nForce3: A Locked PCI Bus?

Published by

Lost Circuits have thrown up their latest article which investigates Nvidia's nForce3 and does the said board have a locked PCI bus?

ONE OF THE DIFFERENCES between the two main contenders in the Athlon64 arena, namely the nForce3-150 and the VIA K8T800 is that the VIA chipset derives its PCI bus by means of a 1/6 divider of the external CPU clock whereas rumor has it that the nForce3 chipset, similar to the nForce2 chipset has a frequency locked PCI bus. In turn, this is often used as an argument to explain the better overclockability of the nForce3 chipset.

There are different ways to prove one way or the other, one way is the use of a Geiger card with the caveat that this PCI card may not function correctly since it is used internally within the system. Honestly, we don't believe that this is a strong possibility. However, another, fool proof way to find out is the use of a high frequency oscilloscope to probe the clock input signal directly at the PCI slot.

We used the Shuttle AN50R nForce3 -150 based board to monitor the PCI input clock at different frequencies using a HewlettPackard HP54512B High Spee3d oscilloscope. Here is what we found:

nForce3: A locked PCI bus?