Changing BP_FID
BP-FID is actually easier, if your cpu isn't locked. BP-FID, like FID, consists of 4 bits, where FID and BP-FID should be equal, otherwise the cpu won't work (unless BP_FID is lower than FID, in which case the BP_FID setting is used for some inexplicable reason -- which conveniently explains Anand's results with being able to underclock but not overclock). So we also have to change BP-FID in order to change the multiplier.
Every signalline of the BP-FID is connected (through the golden bridges on the cpu and a resistor) to either approx. 2.0V or 0V. Also, they are connected via a second golden bridge to pins on the cpu which is described by AMD (very sneaky) in the datasheet as "Not Connected". Next, this signal goes to the cpu that uses it to check if the signal from the FID will be accepted.
In order to ignore the golden bridges of the cpu, follow these easy steps: You create a jumperblock, with wires in the middle that go to the "Not Connected" pins, and on each side you put 0V and ~2V. With these jumpers you can "override" the cpu adjustments.
(the Dutch text can be easily understood when you know that "naar" means "to" in English).
At this moment we presumed that the internal connection between the BP-FID pins on the cpu and the L3 and L4 bridges didn't exist anymore. The only solution was to make a connection around the cpu, by soldering 4 wires at the resistors (the 2 rectangular blocks).
And here comes the magic trick: The connection between L3, L4 and the BP-FID pins still exists, only, it runs via the L1 bridges! Because AMD took this connection away, the BP_FID pins are marked as "Not Connected" in the datasheets. By closing the L1 bridges you can get full control over BP_FID through the mainboard, and that what it's all about!
With a mainboard that already has this function on-board, you are ready. If you don't have this then you still must make it yourself, like at FID, with the advantage that you don't have to pull connectors out of the socket. You find more about this later in the review.
Every signalline of the BP-FID is connected (through the golden bridges on the cpu and a resistor) to either approx. 2.0V or 0V. Also, they are connected via a second golden bridge to pins on the cpu which is described by AMD (very sneaky) in the datasheet as "Not Connected". Next, this signal goes to the cpu that uses it to check if the signal from the FID will be accepted.
In order to ignore the golden bridges of the cpu, follow these easy steps: You create a jumperblock, with wires in the middle that go to the "Not Connected" pins, and on each side you put 0V and ~2V. With these jumpers you can "override" the cpu adjustments.
![]() |
(the Dutch text can be easily understood when you know that "naar" means "to" in English).
At this moment we presumed that the internal connection between the BP-FID pins on the cpu and the L3 and L4 bridges didn't exist anymore. The only solution was to make a connection around the cpu, by soldering 4 wires at the resistors (the 2 rectangular blocks).
![]() |
And here comes the magic trick: The connection between L3, L4 and the BP-FID pins still exists, only, it runs via the L1 bridges! Because AMD took this connection away, the BP_FID pins are marked as "Not Connected" in the datasheets. By closing the L1 bridges you can get full control over BP_FID through the mainboard, and that what it's all about!
With a mainboard that already has this function on-board, you are ready. If you don't have this then you still must make it yourself, like at FID, with the advantage that you don't have to pull connectors out of the socket. You find more about this later in the review.
Next page (Motherboard modifications - 5/7)


