I've been working on a
3D-printed multi-out panel for the toploader and it was working great right up until I tried closing the case back up, at which point the video output stopped working on either of the two video outputs I had wired up (composite and component), and the only output of any kind that I get is a slight audio hum. The screen is just black, not even a momentary blip as I throw the power switch like I would normally see, even with no cart inserted. The power supply and main power rail are fine, but that's as far as I've gotten with my troubleshooting at this point. Any ideas?
Try removing your component video encoder modifications, and un-do everything other than leaving the NESRGB in place. Be sure power is going where it's needed.
From my experiences I've found the second 7805 for the NESRGB isn't really necessary. Try removing it if you are using it just to eliminate a node of failure.
Alright, I'll give that a shot. I wasn't using the extra regulator, I was planning on metering out the main regulator to check total current consumption after the mod was complete and decide at that point whether or not it was necessary, or maybe just replace/upgrade the main regulator instead.
No luck. I even tried removing the NESRGB completely and putting the PPU back in and wiring up the composite video directly, and I get absolutely nothing out. No picture, no sound. Did I get the wiring right? Yes, I know the red wire isn't needed. (R: +5V, Y: Video, W: Audio, B: Gnd)
You can figure out if the CPU is alive by probing the address lines with your audio input. If you hear a bunch of funky square waves, then the address lines are pulsing, which means it's at least doing something.
Started poking around with my multimeter on Hz/duty cycle to see what, if anything, I could find. No signals on the CPU and then it got hot enough I could smell it. So... that sucks. I read 350ohm between +5v and Gnd on the CPU, and about 2.5kohm between +5v and Gnd on the PPU (removed from the console). Something definitely went boom.
Just removed the CPU, and by itself, resistance from +5V to Gnd is 650ohm. The main board, with CPU and PPU removed, shows 850ohm from +5V to Gnd. Is that normal, or is something else on the board blown?
On a known-functioning NES-CPU-07 board, with the power switch daughterboard removed, CIC removed, but the CPU and PPU present, I see 260Ω from Vcc to Gnd, and 250Ω from Gnd to Vcc.
The CPU, not in a socket, reads 660Ω from Vcc to Gnd. The above mainboard without the CPU measure 427Ω. I damaged one of the pins on the PPU when I socketed my NES, so forgive me for not measuring that part...
Hey I've troubleshooted my nes-101 a few times. Do you have any spare parts? To begin with I'd probably verify power. Check for anything that looks blown on the board. After I'd try and take out the nesrgb and put coax back or a simple av mod. PM me and I can try and help you out as much as I could.
Icelvlan wrote:
Hey I've troubleshooted my nes-101 a few times. Do you have any spare parts? To begin with I'd probably verify power. Check for anything that looks blown on the board. After I'd try and take out the nesrgb and put coax back or a simple av mod. PM me and I can try and help you out as much as I could.
NESRGB is out, PPU is back in place, that's the way it was when the CPU temp shot up. I don't currently have any spare parts, but I'm ordering a couple of beat-up-but-functional toasters for parts. I'm also going to order new caps and a new 7805, to start. I probably need to replace them regardless, and then I can rule out the power circuit.
Well, I replaced the regulator and all of the electrolytic capacitors, but when I press-fit the CPU back in place (didn't want to solder it back in just yet because I wanted to test the CPU on another working board first), it heated up again to the point of almost burning my fingers. I had another CPU taken out of another console, so I tried press-fitting that one in, and it didn't heat up at all. My PPU doesn't heat up either, so it might just be a fried CPU. At this point, desoldering the CPU is enough of a pain that I don't want to solder in a new one until I've tested it, so I'm buying a working front-loader that I plan to pull the CPU/PPU out of and socket them both so I can use it as a test board. I was really glad to learn that the 'G' revision CPU/PPU are so common. I was worried it might be a top-loader specific revision. Just gonna take it one step at a time. I can rebuild it. I have the technology!
In general if an IC heats up instantly like that, it's safe to guess it might be toast. If you aren't pressing it into a socket, though, that's not a real test. It's hard to guarantee the needed pins are all making contact. Just solder in a socket!
Yes, I know it's not a real test, but I don't really want a CPU socket in my top-loader. I was more interested in testing the fact that with the CPU out, nothing else gets hot and the power rail seems stable. Then, the minute I press the CPU into place it (and only it) rapidly heats up to the point of burning. Once I get my hands on a working front-loader I'll put sockets in that and use it as a test board, and if the CPU does turn out to be bad, I'll just use the CPU from that front-loader.
qwertymodo wrote:
Yes, I know it's not a real test, but I don't really want a CPU socket in my top-loader.
I highly suggest you do so, that way if the CPU somehow dies again it'll save you a ton of time replacing it.
There are serious clearance issues with the NESRGB adapter board. As in, basically the entire row of CPU pins is within ~0.1mm of the PPU signal pins on the adapter board. Unless I can find a seriously low-profile socket, it's just not going to happen. I'm ordering a few samples of the Samtec SL-series low-profile machine screw strips, but even then I'm not sure how much clearance I'll have.
You might be able to use another socket as a riser for the NESRGB board, giving you the needed clearance.
The low-profile machine screw strips, combined with flush-cutting the bottom pins on the NESRGB adapter board, finished off with a layer of hot glue to insulate them, and I think I can call it safe. And with that, I have it all hooked back up with a new CPU/PPU, and it works! The NESRGB is completely unharmed. Now I can get back to wiring up my 3D printed back panel
Thanks for the help, guys. It was pretty straightforward stuff, but it always helps to have a dialog going while troubleshooting to keep ideas flowing and keep from doing something stupid.
qwertymodo wrote:
my 3D printed back panel
Pictures when you're done?
Oh, that's pretty. Using a Wii component cable?
lidnariq wrote:
Oh, that's pretty. Using a Wii component cable?
Yep (well, Wii RGB cable right now), reasons being that 1) the SNES multi-out doesn't have component, and 2) Wii cables are easier and cheaper to find than OEM SNES cables, and half of the 3rd party SNES cables you can get these days are complete garbage. Wii can do either S-Video/Component (NTSC pinout) or RGB/Component (PAL pinout), in addition to composite (composite video is used as the RGB sync signal). Right now I have RGB wired up, but eventually I'd really like to get RGB and component both working at the same time. I'm working on building a video switching circuit to support the dual output (each of the signal pairs shares the same 3 pins on the connector, and the cable indicates which one to use by shorting a pair of pins to select component).
Here's the clearance issue area:
That's with a super low-profile machine screw strip. A normal socket is much taller, and the pins on the adapter board originally stuck down quite a bit. They were basically touching the corners of the CPU legs.
Comparison photos, because I managed to steal my coworker's tripod long enough to get decent shots. Click for full size. Getting good photos of CRT screens is hard...
Original PPU Composhite output:
NESRGB (RGB output, natural palette), look at the much sharper text, the border around the brown background, and the clearer color of the sky:
NESRGB (RGB output, improved palette). Just... yes. So much yes. The photo doesn't even begin to do it justice:
Also, while I had the thing apart, I took the time to actually measure the current draw with the NESRGB in place, and it ran pretty steadily at 390mA with the game cart and a single controller. With a 1.5A regulator, that should still be plenty of headroom for a PowerPak or anything else you might want to throw at it, without the need for the secondary regulator. Not sure why Mario is blue, the meter must've introduced some noise to the power rail or something weird like that. The game played fine.
Mario is pipe/bush colored. This means the RGB mod is using the background palette with the same number as the sprite palette. There are only four color output pins, not enough to distinguish the background palette from the sprite palette. For this reason, the typical modern RGB mod uses the PPU's composite output to distinguish background pixels from sprite pixels. Something is causing this composite output not to get through.
Sounds about right, from what I was seeing elsewhere. Taking the meter out and hooking the Vout pin back into the mainboard took care of the problem though. So maybe the PPU wasn't functioning for some reason, which might imply that the actual current consumption was higher than what I was measuring, but even then I think it's pretty clear that there's plenty of headroom on a single 1.5A regulator, which was all I really cared about in testing this.
qwertymodo wrote:
This part is so beautiful! did you make it yourself? also the little board that goes on it did you make the whole thing? how are you wiring that thing up to your NES RGB card?