What is the deal with 'long CMOS clear'?

[Guest Kobalt]

The CAPs on DFI boards are high quality, but, they lose their charge as fast as most CAPs of their size do.

 

So what is the logic behind the 'long CMOS clear'? I can see maybe 1-3 mins, but no way it can keep a charge for much longer than that, and no way can it hold a charge for more than 1 hour. These CAPs just aren't rated for that, and they would be quintuple (or more) in size to hold a charge for that long.

 

Did this come about as a result of people not unplugging the PSU, and not removing the battery & not pressing the case power switch (to discharge what little power was left in the CAPs) when doing a CMOS clear?

[red930]

Did this come about as a result of people not unplugging the PSU, and not removing the battery & not pressing the case power switch (to discharge what little power was left in the CAPs) when doing a CMOS clear?

 

I actually wonder about this too, I've never needed more than 10-20 minutes, and typically I only go for 5 max. I'll go longer if I need to shower or something.

 

Dual-tasking baby!

 

But yeah, I'm guessing it's about the 'unplugging the power supply' step that many people don't use.

 

But then again, ExRoadie and others say it can fix things as a last resort, so maybe I've just never come across a situation like that.

[red930]

So what is the logic behind the 'long CMOS clear'?

IT WORKS!

 

It works when a short Clear CMOS doesn't!

 

Now that the short answer is out of the way here comes a detailed explanation.

 

As best we can tell, this Long Clear CMOS works on just about every motherboard since they started using 2Mb chips. It's especially effective on boards that have secondary NVRAM to store data during a Clear CMOS such as the CMOS Reloaded on DFI boards. It takes some DFI users a while to realize that the CMOS Reloaded data doesn't get erased when they Clear CMOS.

 

There are hundreds of posts in this forum alone that attest to the viability of this procedure. Over the past couple of years it's migrated across other support forums and is even included in several vendor's "last resort" trouble shooting guides for top level support. I know because I deal with them almost daily.

 

I've had many dozens of boards that were considered "dead" by other technicians, revived by the Long Clear CMOS. Just a few weeks ago I was able to revive an Intel Server board that arrived DOA according to the software vendor. After setting the CMOS to clear overnight the rig booted right up in the morning and allowed the vendor to get back on schedule. It saved them tens of thousands of dollars to replace the entire rig on short notice since the customer didn't want a "box with problems" for their money.

 

So next time a short Clear CMOS doesn't work, give it a long Clear CMOS and see what happens.

[red930]

ExRoadie beat me to it but might I might I add to the his explanation.

 

With the advent of:

 

- 2Mb and 4Mb (that is Megabit not MegaByte) I/Cs

- longer and diverse trace routes

- denser I/Cs incorporated into the design of various 'newer' (last 3 years) models of motherboards.

- and taking into consideration the varied hardware configurations that may have a affect on dissapation...

 

...capacitance along the circuits has increased.

 

With this realized extra capacitance a new standard of dissipation has evolved and is needed to completely clear the entire circuit of excess energy. Henceforth, a longer CMOS clear-time.

[red930]

2MB and 4MB (that is MegaBit not Megabyte) I/Cs

Yeah, that was a typo that I corrected. My little b got to be a big B when the pinky finger hung on the shift key.

 

Should have read 2Mb!

[Smoken Joe]

I have wondered about this myself but did not see the reason to argue with sucess. It works and that is enough. That has ben the rule like ExR stated sence these boards started.

 

I find it a constant source or amusment when people refer to things as the "DFI street way" like this tecknique when it is simply time honered PC and general computer diagnostic tecnique out of the older PC troubleshooting guides. It is really a complement to the people on this forum that give help.

[Guest Kobalt]

Very intresting.

 

I was looking at the docs for the EEPROM, and according to that, they list :

TPRST VDD stable to Reset Low 1 ms

TKRST Clock Stable to Reset Low 100 μs

TRSTP RST# Pulse Width 100 ns

TRSTF RST# Low to Output Float 48 ns

TRST 1 RST# High to LFRAME# Low 1 μs

 

So in a um.. 'perfect' setup, that is the time needed to reset.

 

It would be more intresting if someone with a 'dead' BIOS had a EEPROM reader, and do a dump of it. Then have them do a 'long clear' and see what changes.

 

I guess another alternative would be to try a new BIOS chip.

 

Well, anyway, thanks for the info.

[twigert]

heh...i learned another new thing today! Now i understand how the long-clear works...

[dr_bowtie]

In case of the NF2 days...

 

It just seems that even though a quick cmos clear should do it...In a DFI sometimes it doesnt...It just seems the board is more picky about holding on to some (but not all) info...

 

Sometimes if you clear the cmos and it wont boot...changing the bios chip will not help at all...until you do the long cmos clear...

 

Everyone knows the DFI-NF2's were easily corruptable and if you dont follow the long cmos clear that when changing the bios chip (or flashing a new) the coruption would go to the new chip also (or flash)

 

I have had 1 board that wouldnt boot after even a 18 hour clear...I left it down for over a week and then out of the blue on day...it booted...so it works...

 

I think it has more to do with the layout and how the cmos works...On other brand boards if you move the jumper to clear and hit the power button....nothing happens...On a NF2 rig if you do that...it will power up...although it wont post...so there is a difference...

[ReelFiles]

Had some old intel boards shelfed up for a couple of years, I swore the bios was corrupted on both of them, well eventually the batteries died. I threw one in a case the other day, added a new battery and voila, there she was running smooth (as smooth as a pIII could be). So it definitely works. Nice to know the reason why, thanks guys.

[old_geekster]

All I know is, "it works"!

 

I was having a "cold boot" problem when I first OC'd my system. I tried clearing the CMOS for several different time periods. None of them worked. So, I cleared it for close to 24 hours and walla, it worked.

 

I have not had another "cold boot" problem since.

[red930]

I've always clear my cmos 5-10 seconds the longest, and i think the long clear cmos is a conspiracy