Worklog Wii U R&D Thread (WURD)

[Lazr]

I've started sanding + scanning the WUP-50 mobo. It is only half of the board, but the other half is pretty much the same as the WUP-01 and the half I am sanding will be useful for a compendium for troubleshooting trims (since the LOLWUT trim line is in this area) and a complete pinout of the SoC.

The scans can be found here!!

ATM there are only the two surface layers. The internal layers will be posted there when i can get scans for them. Have fun!

 

[Lazr]

The Wi-Fi module doesn't work atm, but it is probably due to my crappy wiring for it (SDIO is really bad; those cafeflex' will be incredibly useful for this relocation).

I finally got the relocation to work! Only a few months later but hey, better late than never.
Now the WiFi module technically doesn't need any of the pins going to the SMC/SOC, but there is a pin that needs to be pulled up for the module to function, just like on the BT module. I show that on my Wii U Mini worklog here. The SDIO wiring could have also been an issue, but the current wiring is much cleaner.
Here are pictures of my relocation:

I just scratched off the solder mask on the via that connects to the pin and then tied it to 3v3.
If you want to be extra safe, you could do the same to the SoC pin on the module, but IME its not required.

 

[zal16]

Right. They're huge, they need at least 12V, the 1v one is two-phase (meaning it has two giant inductors), and we don't have full documentation for the controller ICs.

The 12v input they require is the main issue. Going from 3.7V→1v is a lot more efficient than 12v→1v. And managing 4S batteries in a portable is trickier, since you need balancing. Custom regs solve all of these problems

What if the reason for needing at least 12v is because of the 12v regulator?

Because if the console is powered with less than 12v, the 12v regulator stops working and does not send the ok signal to the SMC IC, correct?

It's possible to try to remove this regulator (and spoof your signal on SMC), so that we could lower the input voltage to at least 5v?

 

[YveltalGriffin]

The 12v regulator PGOOD can easily be bypassed (as seen here). The issue is that the two core regulators are designed for 9V - 18V input:

This is very normal for wall-powered devices. Computers use the 12V ATX rail to feed their VRMs, too. Using a higher voltage like 12V means the power supply has to provide less current, which means wires and traces can be thinner.

But in a portable, you would need to put batteries in series to get such a high voltage, and then you would need active cell balancing and high voltage charging, all of which greatly increase design complexity and risk.

I mentioned before that the stock regs are huge, but maybe it's not clear just how big they are. Take a look at the 1V regulator: the solution size is approximately 1360mm² and it requires tons of large components on both sides of the PCB.

Bistro's 1V regulator solution size is approximately 360mm² if you count the caps on the back, which is 1/4 the size of the stock reg. Bistro's regulator probably performs slightly worse than the dual-phase regulator, but that's an OK tradeoff to make it tiny and have it run from 3–5.5V.

With how much work it took to miniaturize the trim, it would be a travesty to settle for the giant stock regs.

 

[Lazr]

Never thought id ever see this, but here we are.

The SMC and RTC are not required for boot!!

Video showing off the console boot:

This shrinks down required board trimming to just the main trim (unless you want to manually wire SLC/MLC) and the relocations required to just SLC/MLC, WiFi, 5GHz, BT, and USB.

I would still recommend relocating these for the time being while the software side of things gets straighten out. Without the ICs, the console is just stuck in ECO mode (standby mode) until you get a lucky reset. You can get rid of the reset fest by doing PRSH Tweaks -> Trigger mcp_recovery in minute, but since the SMC handles the front IO, the only way to do this is with a Pico (you can hear me do this in the video), and the debug vias are covered up on the cafeflex. This can be fixed with a custom build of minute that calls the function that sets that mcp_recovery stuff in PRSH, which I was semi-successful on, but I couldn't figure out why it wouldn't work with the fastboot minute. I will still try, and I will edit this post when I get it working.

The RTC handling the clock is not an issue imo, as an Aroma plugin (Wii U Time Sync) will update the time over NTP, so it should always be accurate when the console is connected to the internet. If you actually play your portables, more than welcome to relocate it if you want to play it on the go (think of it like MX on the Wii).

 

[zal16]

Someone have puctures of a naked and sanded DRC (gamepad) motherboard?
Because i'm searching manners to improve the sound quality of her.

 

[Lazr]

I have meh scans of the top layers with all of the components still. Haven't got much further than that though..

 

[zal16]

Do you know where the volume slider potentiometer tracks go?

 

[Lazr]

Sadly, I do not. That is routed on the internal layers which I don't have scans for (yet).

 

[r3vert]

I was wondering if this project by MattKC would work with these Wii U portables https://github.com/vanilla-wiiu/vanilla, it allows devices like a steam deck to simulate a Wii U gamepad, with this could you make a "headless" Wii U portable that connects to a handheld PC like a Steam Deck?