Slashdot Mirror
Nintendo NES Overclocking Guide
Deven "Epicenter" Gallo writes "I've perfected a process by which to overclock the Nintendo Entertainment System (NES) to run games smoother without slowdown. The NES CPU normally runs at 1.79 MHz, I've reached a stable maximum of 4.2 MHz, about a 230% overclock. The games do not run faster than they should, the CPU never overheats, and most games are perfect up to 3.3 MHz!" Here's the guide on how to perform the modification, along with photos and demonstration videos
I remember getting Megaman for a christmas present many years ago and it had slowdown at some points when there were alot of enemies on screen. Here aa review on the game that mentions the slowdown in it: http://www.nesplayer.com/reviews/mm2r.htm
http://www.livejournal.com/users/metricmusic
Most games were smooth most of the time. But there were a few that, during more intense parts (relatively speaking), tended to slow down. SMB3, as demonstrated in the video on the site, was one of those. With enough objects moving around the screen it did get a bit choppy.
It certainly wasn't a big enough problem to affect the games' playability. It was noticeable, however, on the rare occasions that it happened.
...there would be some very significant side effects to such modifications:
1. NES audio is generated within the 'RP2A03G' (CPU) chip and is based on clock cycles, so doubling the CPU clock will cause the audio to go up an octave (assuming it even runs). The site mentioned in the article actually pointed this out, so it looks like it's legitimate.
2. Games which use cycle-timed code will no longer work properly - Battletoads is the first that comes to mind.
3. Some NES cartridges only used 250ns PRG ROM chips, which is only good up to 2MHz; go any higher and the game may not run at all.
* Q
P.S. If you don't get this note, let me know and I'll write you another.
The time works fine-- it's SLOWING DOWN below the PROPER speed when it lags. The overclocking stops it from slowing down-- it's going the speed the programmers intended. As for temperature, I estimate both to be pretty much indistinguishable, but around 75-85F. Not exactly a heat emergency.
Technically:
All copper lines on a PCB are traces. This includes power and ground lines, excepting large areas of copper, which are usually called planes. Also, there are things called "ground shields", which are actually not connected to ground - they're electrically isolated continuous bits of copper used to provide electromagnetic shielding.
Anyway.
Traces are the copper lines on a PCB - buses are composed of multiple traces carrying a collection of related signals (for example, an address bus is N traces, each carrying one bit of the N bit address.)
---
Mod me down, you fucking twits. Go ahead. I dare you.
(I read with sigs off.)
In my experience, loading program/data to the SPC-700 chip can be pretty slow. And all the RAM in the system allows for more compression, so more time for unpacking..
Other than that, SNES loading times should theoritically be shorter than on NES (if the game was coded similarly).
I'm surprised there wasn't a Zelda demo on the site. Whenever there was a room of those Jumping guys that turned into bats when you stab them, the system would lag like hell if you made too many bats. Also if I remember correctly, those pancake guys (?) that ate your shields never did much for the framerate either. ...Finally, a better solution to killing those guys than the Magic Sword
Partial Credit: The Engineer's Best friend
"Well, the bridge didn't fall all the way down!"
NES carts used special ASICs so the NES could address more memory, and added various other features.
The first NES carts just had 2 ROMs and the 'lockout chip' (which actually is a little CPU, running in sync with the other lockout chip in the system to thwart unlicensed game publishers).
Yes. There used to be PC clones that had a "turbo" switch on - that changed the CPU clock from whatever they were normally clocked at to 4.77MHz so that software that required precise timing loops would work correctly (step forward Sopwith...) on newer machines.
If you find that your CPU doesn't work at above 6MHz, try swapping it for an NEC V20 or V30 (can't remember offhand which) - this was faster even for a given clock speed, went to around 10MHz, and 8086 compatible into the bargain.
I used to have (probably still do somewhere) a Compaq 286 with a 6MHz 287 fitted instead of the 8MHz part. Worked just fine.
Actually, Nintendo used to offer a lifetime warrenty. You could get them to fix your NES if it overheated on its own, for example, due faulty workmanship. I read it the other day in the booklet I got with my SNES. Since the Gamecube and discs read by lasers, however, they've limited this lifetime warrenty to a year.
Actually, you're completely off on this one. It has nothing to do with "moving corrosion"...it has everything to do with moisture.
The reason that blowing the carts worked (and still does for the most part) is that you're applying moisture to corroded terminals. Moisture = connectivity. That's why after you blow in a cart and put it in your system, sometimes it works for 5 minutes and then the game starts blinking. The moisture has evaporated from the terminals.
Bad thing about blowing in the carts (there had to be a bad thing) is the fact that, by applying more moisture to the cart, you expedite the corrosion. It works better now, but down the line it becomes even harder to get them to work.
You can scrub away all you like on most carts and even buy a new pin connector, but your games are gonna all run like crap. It's not the system, it's the carts.
Now if you REALLY want to try your best to get your carts to work, the real way of refurbing them is to get a nintendo bit for your screwdriver, take off the cover, and scrub (and scrub, and scrub) at the connectors with a white eraser. It's the only thing I've found that will do the trick. Also, if you really want to use a cleaner, an anomia based cleaner (like Windex) does a much better job than rubbing alcohol.
Wise men say, "Forgiveness is divine, but never pay full price for late pizza."
Forgive me for linking to the file directly, but here's a direct link to a demonstration video showing the overclocking differences in Super Mario Bros. 3.
;-)
Instead of observing the obvious improvements in fluid animation (and gameplay), listen to the audio differences in both before and after overclocking. The original is exactly how I remember SMB3 to sound, while the overclocked version sounds kind of whacked.
To me, the audio from these classic games is JUST AS important as the video, so I won't be overclocking until a better method is found that won't screw up audio.... but I'd imagine the bulk of people who still have *working* NES units, wouldn't want to mess around with their precious vintage systems anyways
Karma police, arrest this man, he talks in maths....
Size of sprites has no impact on performance, what matters is how much code to handle those objects is executed. You only have a 1 MHz 6502 processor, so if you can't finish handling all objects in 1/60th of a second, you get slowdown.