No POST after writing single bit to CMOS; any solution?

[Behemot]

I have encountered so many boards which storing a single bit in CMOS would render inoperable that I am starting to be crazy about that.

Last time it is my parents PC with DFI LANParty UT nF4 SLI-DR Expert. The symptoms are the same as usually: when you clear CMOS, it works fine. When you store a bit in there, it fails to POST. In real world that means you turn off or reboot the computer and as Windows and possibly every OS writes something in there, even though it would be time update, it does not POST after that. And certainly it is bad to use board in factory defaults with factory time, especially having to reset it manually each time.

This time there is only one difference: I know this board's CMOS has been written to every day at least twice: it has some strange like-a-backup function which saves your BIOS settings each time the board sucesfully POSTs, just before the boot process. And than the OS does the time update and maybe other writes. So that makes me think it is bad CMOS memory?

[SteveNielsen]

Re: No POST after writing single bit to CMOS; any solution?

The only thing I can think of that would normally update CMOS from the OS would be date and time settings. Other updates such as a change in hardware would update it but that is not from the OS booting, that's from the BIOS detecting new hardware.

I think you may be right about bad CMOS memory though, that's the only thing that I think of that could account for the behavior.

[stj]

Re: No POST after writing single bit to CMOS; any solution?

a bad diode could cause low voltage to the ram or false triggering of write-protect.

[Behemot]

Re: No POST after writing single bit to CMOS; any solution?

Which RAM do you mean? BIOS is stored in Flash (probably NOR) for some time. There is section which is usually denied for writing to unles you use programmer (this section is OK, I have flashed the latest BIOS version again with no change in behaviour). Than there is section called CMOS after CMOS RAM which has been used ages ago for this purpose. But it is still Flash, it weares out making the cells unusable.

[stj]

Re: No POST after writing single bit to CMOS; any solution?

cmos is not flash,
it's a couple of k of low-power ram usually stored inside the chipset with the RTC clock.
it's powered by the battery & psu via a pair of diodes.

i'v had to fix this a few times on industrial motherboards.

[Behemot]

Re: No POST after writing single bit to CMOS; any solution?

Chipset? Huh?

[eccerr0r]

Re: No POST after writing single bit to CMOS; any solution?

Some computers have the battery backed configuration RAM ("CMOS") inside the chipset silicon to save costs. Usually the RTC is also integrated. This is versus the external RTC/configuration RAM when things were less integrated. Sometimes it really is stored in EEPROM cells but that's fairly rare.

I'd suggest checking the configuration RAM battery if it has one, sometimes a bad battery could cause that behavior.

[Agent24]

Re: No POST after writing single bit to CMOS; any solution?

I would check battery, battery holder, and battery circuit also. I have many experiences of boards that do not POST with flat or dead batteries.

Sometimes you can get dirty battery holder contacts as well (Happened to me once) or even bad soldering on battery holder (happened to someone else).

Or the diodes etc can fail, of course.

[Behemot]

Re: No POST after writing single bit to CMOS; any solution?

Well, basically no battery would erase the settings going into default as well, wouldn't it?

[stj]

Re: No POST after writing single bit to CMOS; any solution?

no,
no, or low voltage to the sram would cause random values.

a simple test,
reset the cmos,
and SET THE CLOCK.

then power down including killing the 5v standby by pulling the plug for a day.
preferably pulling the psu plug from the mobo incase the psu has a large cap in it.

then boot up and check the clock - the clock section is usually slightly more sensitive than the cmos ram and will show up a problem sooner.

[Behemot]

Re: No POST after writing single bit to CMOS; any solution?

Man you noticed the When you store a bit in there, it fails to POST? How the hell am I supposed to set any clock? IT WON'T POST.

[stj]

Re: No POST after writing single bit to CMOS; any solution?

reset the cmos to start the test.
better yet,
follow the battery + track to the chip with your meter.
or get the datasheet for the chip and see what pin has the vbat supply on it.

[Agent24]

Re: No POST after writing single bit to CMOS; any solution?

Well, basically no battery would erase the settings going into default as well, wouldn't it?

Yes, but some boards don't seem to like it and go crazy when the battery is flat.

Put in a new battery and they POST straight away, then complain about a BIOS checksum error and the defaults having been loaded as per normal.

[Behemot]

Re: No POST after writing single bit to CMOS; any solution?

Strange because the battery itself should also play no role until the stand-by rail stops feeding board with 5 V. I'll have a look into it when I arrive with the PC to the shop.

[Agent24]

Re: No POST after writing single bit to CMOS; any solution?

Strange because the battery itself should also play no role until the stand-by rail stops feeding board with 5 V. I'll have a look into it when I arrive with the PC to the shop.

It doesn't make sense to me either, but I've experienced it first-hand so I believe it.

[stj]

Re: No POST after writing single bit to CMOS; any solution?

it makes sense if there is no diode to feed the standby-power to the sram.

[momaka]

Re: No POST after writing single bit to CMOS; any solution?

I'd say remove the battery altogether and try to power ON the board again. Have not seen a board that would refuse to boot due to no CMOS battery. However, I did have a board get corrupt CMOS settings due to a battery going slightly low. Took me a few minutes to figure how to change the menus from Chinese to English . Then I realized I could just pull the battery out and be done with that. And so I did.

Since I keep a lot of old computers around that I don't get to use more than a few times a year, I always pull their CMOS battery. If I'll be using the computer for only a few days, I don't bother putting in a CMOS battery.

[stj]

Re: No POST after writing single bit to CMOS; any solution?

Have not seen a board that would refuse to boot due to no CMOS battery.

i have - just the other day.
it may have been an Intel LittleFalls2 board.
(965 chipset)

[SteveNielsen]

Re: No POST after writing single bit to CMOS; any solution?

I have seen battery clips so tight they short out with the battery removed. I imagine that could stop booting by dragging the standby voltage down.

[stj]

Re: No POST after writing single bit to CMOS; any solution?

I have seen battery clips so tight they short out with the battery removed. I imagine that could stop booting by dragging the standby voltage down.

no, because there will be a diode inline - sometimes a resistor too.