Making of "Reverse emulating the NES..."

Science

You say we don't want 40 minute videos but... I think we do. Any chance you ever finish up this project completely?

Commenting in the hope you make many more 40+ minute videos.

I would absolutely love to see a part two of reverse emulation and thus turn the Nintendo into a husk.

Here's a thing I encountered with a raspberry pi, I did a project thing... With a pi, and it worked great on my desk, but when I mounted it inside a cabinet with control gear (contactors) it'd false trigger, the RF noise from the contactors coming in and out and the sparks therein were enough to trigger the inputs.

Now yea there's ways to stop that in hardware, and the debounce thing... I tried those and they were kinda lackluster, eventually I had to fix it in code, the contactor spikes were very very short, so I ignored short triggers.

I guess this is testament to why you should use industrially hardened and shielded PLC controllers, but they ain't cheap and for a thing you're not sure is gonna work or not... Eh, big investment.

Ultimately I ended up using optocouplers, debounce (even though they shouldn't bounce) and ignored everything less than half a second long.

That's hacky, maybe things like that are supposed to be... Low level programming itself is really hacky and janky anyway...

Regardless, how do you harden a pi against this? Didn't just happen to the pi either, I had a digital plug in timer... That did it too, and it was just in the same room as some contactors not in their cabinet it was in its own... Seemingly their RF spike is so harsh that just being in the same room is enough, and this thing isn't an Ali express or an amazon thing... But it would still false trigger.

Tl;dr: contactors are noisy as hell and they will trigger micro controllers, and I don't really know how to remove that entirely and in an elegant way.

Edit: also, I wonder if it's possible to take the output from a genesis/megadrive, and force it through the NES, thus kinda... Playing sonic on a NES.

is there a way to use a program to optimize the circuit board using neural networks? What would be more effective, moving elements themselves or moving the connections between them? How do current optimization programs work?

So, something Iʼve been wondering lately: could you on either the NES or the SNES force it to send data during V-Blank? You usually wouldnʼt want to but some information is designed for that; for example, in theory you could closed caption your games with line21 that way.

40 conductor ide is also not straight through. Pin 28 (cable select or spindle sync) is inconsistent, and pin 34 is sometimes only between drive 0 and 1, not to the controller. Because nothing can be simple and easy.

This isn’t meant as an “um actually” moment, this is me having spent a day of troubleshooting to learn this and trying to keep someone in the comments from that

imagine reverse emulating a gamecube or ps1 game on atari 2600

Haha you butchered a Zelda cart?! Surely there are some cheap games noone that plays that you could have used

r u gay or pedo

Able to share your circuit design, outside of painfully copying the diagram from your video? I think this would make for a neat project for others by being able to directly order the boards after you went through all the first steps.
Amazing work.

Bald

not sure to understand and instead of asking chatgpt3 to resume me your video, can you do a NES "upgraded" 16 bits cartrigue game with your technology and place it into a normal NES and it would work and play?

Ideas for further improvements:
Audio
Use the DPCM playback function of the APU.
Low quality, but still better than should be possible.
Samples normally limited in size, but faking an endless sample by making a 1-wide sample and changing it every time it is read should work.
Warning: Controller and VRAM jank
Directly send PCM samples.
Wait, that's what you said at the end...
...except my idea uses a single, "rolled" loop, instead of repeating data as your idea says.
Samples are sent just like in the DPCM example, read from a place, said value is different every time.
While I estimate this will be much slower than your idea, it should be much, MUCH simpler to execute, and failures will only mean missed samples rather than the processor doing whatever it pleases.

Video
*Use a per-scanline interrupt*, which would have multiple uses, like:
Per-scanline "fine" scroll, allowing for the color grid to better match the data.
Use the "emphasis" and "grayscale" bits for a slight palette increase.
Hook for a consistent 15khz sample rate for the audio
Hook for per-scanline palette change?
It should be possible to change it by disabling rendering, and thus ppu fetching, while the ppu should be fetching sprite data.
Might be hard to control what it does after, you might need to set the current address, count cycles or something weird...
Simplify "knocking" procedure, instead of an odd reading sequence, write a specific value to a specific address to enter "palette entry" mode, then read from another address to get the new palette.
Jam in sprites somewhere??

Would love to see this attempted again but with a PICO instead. Still heaps faster than a 6502 but you'd most likely not have timing issues with the gpio pins

Nice rickroll!

It's better to rewatch these old videos than new videos. This and harder drives

One less classic nes Zelda game in circulation 😢

You bet your bottom dollar that I watched this to the end. Every second. I want the rest of your great explanations. This is so fun

You are a genius

Why not use a toploader

Giving a Nintendo "super powers" is stupid

Hmm.. a few years ago I was messing around with an RC car, I wanted to control it with a raspberry pi over a webserver without breaking the RC support or modifying the board at all. I have no electrical experience. So I just probed around until I found a pin that went high when I gave it controls. Stupidly, I connected gpio directly to this, which blew out that gpio pin. After that, I just threw some diodes between the RC board pins and the rpi.. and that ended up working, but now I am wondering if I missed some complexity and was slowly damaging either board.

Chadbovick

why didn't you use a level shifter? they are already doing what you want.

Given the setup of the first video I’m curious whether it would’ve been funny to watch this one first

Great video!
Source code? :)

Hey for the record, the NES has a 7805 linear regulator inside which generates the 5 volts. Increasing the output of the 9V supply to the NES will not necessarily increase the current handling capacity, as the 7805 is the limiting factor. Most of them are rated for peak 1 or 2 amps.

Why not use a 2-way logic level shifter with a single logic-level MOSFET like a BSS138?

Wow 😣 you're boring 🥱

I, in fact, love larger videos. And much more to see the projects being made rather than to just see the final product because then I can learn how to avoid issues the creator found himself in.

Amazing content.
I especially appreciate leaving the errors and the subsequent debugging of those errors and gotchas. I would watch more :)

So, if I'm understanding this correctly the Raspberry does the heavy lifting for emulating and just spits out code that the NES understands? So in this particular instance the NES is a glorified video output.

Have you considered trying to make a custom integrated circuit with software like 'Electric' (that's the name of the program; it's an open source VLSI suite) for some of the components? This is something I've found myself kinda interested in (due to playing waayy too much Kohctpyktop), and would kinda like to know what the process would be like, to design an IC and then send the design off to a fab to make a few.

I have no idea how cheap or expensive it would be, or if it would at all be worth it, but I'd love to see someone smarter and with access to more resources than myself try it.

I enjoyed the 40 minute videos would love to see a follow up.

Sometimes I really realize how much I have left to learn.

Nvidia (with Geforce Now) streams RTX graphics to an android phone, This guy plans to stream graphics and sound to an NES. No reason it could not also be RTX graphics! Loved your video, entertaining and useful tips

oh my god i just unsubbed .... oh wate no i didnt your work is awesome .

This was great. What a brain! Amazing

I know this is mega old by now. But your progression through the electrical aspects of this is great. Everything you describe is what every new EE goes through. Why don't these transistors work like I think, why can't I drive this, what is TTL, etc is great! It's very impressive that your persevered. Many of these problems can be super frustrating and difficult to figure out. As I'm sure a million nerds already (jokingly) pointed out, use less vias!

only just discovering this channel but u are amazing sir

We need more videos like this please

This was amazing.