Racking my brain - 3 960T's, mobo only likes to boot with 1

[I.M.O.G.]

So I've got a Crosshair V Formula here. Running bios 1102.

I have 2 960T's currently, and a 3rd 960T I had tested previously but since returned.

The 1st 960T seems to run fine in the board. Boots every time, no problems.

The 2nd 960T (as well as the 3rd) will barely ever boot. Once in a blue moon it will boot, but if I shut back down it won't start back up.

When it doesn't boot, power comes on, fans spin... But it never gets to the point where the DRAM LED on the mobo lights up. So its failing POST somewhere before that.

In the past 20 minutes, I have installed both the 1st 960T and the 2nd 960T, while changing nothing else, at least a half dozen times each. Every time, I can power on with the 1st 960T without issue, shut it down, and bring it back up - no problem. Sometimes, but rarely, it will start up fine with the 2nd 960T - when it does it runs fine, but it may never turn back on once I turn it off.

Any ideas welcome. Open to suggestions. It may be relevant - my motherboard is, visually, an atrocity. It runs pretty well, but it looks like crap. Covered in vaseline and some remnant art eraser. Some bugs in there too, though I mostly have tried to pick those out (this rig runs in the garage).

This is what I've done so far:

- cleaned the socket with a toothbrush, as best as I can
- melted off as much vaseline as possible with a hair dryer
- checked pins on both CPUs, none bent or missing

This is insanely frustrating, I don't understand how 1 CPU can boot fine every time, but 2 others I've tried OF THE SAME MODEL don't want to POST.

1 960T came from ebay, and the guy said it was working fine when he got it back - I told him I didn't know, but it didn't boot in my system when another 960T worked fine in it. I feel bad about that.

Now I've gotten another 960T, from a member here, and I know it is 100% working without any doubt... But my mobo has somehow become picky, and only wants to run with its favorite chip lately. WTH.

 

[dejo]

I had 3 920 I7's at one time and 1 chip would run in any of the boards and the other 2 chips would only boot in 1 of the boards. never figured out why, was rather strange

 

[I.M.O.G.]

Thanks Dejo, at least I know it isn't the only time anyone has seen something like this. Trying to find a needle that will fit in the socket holes, see if anything might be interfering with good contact in there.

 

[TeknoBug]

Just wondering, do each of the chips have different manufacturing stamps? Malaysia, Indonesia, Mexico, etc?

 

[txus.palacios]

If you want to thoroughly clean the socket, look for tabs around the PGA socket. Remove them and you'll get to the actual socket. Clean that with a toothbrush.

I found that the hard way. I was holding a motherboard, it fell, and I tried to prevent the fall by grabbing it... from the socket latch.

 

[trents]

Matt, take the CPU out and run the board through the dishwasher. I'm serious. Make sure the dishwasher is real clean first and then run the board through the dishwasher on heavy duty setting with heated dry but without any detergent. Then let it dry out real good. Set a household fan to blow air on it for a couple days. Then retry. In the meantime, spray some carb cleaner on the pins of the two CPUs that won't boot and let it dry thoroughly before you try to use either of the two CPUs.

 

[BeepBeep2]

Hey Matt
Try this BIOS before trying anything else and see if it helps...keep in mind this is an AM3 compatible bios, AM3+ CPUs won't work with it.

You'll note a few "special" behaviors with this BIOS, you may have issues with the board not rebooting after saving the BIOS, just save, exit and wait ~5 seconds before pressing RESET and it should come back fine. Outside of that I can't remember any others, I'll check my notes.

EDIT: Looks like just that and efficiency sux but you probably already know that from when I was over in August

Still got your Crosshair IV Extreme?

(DO NOT USE IN "Formula-Z")

 

[Hicksimus]

1/3 isn't that roughly how many 960T's can unlock 6 cores successfully? Are you sure the board isn't gunning for 6 core operation with them? I'm running a tiny screen atm, apologies if it has been mentioned.

 

[caddi daddi]

I was kind of leaning to bad locked cores being unlocked also.

 

[caddi daddi]

bios 1103 is an early fx bios or perhaps a sabertooth bios by accident, i would run the one before that myself.
thats sounds like it's not the problem anyway.

1102 is a sabertooth bios.

 

[I.M.O.G.]

1/3 isn't that roughly how many 960T's can unlock 6 cores successfully? Are you sure the board isn't gunning for 6 core operation with them? I'm running a tiny screen atm, apologies if it has been mentioned.

I was kind of leaning to bad locked cores being unlocked also.

Good guesses. I had thought the same, but asus unlocker is definitely disabled.

bios 1103 is an early fx bios or perhaps a sabertooth bios by accident, i would run the one before that myself.
thats sounds like it's not the problem anyway.

1102 is a sabertooth bios.

1102 may be a sabertooth bios as well, but its definitely a CVF bios as well. You can find it listed here:
http://www.asus.com/Motherboards/AMD_AM3Plus/Crosshair_V_Formula/#download

It's also the bios the current records are set with on the CVF with 960T's on hwbot.org. I usually use the same bioses other good scores have been put up with already, which is why I was on this one. I'm going to slap another bios on now just to see if it makes any difference in the behavior.

Matt, take the CPU out and run the board through the dishwasher. I'm serious. Make sure the dishwasher is real clean first and then run the board through the dishwasher on heavy duty setting with heated dry but without any detergent. Then let it dry out real good. Set a household fan to blow air on it for a couple days. Then retry. In the meantime, spray some carb cleaner on the pins of the two CPUs that won't boot and let it dry thoroughly before you try to use either of the two CPUs.

I'll try that. Kind of working up to it already. Yesterday I stuck the board in a box, with 2 holes - one hole for the hairdryer, the other hole for exhaust. This got the majority of the vaseline off it. There is a decent amount left on there, but I can clean the rest off by hand then give it the dishwasher treatment. Will report back once I get to this.

Hey Matt
Try this BIOS before trying anything else and see if it helps...keep in mind this is an AM3 compatible bios, AM3+ CPUs won't work with it.

You'll note a few "special" behaviors with this BIOS, you may have issues with the board not rebooting after saving the BIOS, just save, exit and wait ~5 seconds before pressing RESET and it should come back fine. Outside of that I can't remember any others, I'll check my notes.

EDIT: Looks like just that and efficiency sux but you probably already know that from when I was over in August

Still got your Crosshair IV Extreme?

(DO NOT USE IN "Formula-Z")

I'll try it.

 

[I.M.O.G.]

If you want to thoroughly clean the socket, look for tabs around the PGA socket. Remove them and you'll get to the actual socket. Clean that with a toothbrush.

I found that the hard way. I was holding a motherboard, it fell, and I tried to prevent the fall by grabbing it... from the socket latch.

Haven't heard of the socket coming apart, I'll see if I can get it apart without snapping anything off.

Just wondering, do each of the chips have different manufacturing stamps? Malaysia, Indonesia, Mexico, etc?

Chip markings that does not work:
HD96ZTWFK4DGR
CCBBE CB 1139FPD
9U92041J10105
Malaysia

Chip markings that do work:
HD96ZTWFK4DGR
CCBBE CB 1142EPM
2Y27123J10923
Malaysia

I'm not sure how to read markings on AMD chips, but just looking at the numbers I'm guessing they were both manufactured in 2011 - the working one in week 42, the non-working one in week 39.

 

[caddi daddi]

just to get it out of the way, why dont you try pre 0813, that was the first fx bios just as a check. I still think it's not the bios.

 

[I.M.O.G.]

Trying 0813 first, then 1703, then beep's 0051.

I don't think its the bios either, but only takes a moment to try anyways.

 

[I.M.O.G.]

0813 still worked fine with the working chip.

Three consecutive hard reboots have posted on the "non-working" chip using bios 0813...

Flashing 1703 now to see if that makes it not boot like it did with 1102.

 

[I.M.O.G.]

Both chips consistently booting fine with 1703 as well.

How do I flash bioses and swap chips that fast? Just toss 'em in the air and roundhouse them into the socket.

Going back to 1102 to see if the buggy no boots come back...

 

[I.M.O.G.]

Same problem back on 1102. Only one of the two CPUs works.

 

[I.M.O.G.]

So ya, this is reproducible. There is some kind of bug in bios 1102 which makes it not work with certain 960Ts on the cvf.

Both the other BIOSes make both chips work fine. Kinda crazy it seems like. But was just a buggy bios.

 

[Robert17]

Seems like a BIOS issue from reading the thread. I found some CPU ID info here:

http://www.cpu-world.com/step_codes/C/CCBBE CB.html

that may also make you scratch your noggin a bit more.

 

[BeepBeep2]

So ya, this is reproducible. There is some kind of bug in bios 1102 which makes it not work with certain 960Ts on the cvf.

Both the other BIOSes make both chips work fine. Kinda crazy it seems like. But was just a buggy bios.

What about 0051? The 1xxx bioses are all tuned for BD, so is anything after 0705* I believe.

EDIT: Whoops, saw you said "both"