Welcome, Guest!

Here are some links you may find helpful

New Idea/Hypothesis for N64 Overclocking

JohnMLTX

New member
Original poster
Aug 18, 2020
1
2
3
Hi all! I'm new here, used to lurk on AG way, way back.

I'll get right to the point. In a discussion with my girlfriend on N64 overclocking and how it hasn't really gone very far in ages, I got to thinking about RAMBUS overclocking, and how that worked on the handful of computers that used it. Add in some research on the R4000 series CPUs, and I'm pretty certain I have a decent angle to approach overclocking in both methods with a new approach.

Basically, replace the clock signals on the board with a separate clock generator. We know the frequencies needed for the CPU, RCP, and RAM, where they feed in, and what values we get out of each component with a given signal. The problem that people have seen, specifically with memory OC, is that the ratios between the CPU, GPU, and memory are really tightly aligned, so things like oscillator swaps don't really work. But since we can just calculate clock speeds that keep the ratios intact, we just need those specific signals. And the N64 has loads of 3.3v rails to tap into.

My hypothesis is that, if we just bypass the traces for clock signals to each component and feed in the necessary signal for the desired output clocks, we can maintain the ratios while independently controlling each chip, rather than relying on the fixed oscillators and clock generators on the board.

First step is now procuring a console to rip apart , desolder the CPU, RCP, and MX8330MC clock generator, and build a dev board. Will update as I make progress.
 

Make a donation