply to
DecadentLinuxUserNumeroUno

Fact: The vast majority of college students have no idea how a flashlight works. The concept of "circuit" is not familiar to them.

Reply to
Bill

'Flashlight' is an app on your smartphone. It works using servers and app stores and for some reason wants to know your location.

--
Spehro Pefhany
Reply to
Spehro Pefhany

Because it wants to be smarter !

Jamie

Reply to
M Philbrook

Was that an invite to a private party of yours?

Jamie

Reply to
M Philbrook

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.