Ivy/Haswell IHS adhesive

Bluntwrapped · 8 Apr 2014 at 11:24

Hey guys, was just wondering if the black adhesive on Ivy/Haswell is sometimes visible beyond the base of the IHS, if it sticks out a bit and can been seen on top of the PCB?

Reason I ask is I just bought a cherry picked 3770k, I'm trying to figure out if it has been delidded and then had the IHS glued back on.

Cheers

wazza300 · 8 Apr 2014 at 11:26

only very slightly from my experience,if its big lumps then its been delidded/re stuck

try it and see,the temps will tell you

Bluntwrapped · 8 Apr 2014 at 12:29

Thanks for the reply, I will do. I'm still at work so it'll be tonight before I can test. Just had a closer look and wondered if this is a sign of being delidded and stuck back.

As a quick test on stock cooling it should be safe enough to boot into windows using manual volts? Say around 1.3v? I'm not going to stress it. Would just be for a few minutes to find out which multi I can boot in to Windows at. Would be much easier than temporarily rebuilding my loop.

wazza300 · 8 Apr 2014 at 14:04

idk how you'd do on stock cooler regarding temps

if its a decent chip you should do 4.5ghz with just 1.212v or round abouts

Jokester · 8 Apr 2014 at 14:21

Got any pics of it?

Bluntwrapped · 8 Apr 2014 at 15:37

The seller says it'll do 5ghz @ 1.28v. If this is true I'll be very very happy

I'll put up a couple of photos this evening. I'm really struggling to get a decent photo with my iPhone camera.

Bluntwrapped · 8 Apr 2014 at 19:29

Some photos. Hard to get a good shot, but you can see the black adhesive along the top and down the side. Is this normal?

I'm using one of these at the moment. my loop needs to be rebuilt and the rest of the build needs finished but i thought i'd have a quick try of booting in to windows @ 5ghz using 1.28v. i wish i had a spare h100i lying around for simplicity.

i know this doesn't mean much but here we go

If difference in core temp is anything to go by this doesn't look particularly bad. I don't know if a delid is worth the risk but I would like to see how far it can go - with some decent cooling

setter · 8 Apr 2014 at 19:47

Drop it down to 4.5ghz ono on the vcore the seller has stated it's capable of doing higher on. Then test with a quick blast on p95 to see what temps hit. Should be a good indicator of wether it has been scalped by the previous owner.

almighty15 · 8 Apr 2014 at 20:15

My 3770k hasn't been de-lidded but it does have a really good even spread between the cores in terms of temperatures.

Doogles · 8 Apr 2014 at 20:38

Where do you get cherry picked ones? Mine will do 4.3ghz at 1.35v

wazza300 · 8 Apr 2014 at 21:01

it lookslike its been delidded to me? seems a lot of adhesive squelching out

if temps are ok then leave it at that

week 18 is one of the best clocking batch

Bluntwrapped · 8 Apr 2014 at 22:21

setter said:
Drop it down to 4.5ghz ono on the vcore the seller has stated it's capable of doing higher on. Then test with a quick blast on p95 to see what temps hit. Should be a good indicator of wether it has been scalped by the previous owner.

Gave it a very quick blast, temps went through the roof :eek:

so I stopped it. Prob a sign it hasn't been delidded?

Doogles said:
Where do you get cherry picked ones? Mine will do 4.3ghz at 1.35v

I said cherry picked but what I actually meant is I got this used and tested. The description and batch number caught my attention. The best place to look would prob be the members market.

wazza300 said:
it lookslike its been delidded to me? seems a lot of adhesive squelching out if temps are ok then leave it at that week 18 is one of the best clocking batch

Going by the above temps I'd say it hasn't had the lid removed but I could be wrong since the cooling is awful.

Just had a quick read of 8Pack's 3770k batch thread. Seems like this could be a gem with some proper cooling :cool:

can't wait to finish the build now.

setter · 8 Apr 2014 at 22:39

With temps like that id say it hasnt been delidded, iirc mine at 4.5ghz on 1.34v was hitting the low 80's, albeit that was with a big dual tower heatsink with x2 high speed/pressure fans.

Bluntwrapped · 9 Apr 2014 at 13:43

Yeah definitely looks untouched, I have just compared it to another few 3770k closeups and a bit extra adhesive is normal. Now to decide if its worth taking the risk to delid. I'm thinking probably not.. but then I don't know if I'll ever be fully satisfied if I don't.. lol

Was thinking a piece of 3lb fishing line might do the job safely, with enough to friction cut through the glue but then again that might cause static.. has anyone tried the hair dryer and vice method?

setter · 9 Apr 2014 at 18:33

Hair dryer to heat it up a little to soften the adhesive, then try the fishing line. Could be a good new method.

thel33ter · 9 Apr 2014 at 20:09

Fishing line is a good shout, else its not too bad with a very thin razor blade, you just need to make sure you know where the IC's are on the cpu so you can avoid them. The vice method always looks really dangerous to me..

wazza300 · 9 Apr 2014 at 20:17

id just settle on 4.5ghz and run it as is

if you must delid then the vice method or two blocks of wood method seems best,razor is too risky

ubersonic · 10 Apr 2014 at 08:31

Bluntwrapped said:
Prob a sign it hasn't been delidded?

Either that or it's been re-lidded. The issue with stock IB/Haswell isn't the TIM that's actually okay, the issue is the is too much of it as the glue holds the IHS higher above the die than on soldered chips. If somebody has been running this with direct cooling (with the IHS off) then they may have used too much glue to replace the IHS and replicated the standard issue or made it worse.

Bluntwrapped · 10 Apr 2014 at 14:35

wazza300 said:
id just settle on 4.5ghz and run it as is

if you must delid then the vice method or two blocks of wood method seems best,razor is too risky

Would be a bit of a waste of a great chip settling on 4.5ghz. Once it's under water temps will be much lower. 5ghz and above is where I'm aiming for.

ubersonic said:
Either that or it's been re-lidded. The issue with stock IB/Haswell isn't the TIM that's actually okay, the issue is the is too much of it as the glue holds the IHS higher above the die than on soldered chips. If somebody has been running this with direct cooling (with the IHS off) then they may have used too much glue to replace the IHS and replicated the standard issue or made it worse.

Yeah, this seems incredibly stupid coming from a manufacturer of microprocessors. They work to the tiniest of micron yet when it comes to putting the lid on they use a massive amount of glue - way more than needed. You would think the results from Ivy would have changed this for Haswell but apparently not..

The seller did say it hasn't been delidded but from past experience on the bay some have a habit of talking BS. I wanted to double check before I decide if I'm going to delid. Looking at the temps above and new 3770k images it doesn't look as if it has been delidded previously.

Just wish there was a 100% way to delid. The razor blade method is way too risky and I really don't like the idea of battering a great chip with a block of wood and hammer.

wazza300 · 10 Apr 2014 at 14:46

you might be ok on water,max I could do was 4.6ghz on air/non delid,and yours is a slightly better chip regarding voltages