Memory Overclocking - Did I break something?

[Candor]

So I was attempting to overclock my DDR4 2400 memory to 2800mhz or better as I'd seen other having success with this. I failed.

The highest speed I achieved was 2666mhz. I did at one point increase the memory voltage to 1.4v briefly but could see I wasn't going to achieve anything higher than 2666mhz and gave up, returning the voltage to 1.2v.

This is where the weirdness begins. My system will no longer boot into Windows 10 with my memory set at 2400mhz. It will only boot with the memory set to 2600 or 2666mhz (voltage on Auto/Default).

I've never come across anything like this with memory before. Although I'm obviously happy to run the memory at 2600mhz... Did I break something?

I have re-flashed the motherboard bios but this changed nothing. Any thoughts?

Gigabyte GA-X99-SLI --- I7-5820K --- Kingston HyperX Fury 2400mhz (2x 8GB dimms).

 

[hat]

Are you sure there's no setting somewhere you may have forgot? Try clearing the CMOS, not by BIOS flashing, but by removing the battery and shorting the cmos clear pins.

I saw you said you flashed the BIOS, but I'm not sure if this would have returned everything to the default settings or not. Clearing CMOS will.

Tip for the future: BIOS flashing can be kinda risky. Usually nothing happens, but if the power cuts out or something while flashing the BIOS you're screwed. Do it if it needs done for an update or something, but it's not necessary just to return to default settings.

 

[Candor]

Yeah I've been flashing things for years and I'm aware of the risks. I initially did the clr_cmos pin and when that didn't resolve the problem I then did a bios flash.

It's just such a strange issue. Like I said I'm happy to run the memory at 2600mhz, I was just trying to figure out if I have damaged the motherboard in some way. It's the only thing I can think of.

I usually sell my previous hardware after a new system is built and I'd hate to sell someone this motherboard in the future while being unsure if there's a fault.

If I have a doubt about it I'll probably have to hang on to it.

Thanks for the reply .

 

[hat]

Hard to believe something has been damaged by running 1.4v for such a short time... and signs of damage would probably be more obvious than refusing to work at standard speeds, only faster speeds. Maybe setting 2600MHz while on auto voltage is raising the voltage a bit, allowing it to work? Maybe you need to enable XMP?

 

[Candor]

Yeah I tried all that too. All I know is the memory worked fine previously at 2400mhz until I attempted the overclocking. Voltage is definitely a suspect. Maybe it's now going low at 2400mhz.

I might try a higher than the normal 1.2v at 2400mhz and see if that does anything.

Ok well using the xmp profile worked. But using manual settings for 2400mhz at 1.2v (default) causes boot failure. Increased the voltage to 1.25v and it worked.

Thanks for your help in narrowing this down. Voltage is the issue. It worked fine on 1.2v-Auto before but now it doesn't. I can live with that, but it's still WEIRD

 

[Hood]

You could try running memory diagnostics, using MEMTEST86, although it doesn't show you the RAM speed, just the total amount and overall bandwidth. That's definitely a strange problem. I have had BIOS settings that wouldn't clear after clearing CMOS, and a buggy BIOS or 2 that wouldn't let you change settings, until you set to optimized defaults, or clear CMOS. Maybe try a long time without CMOS battery (30 min.) - some people have solved problems that way, who knows why. If another hour or 2 of fooling with it doesn't help, I'd say you're running at 2600 from now on.

 

[Candor]

I've just discovered that one dimm slot now doesn't like to have a stick in it - causes a shutdown at post. So I now also suspect that I have damaged a dimm slot after plugging memory in and out for an hour last night when trying to diagnose this problem. Sigh. Happy days, lol.

I once had a computer that wouldn't start when you pressed the power button. The only way it would start is when I blew with your mouth on the motherboard and it would start without touching it. So yeah I've had some strange hardware in my time. But that was 15 years or so ago.

Thanks again for your help.

 

[hat]

I've just discovered that one dimm slot now doesn't like to have a stick in it - causes a shutdown at post. So I now also suspect that I have damaged a dimm slot after plugging memory in and out for an hour last night when trying to diagnose this problem. Sigh. Happy days, lol.

I once had a computer that wouldn't start when you pressed the power button. The only way it would start is when I blew with your mouth on the motherboard and it would start without touching it. So yeah I've had some strange hardware in my time. But that was 15 years or so ago.

Thanks again for your help.

Ouch... I've done something similar. It was Christmas time, and I was young... and I had just got a relatively fast Nvidia 6800XT, along with an Antec SmartPower 2.0 500w power supply, because the one I had wouldn't handle a card like that. It made some kind of high frequency squealing noise... the issue is well known now, but not so much then. Nobody believed me about the noise though, and much like you, I was repeatedly trying my old card, then the new one, then the old one again... because the noise seemed to not be present when using the old card. Ruined my motherboard at the time (or at least the AGP slot), one of the traces was fried...

The next thing I got was a Corsair power supply.

 

[Candor]

Ah good times. I don't know if I'll ever learn to leave things well alone. I'm a tinkerer by nature and can be a bit ocd about getting my hardware just the way I want it.

Like the extra 1 frame per second I might have gained by overclocking my memory, lol.

 

[hat]

Actually memory speed can play a significant part in gaming performance. Maybe not so much from 2400 to 2666, but... slow RAM can hold you back.

What happens if you don't use that bad memory slot? Can you run at 2400 then?

 

[Candor]

Yes I can run 2400mhz on other slots as long as I manually set the voltage to 1.25v or use xmp setting.

Something has definitely changed. I might have to disassemble the system at some point to have a really good look at the slots.

 

[Mussels]

Doesnt that gen intel suffer the problem where CPU mounting in the socket, or a too tight cooler can affect/cut out RAM slots?

Maybe go as far as taking the CPU out, wiping the pads down, blowing dust out of the CPU socket and DIMM slots, and trying again?

I got my 4770k basically free because a thumbprint on the CPU caused 3 of the 4 memory slots to not work

 

[Candor]

Thanks heaps for that info - I hadn't heard about that. Definitely worth a look.

 

[Caring1]

VDDSPD=2.2v to 3.6v
so putting too much through them may have caused an issue, just like too much V's through a CPU prematurely wears them out and they require more volts to run at a stable speed.
But I doubt it.
Check the Ram slots for a bent finger.

 

[Candor]

Ok I've just done a complete tear-down and reassembly. Nothing has changed and I could see no visible damage to the dimm slots.

At this point I must conclude that I have caused damage to something on a hardware/electrical level that I can't see or verify. Considering that I have caused this issue I'll have to suck it up .

I can work around the problem memory slot and the system is working fine.

It just means I won't sell this motherboard on to anyone in the future, given it's compromised state.

Thanks to everyone for their input. It is much appreciated.