Page 1 of 2

IVB TIM on Retail CPUs?

Posted: Mon Apr 30, 2012 3:38 pm
by DPete27
Anybody know if the TIM between the CPU and the heat spreader on IVB was only limited to Engineering Samples, or is it being found in retail versions also?

Re: IVB TIM on Retail CPUs?

Posted: Mon Apr 30, 2012 4:43 pm
by ColeLT1
I will have one in Wednesday (3570k). I de-capped half a dozen AMDs in the past, and IB looks even easier with no resistors? to hit on the top (see pic):
This was my socket939 3800x2 or 4800x2, can't tell:
Image

I have waterloop with an old V1 Dtek Fusion that will work for direct contact, if I de-lid I will do some stock clocked and overclocked capped vs uncapped results if anyone wants.

Dropping it in this rig ASAP:
http://imgur.com/a/tMWR8

Re: IVB TIM on Retail CPUs?

Posted: Tue May 01, 2012 8:23 am
by DPete27
Oh that would be excellent. Please post your findings as soon as possible! I too am looking at purchasing a 3570K in a couple weeks.

Out of curiosity, I tried Intel tech support chat last night and the rep told me that the retail versions will have TIM as well. When I asked if future steppings will go back to fluxless solder, he said there were no plans to go to solder with future steppings. I did make sure the rep knew I was talking about the interface between the chip and the heatspreader, not the heatsink. I sometimes question how much tech support reps really know or if they're just feeding you a load of BS to end the conversation, so it would be great if you can verify.

Re: IVB TIM on Retail CPUs?

Posted: Tue May 01, 2012 8:31 am
by Crayon Shin Chan
Why isn't anybody asking why they are using TIM? Fluxless solder can't be that much more expensive, they must have been using it all the time up until IB.

Re: IVB TIM on Retail CPUs?

Posted: Tue May 01, 2012 10:16 am
by Lucky Jack Aubrey
Crayon Shin Chan wrote:
Why isn't anybody asking why they are using TIM? Fluxless solder can't be that much more expensive, they must have been using it all the time up until IB.


TR apparently did ask that question:

Curious, we asked Intel about the interface between the Ivy Bridge die and the heat spreader. Intel has confirmed to TR that Ivy uses a "different package thermal technology" than Sandy Bridge. The firm stopped short of answering our questions about why the change was made and how the thermal transfer properties of the two materials compare. However, Intel claims the combination of the new interface material and Ivy's higher thermal density is responsible for the higher temperatures users are observing with overclocked CPUs.


http://techreport.com/discussions.x/22859

Re: IVB TIM on Retail CPUs?

Posted: Tue May 01, 2012 10:22 am
by flip-mode
ColeLT1 wrote:
I will have one in Wednesday (3570k). I de-capped half a dozen AMDs in the past, and IB looks even easier with no resistors? to hit on the top (see pic):
This was my socket939 3800x2 or 4800x2, can't tell:
Image

I have waterloop with an old V1 Dtek Fusion that will work for direct contact, if I de-lid I will do some stock clocked and overclocked capped vs uncapped results if anyone wants.

Dropping it in this rig ASAP:
http://imgur.com/a/tMWR8

That's some balls right there :o I wonder if I can do that to my X4 955...

Re: IVB TIM on Retail CPUs?

Posted: Tue May 01, 2012 12:16 pm
by DPete27
Seeing this Fudzilla article saying that retail IVB CPU's cant break 4.4GHz (seemingly even with extra voltage) I'm suddenly very hesitant to purchase one.

Re: IVB TIM on Retail CPUs?

Posted: Wed May 02, 2012 9:43 am
by ColeLT1
DPete27 wrote:
Seeing this Fudzilla article saying that retail IVB CPU's cant break 4.4GHz (seemingly even with extra voltage) I'm suddenly very hesitant to purchase one.


Interesting.

I installed windows on it last night and am installing drivers and updates via logmein ha. Hopefully get to play with it some tonight.

Right now I am seeing core temps around 28-32 idle and 50-55 under prime on stock clock.

Re: IVB TIM on Retail CPUs?

Posted: Thu May 03, 2012 10:45 am
by ColeLT1
So last night I hit the max voltage that I want to use for daily use (4.7ghz @ 1.3v), and the temps are not bad at all (60-70c). I am still debating on uncapping or not, if it is declared that IB is safe for daily use on water at 1.4v (but too hot) I will uncap to cool it off, but if 1.3v is it... then I don't see the point for my setup. I still am going to find the limit of my chip with 1.3v+ and post the data, but 1.3v / 4.7ghz at 70c is great.

Prime blend:
http://i.imgur.com/LvyFa.jpg

Re: IVB TIM on Retail CPUs?

Posted: Thu May 03, 2012 10:47 am
by Flying Fox
Running IBT will give you higher temps, you can make the decision then?

Even CoreTemp 1.0 RC still shows as ES? This is odd.

BTW, you know the Snipping Tool and PrtScn tricks to do screen caps? ;)

Re: IVB TIM on Retail CPUs?

Posted: Thu May 03, 2012 11:27 am
by ColeLT1
Flying Fox wrote:
Running IBT will give you higher temps, you can make the decision then?

Even CoreTemp 1.0 RC still shows as ES? This is odd.

BTW, you know the Snipping Tool and PrtScn tricks to do screen caps? ;)


Prime is fine for temps for me, since gaming will be like 20c lower, but I do like your style. I did notice the ES thing, agreed, odd. Snipping tool is great and all, but I was on my way to work and didn't have time :P I was nice enough to link (not embed) the crappy phone-camera picture.

The way the retention plate holds the chip to the mobo is to grab each side of the heat spreader, I wonder if this would not press hard enough on the cpu once this is gone making poor contact to the socket "pins" or even damage the chip? See here: http://file.ynet.com/2/1203/03/6850171.jpg look on the left and right of the spreader.

The conscious seems to be 1.325 - 1.35v is the most that I want to put to this chip for normal use.

Re: IVB TIM on Retail CPUs?

Posted: Thu May 03, 2012 12:23 pm
by Flying Fox
ColeLT1 wrote:
The way the retention plate holds the chip to the mobo is to grab each side of the heat spreader, I wonder if this would not press hard enough on the cpu once this is gone making poor contact to the socket "pins" or even damage the chip? See here: http://file.ynet.com/2/1203/03/6850171.jpg look on the left and right of the spreader.

I would not worry too much about that. Once the HSF is clamped down the contact should be even better? At least that's my experience with my Wolfdale and Lynnfield CPUs.

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 9:27 am
by ColeLT1
I tried to uncap, but could not get any of my razor or blades to get between the PCB and IHS, I have a feeling the reason we have only seen pictures is the chip was damaged removing the IHS. The AMDs I decapped before had a larger gap, where I could stick a razor blade in each corner and cut the glue. As I kept working the blade all I was doing was scratching up the PCB: http://imgur.com/a/3XlRF

I hit a wall at 4.9, could not get prime to start.

4.8ghz stable, but voltage too high:
http://i.imgur.com/q60XU.png

4.7ghz is going to be my sweet spot @ 1.3v

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 11:05 am
by Ryu Connor
ColeLT1 wrote:
I tried to uncap, but could not get any of my razor or blades to get between the PCB and IHS, I have a feeling the reason we have only seen pictures is the chip was damaged removing the IHS. The AMDs I decapped before had a larger gap, where I could stick a razor blade in each corner and cut the glue. As I kept working the blade all I was doing was scratching up the PCB: http://imgur.com/a/3XlRF


Heat gun to melt/loosen the glue?

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 1:01 pm
by Forge
Do not uncap. The IHS is now part of the CPU hold-down mechanism. If you decap, you will be recapping, or not using your CPU anymore. The CPU socket clamp provides NO downward pressure to the CPU at all with the IHS removed on LGA 1155.

This applies to LGA 1366, LGA 1156 as well. LGA 775 did things slightly differently, but involved putting pressure on the IHS as well.

LGA775 = works with IHS removed, but not well, barely any pressure.
LGA1155/1156/1366 = no pressure on CPU at all, intermittent/no contact on CPU pins.

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 1:04 pm
by DPete27
Looks like you're hitting the same ceiling as Geoff at TR and many other reviewers did. His temps were higher, but he was using air and you're liquid cooling.
http://techreport.com/articles.x/22833

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 1:18 pm
by ColeLT1
ColeLT1 wrote:
The way the retention plate holds the chip to the mobo is to grab each side of the heat spreader, I wonder if this would not press hard enough on the cpu once this is gone making poor contact to the socket "pins" or even damage the chip? See here: http://file.ynet.com/2/1203/03/6850171.jpg look on the left and right of the spreader.


Forge wrote:
Do not uncap. The IHS is now part of the CPU hold-down mechanism. If you decap, you will be recapping, or not using your CPU anymore. The CPU socket clamp provides NO downward pressure to the CPU at all with the IHS removed on LGA 1155.

This applies to LGA 1366, LGA 1156 as well. LGA 775 did things slightly differently, but involved putting pressure on the IHS as well.

LGA775 = works with IHS removed, but not well, barely any pressure.
LGA1155/1156/1366 = no pressure on CPU at all, intermittent/no contact on CPU pins.


TY for confirming.

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 1:34 pm
by Forge
ColeLT1 wrote:
TY for confirming.


No prob. If you've dumbly gone where few men have gone before, you can at least save other folks the trip.

I was actually prouder of one of my other projects, where I capped (not decapped) a T7500. There were special circumstances, though.

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 5:46 pm
by Flying Fox
With potential overclockers (and whoever dumb enough to follow their exploits without thinking much) freaking out over the high temps and this "I can't overclock this by 50% this sucks" attitude. There may be suckers going down that path.

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 5:54 pm
by Forge
Oh, there's potential value in decapping Ivy, but not the way everyone seems to be thinking... If I had an Ivy that would go far on few volts, but just seemed to fry as soon as I tried, I'd measure the crap out of the stock IHS, then get a new custom fitted IHS machined out of copper. It would either have the silicon core 0.001" above the copper level (ring-type shim), or have a very very thin copper top across the original IHS area (better IHS).

With a really nicely done ring-type shim, you can get most of the core protection of IHS, with the better thermal interface of direct contact. I'd be headed there. Instead I'll chase LGA 2011. I think I see part of why there is no Ivy Bridge E now.

Re: IVB TIM on Retail CPUs?

Posted: Fri May 04, 2012 7:24 pm
by Krogoth
Why can't overclockers just accept the fact that Ivy Bridge doesn't response well to overvolting?

I think they just got spoiled by how well Intel has managed to get a decent clockspeed headroom on every silicon release since Conroe. Ivy Bridge defies this trend.

Re: IVB TIM on Retail CPUs?

Posted: Sat May 05, 2012 12:13 pm
by ColeLT1
Krogoth wrote:
Why can't overclockers just accept the fact that Ivy Bridge doesn't response well to overvolting?

I think they just got spoiled by how well Intel has managed to get a decent clockspeed headroom on every silicon release since Conroe. Ivy Bridge defies this trend.


I am ecstatic about 4.7ghz, responded very well to voltage increases, 22nm is just not 1.4v capable. Plus it runs cooler than my nehalem.

My past 3 intel chips:
45nm c2q 2.66 -> 3.4
45nm i7 3.3 -> 4.1
22nm i5 3.8 -> 4.7

Re: IVB TIM on Retail CPUs?

Posted: Mon May 07, 2012 6:51 am
by flip-mode
Just ordered a 3570 K for a CAD machine at work. Going to go with the stock cooler. Plugging it into a p8z77m pro.

Re: IVB TIM on Retail CPUs?

Posted: Mon May 07, 2012 7:57 am
by DancinJack
flip-mode wrote:
Just ordered a 3570 K for a CAD machine at work. Going to go with the stock cooler. Plugging it into a p8z77m pro.


That's the CPU and mobo I'd pick if I were buying IB. Nice choices, if they were yours.

Re: IVB TIM on Retail CPUs?

Posted: Mon May 07, 2012 10:15 am
by JustAnEngineer
I ordered the same processor and motherboard Thursday night for a new build. They should be here Wednesday.

Re: IVB TIM on Retail CPUs?

Posted: Mon May 07, 2012 10:42 am
by flip-mode
JustAnEngineer wrote:
I ordered the same processor and motherboard Thursday night for a new build. They should be here Wednesday.

Excellent, we can compare notes! :P

Re: IVB TIM on Retail CPUs?

Posted: Mon May 07, 2012 11:28 am
by integer
JustAnEngineer wrote:
I ordered the same processor and motherboard Thursday night for a new build. They should be here Wednesday.

Upgrading from the i7-2600k or is this an entirely new build?

Re: IVB TIM on Retail CPUs?

Posted: Mon May 07, 2012 11:34 am
by flip-mode
New build. Upgrading a user from a Dell Optiplex 745 SFF Core 2 Duo E 8500, Radeon X1300 Pro, 8 GB, 80 GB SSD to 3570 K, Radeon 7750, 16 GB (2 * 8 GB), 120 GB SSD, and also putting in a drive for system image backups.

Re: IVB TIM on Retail CPUs?

Posted: Mon May 07, 2012 11:40 am
by Flying Fox
flip-mode wrote:
Just ordered a 3570 K for a CAD machine at work. Going to go with the stock cooler. Plugging it into a p8z77m pro.

Why not go with the "regular" H/P chipset and non-K CPU if this is for a work machine (not likely to be overclocked)?

Re: IVB TIM on Retail CPUs?

Posted: Mon May 07, 2012 11:54 am
by flip-mode
Flying Fox wrote:
flip-mode wrote:
Just ordered a 3570 K for a CAD machine at work. Going to go with the stock cooler. Plugging it into a p8z77m pro.

Why not go with the "regular" H/P chipset and non-K CPU if this is for a work machine (not likely to be overclocked)?

I have no good answer for that other than it just "feels wrong" to buy a locked CPU. It doesn't add much to the budget ($1500 total including software), so there's not a big incentive not to.