Socket 462 Boards + Vcore testing

[Solder Boy]

Hi there I was reading the FAQ about recapping your boards and somewhere it mentioned checking the vCore with a meter before inserting a CPU chip... how the heck do you do that?

Over the last year I have had countless decent AMD chips go south as soon as I put them in an untested used board and of course major swear words are uttered....

I get a lot of Socket 462 boards and it would be nice knowing how to do this test.

Thanks in advance!

[linuxguru]

Re: Socket 462 Boards + Vcore testing

Not all Socket-A boards can support all Socket-A CPUs. In particular, most boards support Thunderbird and Palomino, but not Thoroughbred-(A/B), Applebred or Barton.

In combination with marginal caps and higher Vcore, it's very easy to blow a Barton or Thoroughbred on an older Socket-A board, even if it seems to get past POST initially.

[gg1978]

Re: Socket 462 Boards + Vcore testing

Hi there I was reading the FAQ about recapping your boards and somewhere it mentioned checking the vCore with a meter before inserting a CPU chip... how the heck do you do that?

Over the last year I have had countless decent AMD chips go south as soon as I put them in an untested used board and of course major swear words are uttered....

I get a lot of Socket 462 boards and it would be nice knowing how to do this test.

Thanks in advance!

The thing is, if you don't have a CPU mounted in the socket, chances are that the VRM will be set to be disabled, i.e. output 0 VDC.. The key is that the CPU itself tells the VRM what Vcore it wants through a set of 5 pins called VID lines.. The VID lines are either left open, or tied to ground internally to the CPU package.. If the pin is open, it's a "1", while a short is a "0".. The default state of the pins with no CPU in socket is "11111", which will result in the VRM staying off.. So you need a CPU to set your Vcore output.. I mean you can pull the pins down manually, and check Vcore that way..

I have a motherboard that i brought back from the dead because it's VID inputs either quit working or were disconnected by somebody trying to voltmod the board.. I brought it back to life by hotwiring the VID pins appropriately for the CPU's i was putting into it..

Technically it's safe to run a VRM with no CPU in the socket, so you can probe Vcore, but it won't be under a typical CPU load..