Building A Better NES


The first model of the NES wasn’t all that great; just ask any one of the millions of six-year-olds who independently discovered blowing on a cartridge made it work. The second NES hardware revision, the top loader, was better but only had RF video output. These are the only two pieces of hardware that can play every single NES game, and even with years of hacking NES-on-a-chip devices, there’s still much to be desired.

[low_budget] over on the AtariAge forum decided he’d had enough of these hardware compromises and decided to build the first new NES hardware revision in 20 years. It’s got all the best features from both of its predecessors and a few new features not seen on any existing NES. There’s support for composite and RGB video generators, new and better amplifiers for the audio, no lockout chip, and a top loading cartridge slot to prevent bent pins on the 72 pin connector.

While [low_budget]’s prototype works, it only does so by salvaging the CPU and PPU from a working NES. There’s still much work to be done on the prototype, but even if we’ll have to destroy our beloved NES, we’d love to get our hands on one of these improved consoles.

41 thoughts on “Building A Better NES

        1. The NES has a custom video chip and memory mapper chip that have yet to be 100% cloned. Most clones go with something that works for most games. However, some games use more exotic memory mapping schemes to cram in more data with the cart, meaning clones can’t read the game. Other games use video modes supported poorly or not all on clone chips resulting in a garbled mess for a display.

        2. Here’s a bunch of the random bugs that I can think of in Famiclones off the top of my head:
          * Some get the 25% and 50% duty pulse wave sounds backwards
          * Some force the on-screen layout (“nametable”) memory to always be from the internal memory, and so conflict with games that use more (from whatever source, e.g. Gauntlet, Castlevania 3, After Burner)
          * Some don’t do duplicate fetches from memory when rendering because they’re usually identical; any game that uses the MMC5’s reduced-size attributes won’t work there. (Koei’s military strategy games)

        3. Same reason why emus don’t work. Here a funny fact that kinda sucks. The gameboy color is a gameboy clone but done by Nintendo themselves. It too has some things about it I noticed don’t act like gameboy. I believe the instruction book says that all GB games may not work. I know donkey kong land sometimes bugs out with garbled sprites.

    1. Another reason not to use NES or Famicom clone. They all use one big blob to cover cloned CPU, cloned PPU, and some logic chips so it’s hard to just remove it from the clone board and install in the new NES board.

      nerd tip: PPU handles all the graphic stuff from the cart to the screen.

  1. Love the idea, but why make it top-loading? Surely you could spin a board that you could mount vertically in the original NES case, and still make use of the old horizontal cartridge flap?

    Note that I’m not suggesting to keep with the old spring-loaded tray that bends the 72-pin connector, I’m just saying take your top-loading board and put it on its side.

    1. The problem with front loading a cart into a friction connection is that you end up shoving the entire console back when inserting a cart, unless the console is weighted or you hold onto it. Probably why Nintendo went with the insert and push down to lock system with the first NES (Little force going in, then downward force where it isn’t going to move).

    2. No.

      Just no. I do like the toaster look, the design however is the main reason so many people has problem getting NES carts to work after many years. There is no friction when inserting or removing so the contact never gets scraped and crud builds up more than other cart based systems.

      Also 90 degree connector that don’t use zero force insertion design like NES are prone to failing from the force of being pushed in and pulled out. I’ve seen many TI-99 and C64 computers with broken solder or cracked PCB because they thought that front loading cart design was better than top loading cart design.

      If you still want horizontal loading, buy one and hack it to suit your need. I plan to get one, hack it to 90 degree and use PowerPak or the upcoming Everdrive for NES (amicom version is out though if you have Famicom or will spend another $20 for Famicom to NES adapter) so I won’t have to change game at all and keep it all in a box with no visible slot. I’m thinking maybe making one of the iCade I picked up from Khol’s for $14 into mini NES arcade

      1. The original NES cartridge loading design is an engineering and design FAIL of the same epic FAILness as the Atari 5200 controllers.

        What made Nintendo’s engineers think that a connector which just pressed the contacts onto the PCB’s traces would be reliable in the face of house dust, pet hair and dander, human dander, tobacco smoke residue, air fresheners and every other bit of gunge and grunge and corrosive thing common in a typical home where the consoles would be used?

        All previous cartridge software systems *didn’t* have contact issues because they used a sliding slot connection which cut through any crud and corrosion every time a cartridge was inserted.

        I took apart and cleaned many NES consoles and to everyone who owned one I said “Get one of those cleaner devices that inserts like a cartridge and has the pad you slide in and out to scrub the connector so it stays working.” Not one NES owner would heed that advice. They’d use it until it got dirty and non-functional again then bitch about it.

        “Did you get the cleaner device I told you to get?”

      2. I think you missed my point. I’m fully aware of the design flaw with the spring-loaded cartridge design from the original NES (crud, bent pins, and so on). I’m suggesting that you take the board the guy already made, with the “vertical” cartridge slot (not a 90 degree version), and mount the board on its edge.

    1. Unless you want the HDMI by using a composite-to-HDMI converter (and if you do, what’s wrong with you?) or RGBS-to-HDMI converter and find a source of the ever-increasingly-rare RP2C03s, you can’t get that. (Yet. google: hdmi nes bunnyboy )

      Even after you do that, the NES produces 8-ish bit audio at a 1.8MHz sample rate, and converting that to something that works correctly over HDMI without aliasing or other sampling noise will either require complete emulation of the NES’s audio core or a video-grade ADC to accurately sample the NES’s pins and downsample it.

        1. There is information content in the entire bottom 1.8MHz/2 (nyquist) generated by the NES. Sigma-delta ADCs WILL be confused by it. There’s nothing to disagree with: if you want it to sound right, you will need a brick-wall lowpass at ~20kHz, and that’s best done in discrete time. And for that… you’ll need to sample at >1.8MHz, which is above normal audio ADCs.

          Next time, try pointing out the flaw rather than just saying “you don’t understand X’s contributions to the field”!one!1!!. Piecing together the actual content of the APU using Shannon’s entropy limits is an even worse plan than full emulation of the APU.

          1. The solution is simple: use an analog anti-aliasing filter to kill all inaudible frequencies. Then sample at say 100-200 kHz. You don’t even need that much quantization depth, its a fucking NES.

  2. Nice project. But I don’t understand why he says that the PPU costs $200: you can get a used NES on eBay for 40$ or so, and presumably salvage a PPU from there – so where (and why?) is he getting these PPU’s for $200? Don’t get me wrong, it’s a nice system, but it would be nicer if it wouldn’t cost $500 to buy :)

    1. There are 2 different versions of PPU. The one in NES deck can output video only in composite. The $200 PPU are the second version that are found in Nintendo arcade machines such as PlayChoice-10. The expensive (and rare) PPU can output in RGB format, which would allow people to get cleaner picture than composite. RGB can be converted to component or S-Video and still look better than composite.

  3. I’ve salvaged about 6 NES from tag sales for prices countable in quarteres. About half don’t work (no power in one, bad CPU in another) so I may have to scrap those and combine the working parts into one of these. It just looks so pretty and functional.

  4. Fact check fail hackaday, first version NES(famicom) is a toploader, second version (us / europe is the gray slotin unit) third version is also a toploader, the Eu version?(or us not sure but one of them) do have a RGB connector on the side

    1. You’re assuming this project is intended to be sold or made available worldwide. The person likely lives in USA or Canada, where we got only 2 versions of NES. The front loading is very common around here and has high failure rate due to bad connector, The top loading version is rare and still goes for around $100 but does not offer composite output without a hack.

      Eu version is not designed to play NTSC games and I believe some games do not work right (50hz vs 60hz causing variation on game speed and sound)

      1. Im fully aware of the regional issues due to the multiplyer within the chips and what consoles can be obtained where only pointing out that the fact check is kinda weak, still love the idea that someone had the time to make a new revision. Actually bought myself 1 famicom yesterday in a hongkong flea market :)

  5. Actually that picture is a little old, it’s my first test system.
    I have since made a new PCB version v1.1 and assemble them to use a composite or RGB PPU, since there are differences in the components needed for each.
    The sound has each sound channel separately amplified and output as stereo.
    Auxiliary sound from a PowerPak or Famicom game (like Castlevania 3) can be mixed in as well.
    I mount the PCB in the AG-85 from Polycase and it comes in 3 colors. It seems not everyone is a fan of the case, so that’s why I have just the PCB for sale too.

    Oh and yes, it can be mounted in a toaster. Is that why people want a right-angle cart slot? Problem is, nobody makes one.

    I don’t like removing original Nintendo chips, but there currently isn’t a better way to make a NES clone in my opinion. There was supposedly a project to reproduce the PPU on a FPGA, but who knows if this will be completed.

    1. How can we get into contact with you to maybe get a hold of the board you use? I have an NES I would LOVE to get working 100 percent of the time! Also, what connector do you use for the carts?

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.