Eliminating "CMOS" battery

[Red Squirrel]

Even better idea. Have the CMOS run from a supercapacitor that's recharged every time power is applied to the motherboard. The Dreamcast did something similar.

That could probably work well actually, have it stay charged off standby power even.

 

[nerp]

Even better idea. Have the CMOS run from a supercapacitor that's recharged every time power is applied to the motherboard. The Dreamcast did something similar.

That would be pretty cool, actually.

 

[Heatshiver]

A CMOS battery costs $2. It takes two seconds to replace.

2 seconds on a desktop...

 

[Deders]

Come to think of it, I have opened a Sony Vaio (can't remember which one exactly) where I couldn't find the battery anywhere.

 

[tortillasoup]

this sort of thinking is what lead to USB touchpad mice vs PS/2. PS/2 based stuff is much more reliable/faster than USB crap.

Super capacitor doesn't work, they have strong self-discharge properties so they won't last for months/years at a time w/o being hooked up. Super Capacitors have low density compared with a lithium button cell battery.

 

[nerp]

2 seconds on a desktop...

Sure, but most laptops are made to throw away. Unless you get a ThinkPad and that's no more difficult to open and service than a desktop.

Also, if you have to replace a laptop battery after 5-10 years, that 20 minutes of time isn't such a huge deal, especially when many enthusiasts will crack open a laptop to upgrade the RAM and SSD, as long as they're not welded to the motherboard.

 

[Smoblikat]

Those could be made persistent in Flash, for example. Over-rideable back to defaults with jumpers.

I have no issues with keeping the RTC, but why not make the battery optional? So those of us who don't want a RTC don't have to suffer having to perform maintenance on it?

You dont perform maintenance on an RTC.......

Its a non-issue, as in ive never had to think about it. Out of the hundreds and thousands of computers ive serviced over the years, I have never in my life seen a CMOS battery unexpectedly die or even cause any particularly challenging issues. It takes maybe a minute total to put a new battery in, and put the correct date/time back in. Lets say you live to be 100 years old, assuming you started at birth and each battery lasts 5 years, you would spend roughly 20 minutes out of every century dealing with batteries. I guarantee writing the code to get the clock signal from a (much more latent) server would take longer than 20 minutes.

 

[mikeymikec]

Even better idea. Have the CMOS run from a supercapacitor that's recharged every time power is applied to the motherboard. The Dreamcast did something similar.

I imagine the reasons why this isn't done are as follows:

1 - it probably costs more
2 - it requires more complex electronics
3 - if it ain't broke, don't fix it (though if Asus followed that logic, they wouldn't have decided to completely re-arrange the front panel connector pin arrangement after having it a certain way for decades!).

 

[Lorne]

Seriously, Complaining that you have to open a case up to change a battery every 4 or so years is to often.
You should have them open once a year to blow them out, sharpie a date on the inside the case and change the battery on the 4th cleaning.

The cap back CMOS was tried already, It failed.
I cant remember a lot about other then one of the computer stores I worked in switch MB that used them and our customer service came back screaming at us to switch to some other MB fast.
Problems were they wouldn't hold a charge for more then 16 hrs and then one of the legacy companies (Maybe ISA/VESA I don't remember) made rules that they had to last more then 7 days or so.
It was cheaper just to use the battery.
There are BIOS that save like a static firmware, The data info can only be reset if you hook a battery to it, But again cost is higher.

 

[pitz]

Seriously, Complaining that you have to open a case up to change a battery every 4 or so years is to often.
You should have them open once a year to blow them out, sharpie a date on the inside the case and change the battery on the 4th cleaning.

Some of us work with computers in embedded applications, and its a lot of labour and troubleshooting, at an unwanted time. I can change the batteries with ease in my home computer. But what about a computer used for digital signage that's mounted 25 feet in the air? Or various applications in control systems or surveillance, where a fan-less PC in a small enclosure would work perfectly fine? Or set-top boxes?

PC's are lasting a *lot* longer in a condition that they're perfectly useful. For instance, my i7-2600 cost very little money 5 years ago and will easily be good for another 5 years, meaning potentially 2 battery replacements. The technology exists to allow a PC to acquire time from the network, boot, and run. Yet nearly all PC's shipped today will not operate or will not operate properly and without a lot of manual user intervention with an absent or dead CMOS battery. This is one thing I'd like to see changed.

 

[pm]

As a guy who has worked on the RTC on a modern SOC, I actually looked into this. There are lots of ARM-based computers without RTC's (and their batteries) and they get time as you suggested or similar methods.

The problem, as I understood it, when I asked repeatedly "why do we still use this ancient thing?" was, as I was told when we looked into it, was the same problem you have with all things x86 - which is that there are some legacy uses that still require it, some vendors still adamantly want it, and so when you say "hey, we are getting rid of it", they say "well, we aren't buying that chip then" or worse "we want your chip and we want RTC, we buy lots of your stuff, make it happen". A couple of our customers needed accurate time for systems that didn't get plugged in to the network or needed accurate time prior to accessing the network as part of the network access handshake, and so getting rid of it killed their business model.

Honestly, the RTC is a super irritating chunk of circuitry to put onto an IC - the voltage is super high for a 14nm CPU and getting it working is a very small part "fun engineering challenge" and a very large part "every CAD tool gives me totally bogus errors on this thing and I'm sick and tired of combing through dozens of logs files looking at thousands of errors and triple-checking that they are all bogus" tedious activity. I'd love to see it gone too and I mention that to my management every chance I can.

Patrick Mahoney
Intel Corp.
(not an Intel spokesperson, just a random engineer who works there).

 

[satish0612]

I do not remember my bios password, so i removed cmos battery for whole night but still asking for the bios password.
Please help me to locate the cmos jumper for my sony vaio(SVE15136) E-series laptop.

I already warned you once
for this. If I have to do it a third time,
you won't like the consequences.

We do not provide help for
ways to bypass passwords.
Contact the manufacturer of
your laptop for help.

AT Mod Usandthem

 

[Red Squirrel]

Check motherboard manual (online if you don't have it on hand), there may be a jumper for that. I suppose it's possible that it's stored in some kind of flash.

 

[William Gaatjes]

I think the main reason that bios settings are stored in volatile ram that is powered by a backup battery and not eeprom is quite simple.
Suppose that for some reason your bios settings get corrupted and your board will not boot anymore. The simplest and easiest solution is to take out the battery, short with a jumper any residual charge in the backup power circuit that keeps the volatile ram powered and your up and running again. It is proven and it works. I think that is the main reason.
Why ?
Well, if it is an eeprom where bios settings are stored, you are going to need a means to erase that eeprom if for some reason the system will no longer boot.
That is easy to solve by just designing a system that looks at the voltage level of a special pin.
When that pin is for example cleared, load the default factory programmed bios settings and write the default settings to the eeprom.
When that pin is a high level because of a pull up resistor to 3.3V, as a system, load up all data from the eeprom, do a crc checksum calculation and when the crc value is correct, load the bios settings into the chipset.

Then again, i am fairly certain that volatile ram also store a crc for the bios settings.
Must be a price thing.

It is certainly technological possible to replace the battery for a super capacitor as mention by Rakehelion to power an 32.768kHz oscillator and clock circuitry (Just dividers) for timekeeping.
But when the system is powered off for many days in a row, then even a super capacitor will not be able to keep the circuit running.
Although it is today possible to buy RTC clock circuits that use only for example 250nA or less when in power down but still time keeping. Power consumption low enough that even the leakage current of a super capacitor becomes important again. Lowest RTC time keeping consumption these can be around 50nA.

 

[Pink Jazz]

Isn't on modern systems the CMOS battery only used to power the RTC when power is completely shut off and would otherwise be powered by the +5VSB rail on the PC's power supply? That has been my understanding.