evga 680i cold boot errors

[Windfire]

Symptoms:
I get various errors on cold boot. If it boots correctly all software runs stable and I’ve had no lockups or other crashes when the system boots fine.

On cold boots I can get one of these three:

1 – C1 LED code
Long beep repeats forever until I turn it off or hit reset button

2 – 41 LED code
The following message is displayed:

BIOS ROM checksum error
Keyboard error or no keyboard
Detecting Floppy drive detected…

3 – No LED code, but post screen displays:
WARNING! Now System in Safe Mode
Please Reset CPU or Memory Frequency In the CMOS Setup

Press F1 to continue or DEL to enter Setup

Note: If I enter Setup and then F10 to save and reboot all is FINE.

Note: If I hit F1 I’ve found if I run Nvidia Control Panel the PC turns off(!).


Specs:
eVGA 680i motherboard (122-CK-NF68-A1) (flashed to P28)

CPU: Intel e6600 w/ Thermalright Ultra 120 w/ 74cfm fan

Memory: Corsair TWIN2X2048-6400C4 (2 sticks) running at 4/4/4/12 @2.0 volts

I’ve got the eVGA 8800GTS 320 Superclock (stock speeds)


I have DISABLED the following CMOS settings (standard stuff)
CPU Thermal control
C1E Enhanced holt state
Intel Speedstep
CPU Spread Spectrum
HT Spread Spectrum
PCIe Spread Spectrum SPP
PCIe Spread Spectrum MCP
Sata Spread Spectrum

I’ve also de-coupled memory from the FSB (no overclocking of memory).

I’ve got a mild CPU overclock: 2900MHz.
CPU Voltage: 1.281v
FSB: 1.30v
HT nforce SPP <-> MCP: 1.2v (stock)
nForce SPP: 1.25v
nForce MCP: 1.5v (stock)
Auxiliary: 1.5v (stock)

At 70F ambient temperature I get the following temps using TAT to measure:
Idle: ~28/26c
Orthos: ~41/39c
TAT:: ~45/43c
With most normal games I get around ~41-42c at peak.

In other words, it runs quite cool.


What is going on? Any advice? This is getting kind of irritating.

 

[Skrying]

Run it at stock for awhile, maybe your motherboard just doesn't like the memory and FSB not being 1:1. Hard to say, 680i boards weird issues like this...

 

[Windfire]

Run it at stock for awhile, maybe your motherboard just doesn't like the memory and FSB not being 1:1. Hard to say, 680i boards weird issues like this...

I've removed the OC and dropped voltages to [Auto]--I left everything else the same (disabled features, etc.).

This does not appear to have helped. I've tried cold boots 4 times and all 4 times I got the error about running in Safe mode and am forced to go the BIOS to hit F10 and save to boot correctly.

Any other ideas?

 

[Windfire]

Okay, I've actually went into the BIOS and reset to DEFAULT settings.

Same issue: 1 long beep and the post tells me I'm in "Safe Mode" and that I should reset the CPU or memory frequencies. I have to go into the CMOS, hit F10 and save the settings and reboot--which it does fine. This happens every time now.

I'm thinking the 680i is not happy with my Corsair 6400 C4 memory. That is hard to believe in my opinion--really stupid.

Ideas?

Thanks,
Steve

 

[stevem]

A couple of things to try:
1. Check PSU voltages
2. Unplug USB devices
3. Flash latest MB BIOS (if you haven't)
4. Try single channel RAM/other RAM modules
5. Try another graphics card
6. Install only CPU, single RAM, low end graphics. Disconnect all I/O, HDD, etc. If stable, then add second RAM & try overclocking. Gradually build the system back up.

 

[Malo]

That Corsair memory needs to be run at 2.1 volts.

 

[Kurt]

Check if the CMOS battery is still good/seated correctly.

 

[SugarCoat]

Not to be a jerk windfire but i and a couple others told you so.

what happened to your sensible conclusion? Decided to roll the dice?

http://forum.beyond3d.com/showpost.php?p=910404&postcount=43

 

[Windfire]

That Corsair memory needs to be run at 2.1 volts.

I've tried 2.1v with no change. It behaves the same with 1.8, 2.0 and 2.1.

To be clear, once I've booted into Windows I'm perfectly stable. I've not experienced a single crash or lockup in games (HL2, Oblivion, UT2004).

 

[Windfire]

UPDATE #1: Being ignorant of the true voltage of an e6600 I thought 1.26v was stock. Therefore, I thought I was being generous with 1.28v. It turns out that 1.32v is more "normal". As such, I was under-powering the CPU and OCing 500MHz as well. This was probably part of my problem (that was feedback from eVGA support.

UPDATE #2: In addition to the undervoltage, I figured it had something to do with my RAM so I did some experiments.

First, I removed the second Corsair stick. It booted fine (problem gone).

Second, I replaced the first stick with the second stick. It booted fine again.

Third, I placed the first stick in the second slot (both in black slots)... It worked FINE!


My question: Is it more likely the RAM was not connected "perfectly" before or could it be that the motherboard was picky enough to only want the RAM in this specific order???

I'm going to do some RAM tests to make sure one of the sticks isn't bad.

Has anyone heard of this before?

 

[SugarCoat]

normally on intel chipset boards it doesnt matter what order the memory is in. Dual channel or not it should boot. It sounds like you're now in dual channel since you have both sticks in corresponding black DIMM slots. The only time i've had something like that happen is with AMD systems, some of which need to either be in dual channel, single stick, or nothing at all because it wont boot as it wont like the odd stick. That very well may be a limitation of the 680i board and obviously not a big deal, it should also say in your manual which you should be reading (RTFM) .

All the same its problably worthwhile running memtest86 for at least 10-15 minutes to see if you get a whole bunch of errors.

Make a note of this as well for the future, many times if a computer wont POST it has to do with the memory configuration so one of the first things you want to do is try 1 stick in DIMM 1 and see what happens.

 

[Skrying]

I'd suggest doing two runs of memtest86, or just let it run as many tests as possible over night. I say this because I've had memtest86 not show errors even after one run (though never after two if the first runs didn't have them) and indeed it was the memory causing the issue. Just a safety precaution, I'd say your issue is fixed and if you feel good about it not to worry.

 

[Geeforcer]

UPDATE #1: Being ignorant of the true voltage of an e6600 I thought 1.26v was stock. Therefore, I thought I was being generous with 1.28v. It turns out that 1.32v is more "normal". As such, I was under-powering the CPU and OCing 500MHz as well. This was probably part of my problem (that was feedback from eVGA support.

UPDATE #2: In addition to the undervoltage, I figured it had something to do with my RAM so I did some experiments.

First, I removed the second Corsair stick. It booted fine (problem gone).

Second, I replaced the first stick with the second stick. It booted fine again.

Third, I placed the first stick in the second slot (both in black slots)... It worked FINE!


My question: Is it more likely the RAM was not connected "perfectly" before or could it be that the motherboard was picky enough to only want the RAM in this specific order???

I'm going to do some RAM tests to make sure one of the sticks isn't bad.

Has anyone heard of this before?

Have you tried moving sticks around while still undervolting? Chances are, undervolting was the entirrity of your problem. Also, as Melo suggsted, did you set your memory voltage to 2.1? That's what Corsair says you need to be running yours (and mine, got the same memory) DIMMs for 4/4/4/12.

 

[Geeforcer]

Not to be a jerk windfire but i and a couple others told you so.

what happened to your sensible conclusion? Decided to roll the dice?

http://forum.beyond3d.com/showpost.php?p=910404&postcount=43

Not to be a jerk, but if it turns out to be human error, will I get to ITYS right back?

http://forum.beyond3d.com/showpost.php?p=899573&postcount=16

 

[Windfire]

Thanks for all of your input. I appreciate it.

For now it is working fine. I'll update status if that changes.

As I stated, with 1 DIMM it works fine (either DIMM).

To be clear, and this is what makes the situation really "odd" is that I was getting the C1 memory error and "Safety mode" with the two DIMMS in the same black slots but with the DIMMS reversed.

Example: If I label DIMM #1 as "A" and DIMM #2 as "B" I get the errors if I install them A + B. It works fine if I install them B + A.

Strange eh?

It makes me wonder. We'll see what happens over the next few weeks. If errors resume then I know my motherboard is in trouble or the memory is not working correctly with the motherboard.

 

[Windfire]

Have you tried moving sticks around while still undervolting? Chances are, undervolting was the entirrity of your problem. Also, as Melo suggsted, did you set your memory voltage to 2.1? That's what Corsair says you need to be running yours (and mine, got the same memory) DIMMs for 4/4/4/12.

Yeah, I tried 2.1 with the memory--no difference in the scenario where I was getting the errors.

On that same token, when I boot fine, the voltage works fine at 2.0 too.

I'll bump it to 2.1 to follow the book.


P.S.: I'm wondering if I over-spent on my memory ($130 after rebate) considering that OCing memory doesn't seem to be very helpful. Have you tried OCing the C4?

 

[Windfire]

Have you tried moving sticks around while still undervolting? Chances are, undervolting was the entirrity of your problem.

I'm really hoping that was the case.

However, what doesn't make sense is:

A - It only gave me the error on cold boot

B - The very next warm boot from CMOS (with exact same settings) worked fine

C - I experienced no stability issues in Windows or Games


I would think a true undervolting would cause general instability.

This makes me think that a routine in the CMOS is not working properly for cold boots (something isn't initializing correctly, etc.) where it gets repaired when you go into CMOS (whether you save or not).

Just a thought.

 

[SugarCoat]

Not to be a jerk, but if it turns out to be human error, will I get to ITYS right back?

http://forum.beyond3d.com/showpost.php?p=899573&postcount=16

Whats ITYS?

If its human error thats all well and good but 680i have no shortage of teething issues, and its not all or even a majority human error, they've certainly had far more exotic problems then decent boards based off 965/975 chipsets and the boards themselves are infact to blame. This is evident in the multiple bios releases aimed at fixing multiple severe user problems. What windfire is having happen is actually the best type of problem you can have, the in your face ones that happen right away. What i worry about is some nice hard disk corruption or stability issues a few months down the road which is exactly what both Nforce4 boards i used did to me, they were quite literally terrible quality and nothing i would of recommended to anyone who really wanted a stable computer for any decent length of time. At the release of the 680i when he was inquiring about what board to go with and even now i question how good of a chipset it really is. Especially at ~$200-$250 a pop which is on the high end for a motherboard.

 

[Geeforcer]

ITYS = I Told You So.

Purely from personal experience, right now, 680i is a very fast, stable and robust platform. And yes, it had "teething problems" and then some (RAID controller disconnect) . That aside, it has been a better board than both 965 and 975 board from several manufactures that I have worked. Personal experience, your manage may very, etc.