Confirming delid success.

[Rayniac]

Hey guys I booted up my delidded i5-3570K system for the first time and things seem ok for now. No issues.

However I have read some dude claim that while delidding his system he accidentally scraped his board with a razor and as a result, even though his CPU still worked, he had lost some PCIe lanes.

I also damaged the surface a tiny bit. The damaged area is tiny, hardly even a scratch but you can still see a smalll glint of copper.

I was wondering what should one in my position do to fully analyze my CPU's condition?

If my system booted windows fine can the RAM still have problems? Should I do a RAM check? What about PCIe lanes and other stuff?

 

[MagnusTheBrewer]

What was your goal in delidding? Did you reach that goal?

 

[Rayniac]

What was your goal in delidding? Did you reach that goal?

Better thermals. While they are lower, they're not as good as I anticipated. I hope stress testing will tell a different story. But first I need to get my fresh win 10 install out of the way...

 

[Wall Street]

GPU-Z can confirm that your graphics card is still negotiating over a 16x link and Memtest 86 can confirm that it still can write to all of the RAM (no need to do the whole test, if it can do the first 2-3 patterns it is fine to make sure if can write all bits).

 

[Rayniac]

GPU-Z can confirm that your graphics card is still negotiating over a 16x link and Memtest 86 can confirm that it still can write to all of the RAM (no need to do the whole test, if it can do the first 2-3 patterns it is fine to make sure if can write all bits).

Thanks Wall Street. I'll do that.

 

[MagnusTheBrewer]

Better thermals. While they are lower, they're not as good as I anticipated. I hope stress testing will tell a different story. But first I need to get my fresh win 10 install out of the way...

That was your purpose. Goals are either specific temp readings under load or, a percentage decrease from the original configuration.

 

[Rayniac]

That was your purpose. Goals are either specific temp readings under load or, a percentage decrease from the original configuration.

The purpose of this thread was to learn how to find out if my CPU still works 100%.

 

[MagnusTheBrewer]

The purpose of this thread was to learn how to find out if my CPU still works 100%.

That's a goal.

 

[nathanddrews]

I hope it all works out! I've got an extra 3570K I'd like to experiment with.

 

[Rayniac]

Update: CPU still works fine despite how many times I've opened the lid and applied and reapplied liquid metal to it and closed it and opened again... No errors in memory testing and I still have all my PCI-e lanes.

Still I've been having some unsatisfying results with temperatures First time I applied liquid metal to the die and closed the lid, then used MX-4 between the IHS and waterblock. Temperatures were actually a lot better than what I had at stock but I got greedy. I lapped the IHS and my Cooler Master Seidon 240V and put liquid metal between those two. Suprisingly my temperatures became worse! I tried reinstalling them but that didn't help. Temperatures rose again. Maybe the cooler wasn't on tightly enough, I don't know. Finally I decided to take off the IHS and reapply liquid metal again as I though that maybe the Q tip that I used to spread it left some lint on the die and made the contact worse. This time I spread it with a plastic bag and my finger. Between the IHS and the waterblock I used MX-4 again since I didn't want any more complications and wanted to just be done with it. I also used some washers between the waterblock's mounting screws and nuts to make the waterblock press on the IHS harder because it looked like that was a problem.

That seems to have done it. Running at 4.3GHz and around 1.25V the highest temperature I got with Prime95 was around 70°C. About 67°C is what I used to get while just playing Bad Company 2 with stock clocks and undelidded. So I think I've finally managed to make some definite improvements. Hopefully no further issues arise.

Maybe I can afford to lower voltages more? I don't know if I should go for it. I'm using the offset voltage mode and at 1.6GHz it gets about 1V. Would any less make it unstable at low power state?
One thing is for certain if I have to open this computer one more time I think I'm just going to order a new mobo and a CPU and be rid of the coil whine as well.

Edit: I'm not going to pursuit 4.5GHz because I couldn't boot win 10 even when I used more than 1.3V.

 

[RichUK]

That was your purpose. Goals are either specific temp readings under load or, a percentage decrease from the original configuration.

You're describing an objective, not a goal.

 

[Ed1]

Update: CPU still works fine despite how many times I've opened the lid and applied and reapplied liquid metal to it and closed it and opened again... No errors in memory testing and I still have all my PCI-e lanes.

Still I've been having some unsatisfying results with temperatures First time I applied liquid metal to the die and closed the lid, then used MX-4 between the IHS and waterblock. Temperatures were actually a lot better than what I had at stock but I got greedy. I lapped the IHS and my Cooler Master Seidon 240V and put liquid metal between those two. Suprisingly my temperatures became worse! I tried reinstalling them but that didn't help. Temperatures rose again. Maybe the cooler wasn't on tightly enough, I don't know. Finally I decided to take off the IHS and reapply liquid metal again as I though that maybe the Q tip that I used to spread it left some lint on the die and made the contact worse. This time I spread it with a plastic bag and my finger. Between the IHS and the waterblock I used MX-4 again since I didn't want any more complications and wanted to just be done with it. I also used some washers between the waterblock's mounting screws and nuts to make the waterblock press on the IHS harder because it looked like that was a problem.

That seems to have done it. Running at 4.3GHz and around 1.25V the highest temperature I got with Prime95 was around 70°C. About 67°C is what I used to get while just playing Bad Company 2 with stock clocks and undelidded. So I think I've finally managed to make some definite improvements. Hopefully no further issues arise.

Maybe I can afford to lower voltages more? I don't know if I should go for it. I'm using the offset voltage mode and at 1.6GHz it gets about 1V. Would any less make it unstable at low power state?
One thing is for certain if I have to open this computer one more time I think I'm just going to order a new mobo and a CPU and be rid of the coil whine as well.

Edit: I'm not going to pursuit 4.5GHz because I couldn't boot win 10 even when I used more than 1.3V.

Every chip is different, but my i5-3570k I noticed OC to 4.2 and you don't need any voltage change ,That is on a Asus P8Z77v pro MB, idle voltage is like 0.940 and under load about 1.100.
Going to 4.3 same thing but I upped LLC one notch, load voltage is 1.112 or so.
Moving to 4.4 and just needed to add some in offset, load voltage 1.150. then move to 4.5 and set load to 1.900-1.950. all tested with prime95.
My temps with 4.5 are around 67-68 and in games like BF4, BF1 low 60's with 212evo.

Now can yours run lower, only way is to test, move down slowly and check for stability.
I for longest time was running 4.3 and only lately tried higher cause I didn't want to go to 1.250v as that would put load heat more than I would like (wanted to keep below 70c).

 

[Rayniac]

It sure does seem like my 3570K doesn't clock well. My games started crashing at 4.3GHz with 50% LLC so I took it down a notch to 4.2GHz and upped LLC to 75%. Looks much more stable now. At least there's a high chance I won't have such bad luck with my next chip. *fingers crossed*

 

[IllogicalGlory]

It sure does seem like my 3570K doesn't clock well. My games started crashing at 4.3GHz with 50% LLC so I took it down a notch to 4.2GHz and upped LLC to 75%. Looks much more stable now. At least there's a high chance I won't have such bad luck with my next chip. *fingers crossed*

My 3570k hit 4.2GHz on the stock cooler and 4.2GHz on a Kraken X41. Didn't seem to matter how much voltage I put through it. I could just be inept at overclocking.

 

[Rayniac]

My 3570k hit 4.2GHz on the stock cooler and 4.2GHz on a Kraken X41. Didn't seem to matter how much voltage I put through it. I could just be inept at overclocking.

Did you delid yours?