CPU issue: what do you make of this?

PMc

PMc

There are two (or three) brown stains in the right part. They cannot be removed with alcohol.
 

Attachments

  • IMG_20240623_215916.jpg
    IMG_20240623_215916.jpg
    1.1 MB · Views: 46
I never chased it down but I bet it is the power pins that are discolored.

The power&heat probably acts as an annealer to the pin pads over long term.

You don't see this on new CPU or low hour ones.

Think of it this way. You have a 140Watt CPU with power on how many pinpads?

What looks like a stain is a blob of discoloration from heat. The Core has unequal heat.

That's my best guess.
 
Phishfry said:
I never chased it down but I bet it is the power pins that are discolored.

The power&heat probably acts as an annealer over long term.
Click to expand...
Power pins would make sense - somehow the power must go in. But then, mine doesn't have these stains. Okay, it doesn't run full power all the time.

The problem is, that piece did work fine - for almost a day. Then I did a slight load test, and after some 10 minutes at 50%, then increasing to 75%, it now does no longer run. (Temperature was all more than 20 K headroom to Tjmax.)

So, You're a hardware man, I know that. :) You think this gold stuff can oxidize?
 
I forgot it could be gold pads. I was assuming bronze alloy.
Many times when plating you build up layer for bonding to base material like brass/bronze.
Flash with bronze then plate with gold. Less gold needed.

Can gold oxidize. I think so...

I do think its heat issue not oxidation per say. Maybe not from pins but core above?

It would be interesting to see if discoloration patterns are similar.
 
.
Phishfry said:
Question: Is it a high TDP CPU?
Click to expand...
105 Watt.

Phishfry said:
I forgot it could be gold pads. I was assuming bronze alloy.
Click to expand...
You do likely know more about such kinds of stuff than me. I was thinking it should be gold and it should not be possible to oxidize. But then there is never pure gold used (not even in my teeth ;) ), and the other components of the alloy may behave differently.

Phishfry said:
Many times when plating you build up layer for bonding to base material like brass/bronze.
Flash with bronze then plate with gold. Less gold needed.
Click to expand...
Yepp, that makes sense.

But the important info is, this seems to happen more often, and one cannot conclude from such stains to an inherent defect or previous mishandling.
 
After reading that new CPU's have this stain I have adjusted my opinion.

This is probably some acid cleaner residue used when removing the pin pad mask.

That could permanently discolor pin pads.

Can you imagine the fineness of a LGA3647 cpu mask. A we are just talking the external interconnects.
It would have to use a solder mask right?
Then you got to remove it. Man or machine? Maybe not metal but etched away?
Just thinking about the packaging part is amazing....
 
You must log in or register to reply here.
Back
Top