i7 4790k

[CreativeHiro]

Hello guys

I have the following problem. I bought on ebay a fully funktional i7 4790k to replace my i5 4460. after installing the i7 I saw that the temps were very high even under water so i decided to delid the i7. So first i isolated the resistors next to the die, then replaced intels tp with liquid metal. Then i glued everything with hight temperature silicon back together and installed the i7 again. Started the pc and the temps were fine and the i7 stable. A few days later my motherboard suddenly didnt detected the cpu anymore. So i took everything apart again. Delidded again. And found the problem one resistor was not isolated proberly. So isolating it again. Lm on it. Back in the pc. Same issue a few day later again. So i took everything apart again and thought f**k lm i will use normal non conductive tp again. So i cleaned of the entire cpu and just used tp. Everything back together. Yeah cpu wored again. One week later. Same issue again. So i said f**k it maybe its the mobo. So back in my old i5. It worked fine for a week. So it has to be the i7 wich is brocken. So i used a magnifier to take a close look on the i7 and couldnt find any damaged parts or left over lm. So i thought maybe baking it in the oven helps. After 30 min at 120°C the i7 worked again. 2 days later same issue. So im running out of ideas. Do you have any other ideas what could be wrong?

 

Thy for your help

 

Greetings CreativeHiro

Edit: totaly forgot after several(5 to 50) restarts the i7 is detected again

 

 

 

[CreativeHiro]

Edit:

System:

Mobo: asus z97-a

Grka: asus gtx 970 strix

Psu: 600w be quiet

Storage: 1tb ssd/ 4tb hdd

Cooling: custom loop with one 360 and one 240 radiator 40 mm thick.

[porina]

That oven temperature wouldn't be enough to reflow solder so probably didn't do anything.

 

Don't know what else it might be. 

[CreativeHiro]

5 minutes ago, porina said:

That oven temperature wouldn't be enough to reflow solder so probably didn't do anything.

 

Don't know what else it might be. 

Thy for the answer.

What would be the right temp then?

[Tad Bittoomuch]

Was your motherboard exhibiting this behaviour after switching to the 4790k but before you delidded it?  It's a rather odd that you you progressed straight to delidding after discovering that the temperatures are a bit high - I would have looked for a potential different cause, such as improperly positioned/wrong mounting pressure of the waterblock, or a defective or blocked/obstructed loop or pump.

 

The 4790k itself is a rather hot CPU.  I used to have one and had a 240mm AIO and temps would reach mid to high 70s/low 80s under sustained load.

 

You probably know this, but this is what the 4790k looks like when delidded:

There are two rows of resistors next to the die - I am wondering if, during the process of delidding you accidentally managed to slice one or two of them off.  This could be causing your CPU to not be detected by your motherboard (Is there an error message?).

 

The other possible cause could be that there's too little or too much pressure on the die itself.  This is why delidding your CPU voids the warranty, because the silicon adhesive that attaches the heatspreader to the substrate contributes a non-zero value to the CPU package thickness.  It's possible that the slight change in CPU thickness contributed to the errors you are experiencing with your motherboard.

 

I'd suggest removing the silicon adhesive and simply resting the IHS on top of the substrate, then closing the retention bracket while the PC is horizontal.  Try reseating the waterblock and maybe investigate any possible blockages in your cooling loop.

 

Hope that helps  

[CreativeHiro]

7 minutes ago, Tad Bittoomuch said:

Was your motherboard exhibiting this behaviour after switching to the 4790k but before you delidded it?  It's a rather odd that you you progressed straight to delidding after discovering that the temperatures are a bit high - I would have looked for a potential different cause, such as improperly positioned/wrong mounting pressure of the waterblock, or a defective or blocked/obstructed loop or pump.

 

The 4790k itself is a rather hot CPU.  I used to have one and had a 240mm AIO and temps would reach mid to high 70s/low 80s under sustained load.

 

You probably know this, but this is what the 4790k looks like when delidded:

There are two rows of resistors next to the die - I am wondering if, during the process of delidding you accidentally managed to slice one or two of them off.  This could be causing your CPU to not be detected by your motherboard (Is there an error message?).

 

The other possible cause could be that there's too little or too much pressure on the die itself.  This is why delidding your CPU voids the warranty, because the silicon adhesive that attaches the heatspreader to the substrate contributes a non-zero value to the CPU package thickness.  It's possible that the slight change in CPU thickness contributed to the errors you are experiencing with your motherboard.

 

I'd suggest removing the silicon adhesive and simply resting the IHS on top of the substrate, then closing the retention bracket while the PC is horizontal.  Try reseating the waterblock and maybe investigate any possible blockages in your cooling loop.

 

Hope that helps  

Thy for your detailed answer.

These two rows of resisors are not demeged or knocked of i did check that multiple times. Also the loop should be fine. I only have a clear premix in it.  That there can be to much mounting pressure due to the glue is new for me. So I will try if this helps.

[porina]

24 minutes ago, CreativeHiro said:

What would be the right temp then?

I don't know the typical number without looking it up, other than I know it is above 120C for old lead based solders. Newer lead-free ones melt at even higher temperatures.

 

I'm not sure this is a meaningful route to go down. The oven trick is basically to melt the solder and allow it to reform, hopefully fixing any breakages that might not be visible. On a CPU, I'm not sure what they use to bond the die to the substrate, but the surface mount components do look like they may be soldered. If you have high enough magnification to see them you can at least do a good visual check of them first. 

3 minutes ago, CreativeHiro said:

That there can be to much mounting pressure due to the glue is new for me. So I will try if this helps.

I forgot about that, in part this is Haswell, and the mounting pressure I only recall reported on Skylake and newer as they used a thinner PCB for those which could bend under high pressure. Still, it is worth trying to alter the pressure to see if it helps.

[CreativeHiro]

33 minutes ago, porina said:

I don't know the typical number without looking it up, other than I know it is above 120C for old lead based solders. Newer lead-free ones melt at even higher temperatures.

 

I'm not sure this is a meaningful route to go down. The oven trick is basically to melt the solder and allow it to reform, hopefully fixing any breakages that might not be visible. On a CPU, I'm not sure what they use to bond the die to the substrate, but the surface mount components do look like they may be soldered. If you have high enough magnification to see them you can at least do a good visual check of them first. 

I forgot about that, in part this is Haswell, and the mounting pressure I only recall reported on Skylake and newer as they used a thinner PCB for those which could bend under high pressure. Still, it is worth trying to alter the pressure to see if it helps.

So i already took my system apart again. And made a picture of the 2 rows of resisors. They look not damaged in my my eyes. But the mounting pressure seems a good reason. Because when i testet my cpu without glue between delid and reassembeling it only booted when the heatspreader was perfectly centered.

[Tad Bittoomuch]

9 hours ago, CreativeHiro said:

when i testet my cpu without glue between delid and reassembeling it only booted when the heatspreader was perfectly centered

Did you experience the motherboard not detecting the CPU in this configuration?  Having the IHS held in place by the socket should be fine for normal operation, but I'd be interested to see if you start seeing the error after a few days.

 

My other questions are:

• Is there an error message when the CPU is not detected, i.e. what happens on the display?
• Does this occur when logged into Windows or is it only when you turn on the computer?

[CreativeHiro]

10 hours ago, Tad Bittoomuch said:

Did you experience the motherboard not detecting the CPU in this configuration?  Having the IHS held in place by the socket should be fine for normal operation, but I'd be interested to see if you start seeing the error after a few days.

 

My other questions are:

• Is there an error message when the CPU is not detected, i.e. what happens on the display?
• Does this occur when logged into Windows or is it only when you turn on the computer?

Yes thats the strange part. When i let the IHS held in place by the socket the cpu only is detectet when the IHS is perfectly centered. Which makes sense due to the mounting pressure the cpu needs. But when i turn it off again wait a few min and want to turn it back on the cpu is no longer detected. But sometimes it works for a few days and then suddenly the cpu is not detected anymore when i want to start the pc. Also this issue only occures when i turn on the pc. As soon it is running it runs stable at 4.5 ghz all core running stress tests.

An error Massage is not showing up. I only have these indicator leds which shine red when something is wrong. And the one under the cpu shines when the cpu is not detected. At least the manual says that there is no cpu detected when this led shines.

[ZzLy]

23 hours ago, CreativeHiro said:

Hello guys

I have the following problem. I bought on ebay a fully funktional i7 4790k to replace my i5 4460. after installing the i7 I saw that the temps were very high even under water so i decided to delid the i7. So first i isolated the resistors next to the die, then replaced intels tp with liquid metal. Then i glued everything with hight temperature silicon back together and installed the i7 again. Started the pc and the temps were fine and the i7 stable. A few days later my motherboard suddenly didnt detected the cpu anymore. So i took everything apart again. Delidded again. And found the problem one resistor was not isolated proberly. So isolating it again. Lm on it. Back in the pc. Same issue a few day later again. So i took everything apart again and thought f**k lm i will use normal non conductive tp again. So i cleaned of the entire cpu and just used tp. Everything back together. Yeah cpu wored again. One week later. Same issue again. So i said f**k it maybe its the mobo. So back in my old i5. It worked fine for a week. So it has to be the i7 wich is brocken. So i used a magnifier to take a close look on the i7 and couldnt find any damaged parts or left over lm. So i thought maybe baking it in the oven helps. After 30 min at 120°C the i7 worked again. 2 days later same issue. So im running out of ideas. Do you have any other ideas what could be wrong?

 

Thy for your help

 

Greetings CreativeHiro

Edit: totaly forgot after several(5 to 50) restarts the i7 is detected again

 

 

 

120°C is not enought to do anything. Solder melts at around 250°C so you need to do at least that.

[Tad Bittoomuch]

If I understand correctly - this only happens when you turn on the PC.

 

If you can get it to boot to Windows, does the CPU 'stop' being detected?  I.e. how long can Windows run for without an error related to your motherboard/cpu problem?

 

Are you getting the same issues with your old CPU?

[CreativeHiro]

18 hours ago, Tad Bittoomuch said:

If I understand correctly - this only happens when you turn on the PC.

 

If you can get it to boot to Windows, does the CPU 'stop' being detected?  I.e. how long can Windows run for without an error related to your motherboard/cpu problem?

 

Are you getting the same issues with your old CPU?

Yes this only happens when i turn on the pc.

I can run windows for 8 hours. Didnt tested it longer yet. In windows it hat normal temps and boostclocks and runs absolutly stable. Even at a 30 min cinebench r15 loop.

With my old cpu i dont have any Problem. Its starting every time normaly. 

[Cavalry Canuck]

Really stupid question, but here it goes...

 

Did you flip it over to make sure nothing got on the pin contacts? Same question for the pins on the mobo.

[CreativeHiro]

3 hours ago, Cavalry Canuck said:

Really stupid question, but here it goes...

 

Did you flip it over to make sure nothing got on the pin contacts? Same question for the pins on the mobo.

Yes i did