Is the stock Core I5 Cooler Pathetic, or do I have a problem?

[henilgas]

CPU: Intel Core i5 750 - No overclock
Cooling: Stock intel cooler
Mobo: Gigabyte GA-H57M-USB3, Intel H57
GPU: 1GB XFX HD5770, PCI-E 2.0(x16)
RAM: 4GB (2x2GB) Corsair DDR3 PC3-10666
OS: Windows 7 Home 64 Bit
HDD: 1TB Samsung HD103SJ Spinpoint F3
Case: Antec Sonata III Case

- Assembled PC, no issues
- Booted, installed Win7
- Installed all the various drivers and mobo utilities.
- Installed latest Radeon drivers
- Hmmm, my CPU is idling @ about 45C. That seems pretty high.
- Hmmm, the stock fan is quite noisy and busy
- Download Prime95, run default torture test
- Temps. immediately start spiking 65, 70, 75, 80...
- Beep beep beep beep beep
- Stop test, has been running for all of 10 seconds.
- Back to idling @ 45C.
- Leave running overnight to settle in.
- Still idling @ 45C.

- Reopen case, carefully remove stock cooler.
- Thermal pad has melted evenly enough. doesn't cover the entire die though, hmmm.
- Reseat cooler, being careful to ensure that it snaps in.
- Reboot, now idling @ 42C.
- Rerun torture test.
- Beep beep beep beep beep
- Test stopped before exceeding 70C (max safe temp for I5 750).
- Took about 5 seconds to reach this value

- Go and buy some Akasa TIM remover, and paste.
- Carefully remove cooler.
- Scrape/wipe melted TIM pad from CPU/Cooler
- Painstakingly apply the thinnest possible layer of paste.
- Reseat cooler.

- Reboot, now idling @ 20C. Yay!
- Idle temp slowly climbs to 38C. Aww.
- Rerun torture test.
- Takes about 40 seconds to reach 70C, but was steadily increasing.
- Test allowed to continue. About another 60 seconds to reach 80C.
- Test stopped.

- Head scratching.
- Internal speaker gives random beep. It's been doing that every so often, no idea why.
- Some googling suggests it may be a sensor glitch with the I5 750. Unrelated problem.
- Walk away from machine and leave idling for a few hours.
- Come back and machine has rebooted, telling me it has encountered a stop screen.
- Oh FFS.

...can anyone tell me what the story might be with my CPU? Is this just a hot CPU? Is the stock cooler just that pathetic that it can't run the CPU at load? Bear in mind, nothing in this box is OC'd. And while it idles quietly enough, as soon as it's asked to do something relatively mild (like opening a browser), there's a lot of varispeed noise from the fan and it spins up and down. I have made no BIOS changes (F5 GB BIOS) other than upping the CPU temp alarm from 60C to 70C.

Is it time to send these components back?

As as aside, does anyone know what the story with the random beeping might be? Nothing interesting appearing in the event logs...

/HG

[Phage]

Hmmmm you did everything obvious. The next thing would be to check with another cooler or case. I assume the case exhaust fans are working ? Does it still happen with the side off ?

[henilgas]

you did everything obvious.

...this is actually useful to know. It's been a few years since I put a box together, and things have moved on a lot since then.

The next thing would be to check with another cooler or case.

...don't have another case to hand. The Antec Sonata 3 is a pretty well regarded case though, and the case fan is working and set to high. Moreover, the PCH temp is pretty steady, at around 40C. It really seems to be a CPU specific temp problem.

Nor do I have a suitable cooler to hand. My decision is whether or not to invest more cash in a shiny new cooler, or to cut my losses now and start returning stuff. The stock cooler I have seems mechanically sound, so I assumed it would be capable of cooling a non OC'd CPU at load.

If it isn't designed to handle this load (which is pathetic, frankly), then yes, it's time to get a new cooler. If it is designed for this, then I either have a dodgy cooler or a dodgy CPU (or, I dunno, dodgy mobo sensors?).

I assume the case exhaust fans are working ?

There's only the one. It works just fine, has 3 speeds. Setting it to the highest speed doesn't make a lot of difference.

Does it still happen with the side off ?

Yes, again, doesn't make a lot of difference. Actually, I haven't tested this since I reapplied the paste.

But these questions do lead me to reason that the case temp is reasonably ok. Therefore any temp problem has to exist between the CPU and the cooler.

Hmmm. Returning a CPU/cooler combo is pretty easy...

[Phage]

Aye - you could waste hours otherwise.

[KidChameleon]

The very rapid increase suggests the heatsink is not making contact. If you're absolutely sure the heatsink is installed properly and is getting enough pressure against the CPU, it could be that the heat spreader is not making contact with the cpu die. That would probably mean an RMA, but you might want to try other stuff first.

My standard i5 heatsink cooled mine to under 30*C at idle and up to 65-70*C after 10 minutes of encoding video, but around 50*C for normal tasks like gaming. But when heatsinks are making full contact it should take a couple of minutes for the temperature to get very high (60*C+ with a stock heatsink).

[Terbinator]

The original cooler is pap - I got an overheat issue with my stock i7 860 one when running the Linpack burn in test. Yes, it was fitted correctly in a well ventilated case. IMO, buy a third party one, an Arctic Cooler Pro 7 (version 2) would be a decent replacement.

Edit: or Kid Chameleons point is worth looking over - make sure the cooler is installed properly, don't worry about applying a bit of force to make sure the pin is all the way through the motherboard until it clicks - rather than just the pin head sticking trough - do this for all 4 pins.

[Georgy291]

i got a i5 750, never even tried the stock cooler, i bought a venomousX right away.

And it idles at room temp and loads at....not sure never tried

atm my i5 is running at (cough cough scan people cough) 3.8ghz and it idles at 40C under full load is at 55C

CHECK YOUR CORE VOLTAGE

sorry for caps but do so in bios, might be set too high, either clear cmos and then re do all your tests or just check it. it should be aroun 1.0 1.1at most

[Englander]

I was so shocked when I saw the paltry amount of metal included in the stock heatsink that I had to take a picture and show it to my friend.

Definitely go for an aftermarket cooler!

[henilgas]

Thanks everyone for the input. An update:

- Reopened case.
- Removed heatsink. Cleaned off yesterday's freshly applied paste.

EDIT: and yes, the cooler is surprisingly insubstantial. Once I get this build stabilised, I'll try and pick up a more effective, quieter cooler. One of the objectives of this build was for it to be relatively quiet and low power. The stock cooler is by far and away the noisiest component in the box. Anyway...

- Reapplied paste to CPU using finger wrapped in plastic rather than spreading with plastic card.
- Used slightly thicker layer of paste this time.
- Reseated cooler, using a wee bit more force this time, definitely got clicks all round.
- Rebooted. Left case door off.
- Idling @ 38C. 100% load @ 68C, stayed stable at this temp for 5 mins.
- Yay!
- Replaced case door. Restarted prime95.
- Prime95, Core 2 thread fails with a hardware error.
- Hmmm. That doesn't sound good. Random error?
- Restart Prime95. Core 2 thread fails again, hardware error.
- Text of error: FATAL rounding error. Got 0.5, expected value less than 0.4.
- Hmmm.
- Fire up Chrome to google error.
- Chrome crashes with sad face.
- WTF?
- RealTemp crashes.
- Event viewer, a bunch of components and services have crashed or errored recently.
- RAM issues? Maybe all the CPU temp stuff was a red herring.
- This might explain yesterday's spontaneous reboots and the intermittent beeping.
- Currently running Memtest. So far 400,000 errors.
- It all seems to be focused in one 2GB stick.
- Bad RAM stick, bad installation, voltage issues, or dodgy mobo?

...I'm going to let Memtest finish, remove of the sticks and rerun it. Any other advice would be appreciated.

/HG

[Terbinator]

What voltage is the RAM designed to use ?

[henilgas]

What voltage is the RAM designed to use?

It's this stuff:
CMV4GX3M2A1333C9 - 4GB (2x2GB) Corsair DDR3 PC3-10666 (1333) Non-ECC Unbuffered, CAS 9-9-9-24, 1.50V

...I haven't made any changes to RAM voltage in the BIOS. And I think it's at 1.5V.

[Terbinator]

Okay, I would test 1 stick of RAM at a time so you can see if one, or both, are faulty. My guess would be that your RAM is borked TBH - at least you've got the temperatures sorted now

[PetrJoine]

I got the i9 64gb 2080 OLED and so far I really enjoy it. For sure it will throttle at 5ghz, I have mine set at 4.3ghz all cores with -140uv and temps max around 90, The keyboard is so much improved over the old, and gaming several hours the keys are nice and not hot at all.
GPU OC at 155core 400mem is the 3d mark I got.