Misbehaving BIOS?

Associate
Joined
2 Nov 2006
Posts
141
Location
Kent, UK
I am a bit stumped. I have an ASUS P6X58D-E mobo, an i7 980X CPU, nVidia 460 GTX GPU and 12gb RAM.

Now the problem I have is every once in a while the machine will not boot up, all the noises are heard and all the lights come on. I know when it will work since I hear a distinctive beep which means it sees the GPU and boots, if I don't hear this I know I am in trouble.

Usually the way I have found to fix the machine is unplug the monitors and the two SATA drives. Then switch it on then off again. Then I plug everything back in and it tells me it cannot see a bootable drive and I have to go into BIOS and select the drive with Windows on it since it seems to change all drive priority settings.

Does this seem like faulty BIOS or ropey mobo or both? I have the latest BIOS installed I am puzzled why it does this every once in a while nothing seems to aggravate it.

This machine was an Overclocker custom made machine which was originally set to be overclocked and I had problems with my £D program because of this. I am wondering if the overclocking has cause this booting anomally.

Anyone got any ideas?
 
wazza300

Yes it does it from a cold boot. Prime example today, I was using the computer all morning went out after switching it off and then when I cam back an hour or so later it wouldn't boot. Got a 600watt PSU in it I would have thought that would be enough?
 
you can either add more voltage to imc and or dram voltage or test with just 6gb of ram and see if it still does it from cold

How do I add more voltage to imc/dram and what is imc? I must admit that I never had this problem until I went from 6gb RAM to 12gb.
 
Well I quickly read up on these settings from the manual and checked the OV_CPU and OV_DRAM_BUS jumpers, both where set to normal which was a surprise since it mentions that these can cause the machine not to boot.

I haven't yet looked at the BIOS settings but will do today and post a picture but I think that the QPI/DRAM Core Voltages are set to default since I reset the BIOS when I first encountered these problems.

The RAM is 6x2gb Corsair DDR3 XMS3. When I looked this up using CPU-Z I saw Voltages from 1.5 to 1.65 in the Timing Table does this help?

BTW this machine is not overclocked anymore since it caused my 3D app to crash.
 
Here is a grab of my BIOS under AI tweaker section, as I thought the settings talked about are at their default values. I must admit I am baffled that you have to change voltage setting when installing more RAM. I have had and used many computers and never had to do anything like this when upgrading RAM…

BIOS%20Settings.jpg
 
you need to set qpi/dram voltage to atleast 1.35v,maybe more

last time I ran 12gb at 1600mhz I needed 1.3750v qpi/dram

then set dram voltage o whats printed on the ram either 1.55v or 1.65v

x58 is funny when it comes to memory and auto doesn't always set the right voltage,modern sandy/ivy/haswell rely on a fixed bclk so it much more stable for memory

Cheers wazza300, I will check this out :-)
 
It's flipped out again but I didn't change the voltages. I took out 6gb RAM and it still didn't boot... Then I took all the RAM and it didn't squeak like it should do when no RAM present. So I put all the RAM back booted pushed the DVD eject button as it booted and it was fine???

If I up these voltages I won't damage my computer will I? I also noticed that it said that the over locking setting aren't working and I either need to set the BIOS manually or use defaults??? Something isn't right...
 
Just been looking through the manual and the QPI/DRAM Core Voltage setting seems straightforward whereas it states that setting the DRAM bus voltage to more than 1.65v may damage the CPU. My memory is 1600 1.65v so now I am nervous about setting that value... If it kills my puter I won't be happy right in the middle of a big project...
 
Well I have taken the plunge and set QPI/DRAM Core Voltage to 1.35 and DRAM Bus Voltage to 1.65 although it went up in .02V increments it went from 1.64 to 1.65 with the message 'According to Intel CPU SPEC, DIMMS with voltage setting over 1.65V may damage CPU permanently!!'

Now I am worried, I only just bought this CPU and it's great, don't want to shag my system just yet…

Actually, having a little look on the web I don't think I should have Corsair XMS 1600 1.65v memory in my computer anyway, I think that was installed to allow for overclocking which caused it to crash my 3D app and cause my boot problems. I can buy some 1.5v memory which is below that safety limit of 1.65v so I might get 12gb RAM in 3 x 4gb DIMMS. That sounds better :-)

12gb kit Ballistix


What do you reckon peeps?
 
Last edited:
Thanks Wazza, it tripped out on me again this morning before I committed to the changes you recommended so I took three sticks of memory out and it booted OK with 6gb RAM although I had to reset the boot order again.

Just out of curiosity I snapped these from CPU-Z, it is showing the Voltage as 1.5V??? and the XMP-1600 as 1.65V I don't understand the last bit - I am more of an animator than techy :-)

CPU-Z_Mem_Readings.jpg
 
IMC = Internal Memory Controller?

So what you are basically saying is that to run the memory at 1600MHz you need a voltage supply to allow this i.e. 1.65v?

the 980x has a better imc so you might be able to lower dram/vtt later on and uncore freq is twice your memory multi or maybe one or two clicks less as its a 980x cpu

so 8x ram multi = 16x uncore freq,or maybe 15x or 14x whatever is the most stable

Sorry, this bit is over my head…

So why are Crucial selling memory which needs only 1.5v and runs at 1600mhz? Whenever I used to upgrade other machines I never had any of these problems - REMEMBER - I am not overclocking this machine.
 
OK, I took out 6gb of RAM without touching the voltages and the machine played nicely (plucking up the courage to up the voltages as recommended). Kinda figured that it worked all the time with 6gb, so left it at that for a while.

Then it done it's usual after two weeks and suddenly wouldn't boot. So I took all the RAM out and booted waiting for the sound of no RAM present (one continuous beep followed by two short beeps) but it didn't happen. Tried several times but my logic board thinks it still has RAM in it - huh?

In the past I found that once it beeped it was OK to put the RAM back and then it would work.

Is it shagged?

Actually I noticed a button on the mobo which is called MEMOK! It says if it is on all the time the memory is incompatible which it is when it doesn't boot. Is it worth testing with this button or a waste of time? I notice that it switches off on successful boot.
 
Last edited:
Yeah, but why doe it not beep when no memory installed at all? I was going to get a SSD drive for this machine but now I thinking replacement mobo it seems a bit dodgey.
 
Yes... Once it beeped I put the memory in and it behaved…

Surely it shouldn't do this? Does it sound like a dodgey mobo? I am also thinking that when I first bought it in April 2011 it was over-clocked perhaps with these memory settings you suggest which crashed my 3D app, perhaps if I upped the voltages it would start causing the same problems…

Gonna check with Overclockers to see if they touched the RAM and up the voltages anyway, if it dies I guess I have to buy an i7 3970 bundle :-)
 
Last edited:
Checked with Overclockers and they didn't touch the RAM so I am still wondering why the machine doesn't 'BEEP' when no memory is present in which case have I got a dodgey mobo?
 
OK. Machine wouldn't boot even with 6gb RAM and no voltage tweaks. Turned machine completely off then back on again - it booted. Set the voltages as recommended and rebooted with 6gb RAM it was slow but made it to Windows. Turned the machine off and installed all six RAM modules - won't boot at all.

Does this sound like a shagged motherboard? I am getting so ****ed off with this problem :-(
 
It's still under warranty, but I don't know what that they will do they don't make these boards anymore.

Thanks for all your help wazza - I appreciate it :-)

Still wondering if it is those chips being 1.65v - the printed voltage on the mem sockets says 1.5…
 
Last edited:
Well I'll go to the foot of my stairs. I put all the RAM back in and left the machine for two hours. Came back and it booted up fine...

Will somebody tell me what is wrong with my system?

Seems like the voltage thing doesn't make a difference it still has boot problems. I think I will opt for getting 1.5v RAM 3x4 and see if that behaves. It doesn't matter since it will mean I can upgrade the computer to 24gb later rather than the 12gb I have now.
 
OK, spoke to any engineer and when I told him the symptoms he said that the board is cracked (although you cannot see it) and thus the memory is not being detected.

Does this sound like the fault
 
Back
Top Bottom