Upgraded CPU, now system locks up

[Tech Tweaker]

So, I purchased a faster CPU (Intel E6550 (SLA9X)) for the 775 rig in my sig earlier this week to replace the old E4400 (SLA3F).

Just arrived today, took the old one out, put the new one in, re-TIM'd the HS and proceeded to lock it down.

Went into bios and set the FSB manually to 333 just to make sure it was set right, then loaded up XP. Now at this point everything seems to be going fine, for about two or three minutes, maybe less, at which point the system completely locks up/freezes and I have to reset and let it cool down before it will attempt to boot again.

So after fighting with it for a half hour or more I start grabbing at straws and feeling around in the case to see if anything is overheating perhaps, I feel the HSF for the CPU and it feels alright (a little hotter than with the E4400, but not bad), then I check the NB HS which is one of those Zalman ZM-NBF47 passively-cooled Aluminum heatsinks (Houston, we have a problem), I swear if it were flat I could probably fry an egg on it.

With the slower processor it would barely even be warm to the touch, and took a long time to warm up, with the new one though it is HOT, and warms up very quickly.

I'm thinking about going with one of these northbridge heatsinks, I already own one of them which I've found to work great, but it is unfortunately mounted to another board.

Anyone have any other thoughts or suggestions?

 

[EarthDog]

Bios update to handle that CPU? Its faster FSB and all but... I wouldnt expect it to be that much different temp wise.

 

[Tech Tweaker]

Bios update to handle that CPU? Its faster FSB and all but... I wouldnt expect it to be that much different temp wise.

Already has a new enough Bios version to support this CPU. This one was supported since version F3 and I'm running F7.

Yeah, I wouldn't think so either, but there seems to be a considerable difference. I can't say for certain the exact temps of the NB, but with the E4400 the CPU idles at 28-33°C, while the E6550 idles at 45-47°C with a stock Intel HSF. That was the one thing I managed to check before the system overheated and crashed.

 

[OnDborder]

Are you sure one of the push pin retainers on the stock hs aren't bent back?
You know the kind of clear plastic hook things that insert into the mb mount hole.

 

[Tech Tweaker]

Are you sure one of the push pin retainers on the stock hs aren't bent back?
You know the kind of clear plastic hook things that insert into the mb mount hole.

Just checked and they all look normal from the backside of the motherboard with all of them going through the mounting holes properly.

 

[EarthDog]

Yeah, sounds like a bad mount to me... it shouldnt be that toasty...

 

[Tech Tweaker]

Yeah, sounds like a bad mount to me... it shouldnt be that toasty...

Four to five times in a row?

I seriously remounted it four or five times, just in case that was the problem. So, I highly doubt that I could be that bad at mounting heatsinks. Every time all of the pins were fully engaged/locked in.

 

[RJARRRPCGP]

Looks like you need to take out the board and inspect it for bad mounting and dirt.

I really doubt it's BIOS-related. It's not a 45 nm.

 

[EarthDog]

If you push down firmly on the hs/f do the temps drop? WHen you touch the hs/f when the CPU is that warm at idle after a while, is it at least warm to the touch?

 

[Tech Tweaker]

If you push down firmly on the hs/f do the temps drop? When you touch the hs/f when the CPU is that warm at idle after a while, is it at least warm to the touch?

Nope, no change.

Quite warm in fact.

 

[davedree]

was the processor 2nd hand? and what actual temps are you seeing in the bios?

 

[Tech Tweaker]

was the processor 2nd hand? and what actual temps are you seeing in the bios?

Of course, I buy most of my stuff used.

Don't know, didn't check.

 

[Tech Tweaker]

Well, I have rectified my issues it seems.

As it turns out, you can't use a HSF meant for a C2D E4x00 series processor (200MHz FSB/800MHz effective bus speed) with a C2D E6550 processor (333MHz FSB/1333MHz effective bus speed) while using a passive NB cooler and expect it all to work properly.

I traded up to a Zalman CNPS8700 NT cooler and an actively cooled Vantec CCB-A1C NB heatsink and now all is working fine with the E6550.

Initially I didn't really think there'd be that much difference in heat output between the two processors, so didn't assume I'd have to upgrade my cooling to run the faster processor. Fact is though, upgrading the cooling made a great difference and turned out to be necessary.

The only thing I don't understand though is why the E6550 runs cooler with this cooling setup than the E4400 did. Both idle at 29-32, but the 6550 does it with much lower fan speeds on the CPU HSF (1000+ RPM for the E4400, and 800-900 for the E6550). Right now I've got three programs open (HWMonitor, RealTemp, and Firefox (with five tabs open)), and I'm sitting at only 32°C with 822-823RPM on the fan, with the E4400 the fan would be up to 1100+ RPM by now and it'd be in the mid to high thirties or maybe even low forties.

Also, socket temps are lower at idle (27-30°C for 4400, 23-24°C for 6550). (Haven't tested for load temps yet.)

Another interesting thing is that HWMonitor is actually able to report the correct temps on this chip, not sure why.

The system's been running and been stable for over an hour now without any issues. I've opened several programs, checked my email, surfed around this site and a few others, and not a single hiccup.

 

[SamSaveMax]

Why? Yes....you've answered to your own questions. Better specs: bigger fan + copper vs smaller fan spins higher and non copper.

The only thing I don't understand though is why the E6550 runs cooler with this cooling setup than the E4400 did. Both idle at 29-32, but the 6550 does it with much lower fan speeds on the CPU HSF (1000+ RPM for the E4400, and 800-900 for the E6550). Right now I've got three programs open (HWMonitor, RealTemp, and Firefox (with five tabs open)), and I'm sitting at only 32°C with 822-823RPM on the fan, with the E4400 the fan would be up to 1100+ RPM by now and it'd be in the mid to high thirties or maybe even low forties.

I traded up to a Zalman CNPS8700 NT cooler

 

[Tech Tweaker]

Why? Yes....you've answered to your own questions. Better specs: bigger fan + copper vs smaller fan spins higher and non copper.

You don't seem to understand my point, I ran both chips with the same exact fan (Zalman CNPS8700 NT), and the faster one with a higher FSB runs cooler.

I'm not stupid enough to not be able to figure out that a bigger fan with a copper heatpipe would run cooler, that's fairly obvious.

I ran the E4400 under it first for a couple of weeks to test the cooler out and see what kind of temps I got with it, and I was expecting the E6550 to run hotter, but strangely I got the exact opposite result. In fact running overnight the 6550 got down to 24°C, the 4400 never got below 28°C. Now I'm using the same thermal paste as I did with the 4400 and the same application method.

 

[RGone]

What appear to be lower temps with a different cpu, does not really mean that much since the sensors from Intel and now even AMD are designed to be more accurate at greater temps because that is where the cpu is using temp to throttle for safety and economy. Even cpus of the same model may show a different idling temp on the same cooler. Heck that could happen based on the IHS that is attached to the cpu. So a couple of degrees or so at idle don't mean much as it has come to be known.

 

[jason4207]

I think the E4400 uses more of a TIM like substance between the CPU DIE and the IHS, and the E6550 uses more of a solder like substance that wicks heat more efficiently to the IHS.

And +1 to what RGone said as well; these sensors have never been that accurate at idle temps as reporting accurate temps to the user was not what they were designed to do. Uncleweb and friends did extensive testing on these sensors years ago soon after RealTemp was first released. They even went so far as to drill a hole in the IHS and install a thermal probe.