Temperatures strange after delidding i7-3770k

[DropTheBase]

Hi guys, I was hoping you could give me some advice.

I recently delidded my i7-3770K and the process went very smoothly. No scratches or anything.

However, my temperatures after delidding are behaving funny. At idle, there is a 15 degree temperature difference between the first and last core (28, 23, 23, 15). When at max load, there's a 45 degree difference. The temperature fluctuates very rapidly depending on what task I'm doing.

When I stress test (which I don't run for more than a few seconds), my first core jumps up to 99 degrees.

Based on how smoothly the delidding process went, I'm doubting that I nicked the die or anything (I didn't notice any scratches). I figure if I did something like that, the CPU just wouldn't work at all.

I am thinking that I must have applied the thermal paste incorrectly, which might explain the difference in core temps. I used Coolaboratory Liquid Pro. And I applied it using the half rice grain method.

Putting on the IHS wasn't so smooth, I guess. it moved around a lot before I finally locked it in place. Maybe THIS is when something went wrong?

Any advice? I'm leaning toward taking off the lid again and checking out the thermal paste. I heard it's hard to remove the Liquid Pro paste. How should I go about removing it? I'm scared to damage the cpu die. Will just alcohol and swabs work?

Thank you very much. Any help is appreciated.

 

[ozan]

I de-lidded too, yet couldn't trust rice method to spread nicely. I spread it with the provided earbud and it works fine. I have 5 deg difference between cores at max and done it 2 times with no difference in between. İf you've done the de-lidding recently, all the better. It wont have eaten into the ihs and the process will be smoother.

Alcohol is useless in removing liquid pro. You need to wipe it off till the die and the ihs are totally free of any leftover material. If you have touched it, then you may want to clean the die with alcohol.

Congrats on de-lidding

 

[Luis_GT]

you have to spread it with the provided earbud, first time I tried it with the rice setup and it didn't spread well. Once I spreaded it, temps where closer, though I still have 1 core that's considerably cooler than the rest, but it was that way even before delliding.

 

[bluezero5]

I see a few reasons for that.

1, pressure of heatsink not strong enough, resulting in air bubbles and poor contact, this can be partially fixed by screwing down the heatsink tighter.

2, Poor TIM spread, which spread method you used? To be save, you should always use the thin film method, cause we are no longer talking about the IHS, you NEED FULL contact between the die and IHM, if you used the pea method, you will have edges of the DIE without TIM, and heat can spike there (and can damage). This is why I ALWAYS tell people to use the thin film method.

Good luck!

 

[DropTheBase]

After reading these responses, I am almost certain that this is my problem (lack of TIM spreading), based on the temperature behavior.

Hopefully I didn't damage the CPU too badly; I never ran the computer for more than an hour total.

This has given me a lot of hope! I'll post back with the results!

Thank you again.

 

[DropTheBase]

Hi guys, I just wanted to thank you for the great advice!

I took off the lid and the thermal paste had barely spread at all! A huge portion of the die had been without thermal paste.

As for cleaning it off, there was some that would NOT come off. The instructions said to use their abrasive polisher to get it off. I did not realize how abrasive it was, one swipe and it scratched the die tremendously. Not wanting to completely sand it down, I just left it as it was and used the spread method.

Anyway, the results are phenominal! (They should be considering I have 7 120mmx120mm worth of radiator!) There's a gradient in core temperatures, probably because some of the liquid pro was still on the die

In the future, maybe I'll actually clean off the liquid pro properly by being rougher on the die.

In the mean time, I'll be overclocking!

---

Also, safety tip! I was putting on the cap for the syringe of liquid pro, it went through the cap and directly into my finger. Luckily my thumb wasn't on the plunger otherwise I would have directly injected it into myself (in addition to the tiny bit that made it in). BEWARE of the needle!

 

[bluezero5]

awesome!!!!

looking very good.

 

[Stewart@MSD]

Good work mate, I was going to add that there is no way the grain method will work with liquid pro, but I see you have rectified it yourself.

Your further attempt to replicate a T1000 did make me giggle to myself I must admit. ROFLOL

 

[LoneWolf121188]

Very nice! Glad you got the temperature asymmetry worked out. I should add that piece of advice to my stickied thread (are people still using that?).

 

[Zenos]

Congratulations.