Overclock.net banner

1 - 20 of 29 Posts

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #1 ·
I've been running a few CPU's lately (3570K, 2500K, 3770K) and lastly a 2700K. All the CPU's (except for the 2500K and 2700K) where de-lidded, lapped IHS and die and ran with CLU on an H110i with x2 140mm at full blast.

I've pushed them all pretty hard (1.6v and up) and I'm having strange temp issues with this 2700K however. The 3770K runs at about 93-94 under IBT at 4.8ghz with 1.656v and it'll hold that temp for hours. Same with the 3570K and 2500K.

The 2700K clocks really well, but temps are strange. 5ghz at 1.508v under IBT I'm getting into the high 90's and even touching into the 100c range. It's not lapped, but I am using CLU with the H110i. It's temp spiking like normal paste would, and 1.5v = almost thermal throttling?

Would it be possible the solder that holds the IHS on has degraded over time?
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #2 ·
Okay, well I de-lidded the 2500K very easily... First soldered CPU and it went really smooth. Two razor blades on opposite ends stuck under the IHS, held one of the blades with a pair of pliers and heated the IHS with a long reach lighter and pow... Off came the IHS lol.

Razor bladed the solder off and lapped the CPU die a little with some 3000 grit to clean up the small amount of Indium left. Dropped it in a spare board, using a pre-lapped IHS off of the 3570K and it booted first try.

Problem is, I don't have any CLU left and USPS dropped the ball on delivery. I have two packages that was supposed to be here the 23rd and according to them it hasn't left the sellers state yet. Anyways... I'll update back with the 2700K before and after temps once I get it de-lidded. We'll see if the solder degrades over time or if I just wasted a bunch of time.

So far I've been running it since yesterday at 5.3ghz. 1.608 -1.616 load voltage. Rock solid bench / gaming. Only IBT hits the thermal limit and throttles.

2475546
2475547
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #3 ·
Interesting. So I delidded, lapped the CPU die as well as IHS and used CLU. Now I'm getting higher temps... That's a first for me. 1.504v gets me into the high 90's in IBT at 5ghz. Something isn't right. The CLU was really weird I noticed too. I might have to get some more because this stuff was really stiff and would not stick to anything.
 

·
Registered
Joined
·
2,086 Posts
Interesting. So I delidded, lapped the CPU die as well as IHS and used CLU. Now I'm getting higher temps... That's a first for me. 1.504v gets me into the high 90's in IBT at 5ghz. Something isn't right. The CLU was really weird I noticed too. I might have to get some more because this stuff was really stiff and would not stick to anything.
You delided a soldered cpu lol? That's probably why your temps are worse. Gotta clean that die perfectly.
 

·
Registered
Joined
·
2,086 Posts
Interesting. So I delidded, lapped the CPU die as well as IHS and used CLU. Now I'm getting higher temps... That's a first for me. 1.504v gets me into the high 90's in IBT at 5ghz. Something isn't right. The CLU was really weird I noticed too. I might have to get some more because this stuff was really stiff and would not stick to anything.
I re read your post. K so go direct die with liquid metal and send her. Other than that what kinda life span you getting outta them chips running them that hot with that super high voltage
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #6 ·
Oh believe me I'm not worried about life span lol. None of these are really good clockers anyways.

I've never had CLU that didn't spread. It's like it wouldn't stick to the surface at all. Once I put the block back on it would just squish out the side. When I took the block back off, the center of the IHS was almost bare.

Yeah, the 2700K was die lapped as well. Mirror finish = no stick for CLU and roughed up with 600 grit and the CLU still didn't stick. Never had that issue with the 2500K, 3570K or 3770K. Very strange. I ordered more paste anyways, so I'll try it again.

I ran the 3770K at 4.8ghz with 1.656v and it got to the low 90's after about an hour of IBT. The 2700K rockets up to the high 90's in seconds at 1.504 voltage.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #7 · (Edited)
This 2700K is a strange one. 5ghz at 1.5v using arctic silver 5, I'm getting 85-86c max under IBT after about 10minutes at full load. 5.1ghz at 1.52v = 105c instantly... Very strange.

Here is a recent pic of my 3770K with the same cooler, same board, ram, case etc. Using CLU. And... I'm using the same IHS from the 3770K as well on the 2700K.

3770K @ 4.8ghz with 1.640 peak voltage I'm I'm not even getting to 100c after 50 standard IBT runs.
2476327
 

·
Registered
Joined
·
2,086 Posts
This 2700K is a strange one. 5ghz at 1.5v using arctic silver 5, I'm getting 85-86c max under IBT after about 10minutes at full load. 5.1ghz at 1.52v = 105c instantly... Very strange.

Here is a recent pic of my 3770K with the same cooler, same board, ram, case etc. Using CLU. And... I'm using the same IHS from the 3770K as well on the 2700K.

3770K @ 4.8ghz with 1.640 peak voltage I'm I'm not even getting to 100c after 50 standard IBT runs. View attachment 2476327
them 5ghz plus runs are fun as hell :p i popped a kaby lake quad into my x299 board and slammed it with vcore on water.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #10 ·
Yes they are lol. I ran the 2700K at 5.3 before delidding. It wasn't stable (too hot) but never crashed while gaming or benching. I know it'll do 5.3 stable and stay under 105c, its just getting the mount issue fixed. I'm not 100% sure what the issue was, but as soon as more CLU shows up I can put it back in and see whats up.
 

·
Registered
Joined
·
2,086 Posts
Yes they are lol. I ran the 2700K at 5.3 before delidding. It wasn't stable (too hot) but never crashed while gaming or benching. I know it'll do 5.3 stable and stay under 105c, its just getting the mount issue fixed. I'm not 100% sure what the issue was, but as soon as more CLU shows up I can put it back in and see whats up.
Had my 7940x at 5.2 last night on water before I pulled it.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #12 ·
I think I screwed the chip up delidding it... It still runs fine, I'm sitting at 5ghz with 1.496v under IBT and I'm getting into the low 90c range after about 100 runs on standard.

Out of all the CPU's I've ran (2500K, 3570K, 3770K) I could easily run in the 1.6-1.65v range and keep it under thermal throttle. This 2700K will peg the 105c limit with 1.55v within a few seconds.

The new CLU spread fine, no issue there. Pump is working fine, fans are working fine, fins are clean. I dropped the 3570K in for fun, and I was at 5.1ghz with 1.66v running fine. High 90's but no throttling. Ah well, I guess I "kinda" broke the 2700K.
 

·
Registered
Joined
·
2,086 Posts
I think I screwed the chip up delidding it... It still runs fine, I'm sitting at 5ghz with 1.496v under IBT and I'm getting into the low 90c range after about 100 runs on standard.

Out of all the CPU's I've ran (2500K, 3570K, 3770K) I could easily run in the 1.6-1.65v range and keep it under thermal throttle. This 2700K will peg the 105c limit with 1.55v within a few seconds.

The new CLU spread fine, no issue there. Pump is working fine, fans are working fine, fins are clean. I dropped the 3570K in for fun, and I was at 5.1ghz with 1.66v running fine. High 90's but no throttling. Ah well, I guess I "kinda" broke the 2700K.
just a high leakage chip
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #14 ·
Well the 2700K is dead... I lapped it a little too much I think . I booted it up, 5.2ghz at about 1.55v and it ran about 10 minutes with reasonable temps and then locked up. After that, no CPU detected. Oh well, time to grab another 2700K I guess.

Everything else is fine however. Running the 3570K at 1.656v and in the low 90's. Does 5ghz all day, but 5.1ghz I get WHEA's at even 1.704v.
 

·
Registered
Joined
·
2,086 Posts
Well the 2700K is dead... I lapped it a little too much I think . I booted it up, 5.2ghz at about 1.55v and it ran about 10 minutes with reasonable temps and then locked up. After that, no CPU detected. Oh well, time to grab another 2700K I guess.

Everything else is fine however. Running the 3570K at 1.656v and in the low 90's. Does 5ghz all day, but 5.1ghz I get WHEA's at even 1.704v.
jesus you are just torturing them chips. 1.7 through ivy on water?!?!?!?!?
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #16 ·
They are pretty cheap CPU's tho. Yeah, 1.704v under IBT. I was 1 degree below thermal throttling for 50 runs (on standard) with IBT lol. Right at the limit. Still threw some WHEA's tho so I dropped it back to 5ghz. I wish I wouldn't have delidded the 2700K, that chip did 5.3ghz semi stable but ran hot. After the delidd, it wouldn't even do 5.2ghz and ran even hotter.
 

·
Registered
Joined
·
2,086 Posts
They are pretty cheap CPU's tho. Yeah, 1.704v under IBT. I was 1 degree below thermal throttling for 50 runs (on standard) with IBT lol. Right at the limit. Still threw some WHEA's tho so I dropped it back to 5ghz. I wish I wouldn't have delidded the 2700K, that chip did 5.3ghz semi stable but ran hot. After the delidd, it wouldn't even do 5.2ghz and ran even hotter.
yea lol
 

·
Graphics Junkie
Joined
·
3,370 Posts
Try prime95 or some other CPU stress test instead. Sandy bridge will get hotter in ITB than any other stress test.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #19 ·
Try prime95 or some other CPU stress test instead. Sandy bridge will get hotter in ITB than any other stress test.
Oh I know. That's why I run IBT, it's more of a how hot is it going to get than a stability test. With Sandy it's harder because there are no WHEA errors to go off of. Ivy I just IBT to see where I'm at with thermals, then go off of WHEA errors to get stable.
 

·
Registered
Joined
·
1,044 Posts
Discussion Starter · #20 ·
Well die lapping the 3770K did wonders lol. Very high centered while lapping. I used 400 grit dry to lap it. Left it with a 400 grit finish.

Used the smallest amount of CLU I possibly could between the die / ihs and ihs / water block. And.... I broke one of the studs off mounting the block cranking it down too hard...

Lapped water block side of the IHS as well. Left it with a 400 grit finish. The inside of the IHS was just roughed up with the pad that came with the CLU. Same for the H110i block, just roughed with CLU pad. Obviously everything was very well cleaned with 91% isopropyl priory to assembly.

I ran it to 4.9ghz but it needed 1.704v just to stay out of WHEA-ville with a quick and dirty bench / temp run with IBT. 4.8ghz I can probably go a little lower, but that thermal head room! At 1.704 I was only in the low / mid 90's. 73F ambient.

2477152
 
1 - 20 of 29 Posts
Top