Sourcing a Nintendo multi out socket?

Kevtris hasn't set a price on the Hi-Def NES kit yet, but there is also still a chance we will work out how to mount an HDMI connector to the rear panel and then will want a special one made for the nes-101 and hvc-101. My next order will be for ppl still using rgb with crt or upscalers. About half of the list decided to wait for the Hi-Def NES instead of using the nesrgb.

 

That's what I assumed and hopefully drive prices down for the NESRGB at least which will be better overall for people who have shyed away from it due to price. When is the ETA on the HDMI board?

 

No eta yet, but will prolly start doing more extensive testing in a couple weeks.

 

This thread inspired me to try my hand at 3D modelling because I wanted to make a few minor improvements to the already awesome panels you guys had made here. Most visibly, I wanted to use a simple momentary tactile switch for palette switching, in order to greatly reduce the footprint of the switch on the otherwise rather cramped panel (and I didn't want to have to drill holes elsewhere to accommodate it). I have (hopefully) one last revision to print in order to resolve a few minor measurement errors, but so far I'm very happy with the results. Just press the tact switch to cycle through palettes, then hold the button for about a second to save the current selection as the default. It uses a JST ZH connector to give it a nice tight fit in a compact footprint.

I'm using the Wii connector, because I want component out, which is not present on the SNES connector. I am also in the process of working on a video switching circuit so I can utilize the Wii connector's built-in switching behavior to enable either RGB/Component or S-Video/Component on the same pins (the first option is the PAL pinout, the second is NTSC), but that might be a rather long way off... I don't fully understand the analog video specs and the component requirements for properly driving the signals.

 

I have seen this in Nintendoage and it's very nice especially the switching of the pallets. If you don't mind me asking what did you use to achieve the switching? Some arduino or pic? Can you share the code? I have a genesis project that could benefit so much from this. Also a heads up don't use anything with aluminum powder mixed in so you don't get shorts.

 

Looks like you are using Alumide. If you are, you may want to reconsider. Alumide makes the part brittle. So areas such as the two mounting tabs and the two small socket clasps may be prone to snapping. Also, like what Helder mentioned, the aluminum power in the Alumide maybe a safety hazard. There's always a potential for the metal powder to flake off and get on the electronics.

If you want something other than the raw white color, get your part dyed gray from Sculpteo if you already used Shapeways. They cost about twice as much, but at least the color matches the lower half of the case.



Those pictures are from my own install. I have that printed from Sculpteo. I also incoporated a spot for mechanical switch.

 

I started trying to implement the switch with a bunch of discrete components until I realized that a PIC10F322 was $0.65 in single units, and there was no way I could beat that price using multiple discrete components (debounce, counter, decoder). So that's what I'm using. The selection is saved to the high-endurance flash on the PIC (I chose to require a long-press to save rather than saving every switch to reduce writes to the flash, even though it is high-endurance rated). The code also does software debounce for both press and release using a shift register (a trick I picked up from Verilog), so no glitchy multi-presses.

Here is my PIC code: http://pastebin.com/b6d3ZUW2

It's pretty short and straightforward, but here are some pointers for modifying the code for other uses.

The amount of time to differentiate between a short and long press is determined in button_pressed() by the value assigned to TMR0. The lower the number, the longer the time. I think 239 is 0.5 seconds, but I don't remember exactly. You'd have to check the datasheet and calculate it based on the prescaler. Or just play around with it.

Your behavioral changes will probably go in the output_state switch block. If you want to add additional states, you'll also want to change the rollover in switch_output()

The code uses something like 75% of the program space, so if you want to add much more to it, you might have to increase to a larger chip, but I really liked the SOT-23 package, since that was exactly the number of I/O's I needed.

Here is the Microchip App Note with the Flash.h/c and HEFlash.h/c files, as well as instructions on setting up the project config to reserve the high-endurance Flash segment (use --ROM=default,-01f0-01ff for the PIC10F322)
http://ww1.microchip.com/downloads/en/AppNotes/00001673A.pdf

derp... I didn't even consider that. I did use the aluminum powder for my latest print and guess what? I blew up my top-loader a few weeks back. My original thought was that I'd killed the power circuit trying to run the NESRGB without the extra regulator, but now I wonder if it wasn't the back panel. I did run my multimeter all around the thing and it doesn't show any resistance (as in no conductivity, not 0 ohms), so perhaps it's not really so much of an issue after all? Shapeways claims it's not conductive... so I'm not sure what to think there.

 

Aluminum is conductive. But I suppose what Shapways is claiming is that since it's not solid aluminum, rather a mixture of nylon and aluminum powder, you don't get a continuous transfer of electrons. But since the part is literally sitting right next to the power supply, I still think it's not ideal to have the aluminum powder in that proximity. Aluminum power can be abrasive too, so there is also a potential for the multiout plug to be worn down prematurely.

 

I suppose that makes sense, but like I said, I can take a multimeter to my alumide panel and am completely unable to measure any level of conductivity whatsoever, even with the leads ~1mm apart. And also, it's not like there aren't plenty of other conductive surfaces in close proximity to that same area, such as the RF shield or the large exposed ground plane underneath where the RF module used to be. I think it will be ok. At least, it will be once I can get my hands on a new CPU...

 

What I'm trying to explain is that the metal powder is only encapsulated by the plastic and not fully fused; as normally plastic and metal don't naturally combine. The concern is that the metal powder may come loose and if there is enough collecting at one spot on exposed PCB, there may be a chance of shorting.

 

Finished up the video switch PCB that forms the second half of my project here. Basically, the Wii multi-out port shares the same 3 pins between RGB/S-Video/Component, and switches between them if two pins are shorted together (PAL gets RGB/Component, and NTSC gets S-Video/Component). With this extra add-on board, it enables this behavior so the Wii multi-out can be used to its fullest extent. You do have to choose between RGB or S-video, but it will properly switch between that and Component depending on what cable you hook up to it (Component cables internally short those pins, RGB/S-video cables do not).

This also required me to build my own component Pb/Pr encoder. The colors aren't perfect, but they're pretty dang close, and I'd say I'm happy with them.

RGB:


Component:


I did mess up the PCB dimensions a bit, and it butts up against one of the screw posts for the cart slot, so I can't put the top of the shell back on, but I'm revising the outline and then it should be good to go.

 

Sounds like you got a winner on your hands, this will likely get more people into the wii sockets and cables as they've had little interest shown.