Broke a pin off my CPU - still works fine?

W

WCES Ryan

Gawd
Joined
Mar 28, 2005
Messages
723
At the risk of terrible embarassment, I am just too curious and have to post this.

About a week ago, my gf's computer died. I determined it was the motherboard, so I bought a replacement, no big deal. The fun came when I was putting it back together, the CPU wouldn't fit in the slot, and I saw that there was a couple of bent pins. I was able to straighten most of them, but after a few tries trying to get one pin straightened out, it broke off. I figured what the heck, I'd try it anyway, and was surprised to find that the computer still works like normal.

So that begs the question: what was this pin for if the CPU still works after it was broken off? The chip is an AMD quad core, Phenom 9550 I believe. The pin was on the outer edge, close to the corner where the little arrow is that you use to orientate the CPU when you put it in the slot.

Thoughts?
 
Look up a pin diagram for the chip and figure out what it was. If it was something fairly insignificant like a grounding pin, then it doesn't really matter. It really depends which one you break off.
 
The AM2 pinout documentation is not available, however it is very similar to socket 939/940. There's three types of pins on the edge near that corner: memory addressing, hypertransport control, and voltage. Since your cpu still works without errors, it's probably not an addressing pin. So that means either you snapped off a voltage supply pin (overclocking will be less stable), or a hypertransport control pin (one less lane used). A hypertransport error would be hard to spot, but you might see slower PCIe transfers or a slower max disk transfer rate.
 
Could be a grounding pin. I don't remember the layout off-hand, but processors have a LOT of redundancy in their grounding circuits.

And just a personal pet peeve of mine, "orient." Not "orientate." Sorry.
 
mavalpha said:
Could be a grounding pin. I don't remember the layout off-hand, but processors have a LOT of redundancy in their grounding circuits.

And just a personal pet peeve of mine, "orient." Not "orientate." Sorry.
Click to expand...

That's the thought that went through my head when it happened, I just figured eh it's probably just a ground or something, no big deal. Maybe I was right.

My humble apologies, I thought orient at first too, but remember having heard the word orientate used in conversations before.
 
WCES Ryan said:
My humble apologies, I thought orient at first too, but remember having heard the word orientate used in conversations before.
Click to expand...
Ah, so you heard people conversating about orientating... I see now. :p

Glad it works, and yes- the odds are heavily in your favour that it was just a ground. I'm not sure offhand, but I'd venture a guess that over 25% of the pins in any given socket are there for ground. Considering that processors can get to 150 watts with less than 1.5 volts, we're talking about over a hundred amps flowing between the socket and the CPU- that'd be a lot of current for just a few pins.
 
a lot of redundancy pins on processors (either amd or intel). can be ground, vcc, ddr, etc.
used to retain signal integrity as well as address contact issues (on sockets)
 
I've never snapped a pin off a CPU, but I bent a couple of the contact pins in the 775 socket before. Those things are very delicate. I'm almost certain they weren't making good contact afterwards. But, to my surprize, the machines still functioned normally. *shrug*

I wouldn't sweat it.
 
I did this and ruined a brand new Athlon X2 4800 (a $1000 cpu at the time). Good luck and I hope you have a different outcome
 
You must log in or register to reply here.
Back
Top