Worklog Nintendo Vegas

[YveltalGriffin]

After months and months of procrastination, I've finally begun assembling a prototype of my Wii motherboard redesign, which I'm calling Vegas. The name is an homage to Nintendo's naming scheme for the Wii's chipset (Broadway and Hollywood, both famous hubs of entertainment in the US). I designed the proto board in Altium, referencing both the compendium and additional mobo scans I took. While I tried to replicate the Wii's stackup as closely as possible, I couldn't get the GDDR3 impedances to where I would've liked. Worst case, I can try tweaking the GDDR3's ZQ termination resistor.

This ready and willing RVL-CPU-40 was the donor motherboard. The first step was to desolder the Hollywood and Broadway. I bought some neat BGA hot air nozzles that match the size of the Wii GPU and CPU to make things easier. (Flux added around perimeter of Hollywood after pic was taken)

I'm pretty happy with my workbench setup for this process. The Omnivises, hot air stand, and vac pick-up tool really streamline things.

Next is reballing. The overall process is as follows: clean BGA pads with 63/37 solder, flux, and solder wick; apply thin layer of tacky flux to the bare chip; use reball jig and stencil to apply solder balls; reflow under gentle hot air. The magnetic jig I'm using is pretty good, but it's still challenging to get the chip level and flush with the underside of the stencil. Speaking of stencils, I got an assortment of video game console stencils that included Wii-specific ones. No fooling around with universal stencils here.

After a couple tries, I mostly got the hang of the process and had a sexy reballed Wii chipset on my hands (and solder balls ALL over my workbench). Luckily, I bought fresh GDDR3 for this project, so I didn't have to desolder and reball the original RAM.

Hopefully all these image embeds don't break... Next post will cover assembling the PCB!

 

[Wesk]

 

[YveltalGriffin]

Solder paste stenciling was an absolute nightmare... For some reason, JLCPCB's stencil was warped right in the middle between the GPU and CPU. I eventually managed to get the CPU and GPU pasted properly but had to wipe off the RAM pads and do them separately in a second pass. I've gotten dozens of stencils from JLC and haven't had this issue before so I have no clue why this particular stencil was so screwy. Foreshadowing...?

I placed the GPU, CPU, and RAM and popped it into the reflow oven. A few minutes later and the board was looking pretty good!

Next I transplanted the NAND and AVE from the original mobo. I figured I should rely on the reflow oven's controlled temp profile for the BGAs and do the ICs that are easy to solder by hand manually.

Being the rather lazy person I am, I did not delve into the land of reference designators and Shank's Super Thread to buy a fresh BOM of passives for the proto board. Instead I just transplanted passives from the original mobo. After tacking everything down, I went back over with SMD291 and hot air to get all the joints solid.

Once all the passives were transferred, I put the board into the ultrasonic cleaner for a while. Here's how it turned out:

Unfortunately, here's where my luck ran out (for now). When I checked the voltage rail resistances I found ALL the main rails were shorted together. I reflowed the Hollywood a few times with copious amounts of flux and jiggled it around, but it didn't change anything. Having already taken some glamour shots of the sexy assembled PCB, I desoldered the Hollywood and the short between 3.3V and 1V went away. There must have been some seriously major solder blobbing going on under the BGA for 3.3v and 1v to be shorted together...

I didn't enjoy removing the GPU right after finishing the assembly, but such is life with experimental boards. Debugging is inevitable with projects like this! A 28x28 BGA is no joke especially in the suboptimal conditions of a hobbyist's home lab. Thankfully I can just rework the BGAs onto this same mobo with hot air while keeping the passives, NAND, and AVE on the back.

I think I'll try reballing the GPU with slightly smaller solder balls (0.5mm instead of 0.6mm) to reduce the chance of shorts under the chip. As mentioned above I'm also going to rework the BGAs back onto the mobo with hot air, which should let me monitor the reflow a bit more closely. Once the solderability issues are sorted out and the chipset is on without shorts, I can begin the pre-powerup checklist!

 

[supertazon]

Truely a work of art, the Sistine Chapel of Wii portablizing. Fingers crossed you get the issues ironed out!

 

[ScrambledFrequency]

Holy crap that is nice!
How come you decided to do a custom PCB?
It doesn't look any smaller than an OMEGA trim.
Just out of curiosity, I'm sure you know better than I do!

 

[StonedEdge]

Holy crap that is nice!
How come you decided to do a custom PCB?
It doesn't look any smaller than an OMEGA trim.
Just out of curiosity, I'm sure you know better than I do!

 

[buzzyc]

In practicality sense, all the test pads for BT, USB, Voltage rails would be nicer to work with.
In reality, getting to the point of the Vegas undermines the ease these test pads will bring
But, pretty!

 

[YveltalGriffin]

Yep, a single custom PCB will never be able to beat the OMEGA when it comes to size (especially since OMEGAs "stack" the NAND and AVE under the RAM and CPU.) But it'll open the door to more integrated builds, like a single-PCB portable with regulation/charging, GC+, and direct drive all on one board.

I'll admit it's not super practical-- making an OMEGA is a one-evening job with no reballing required. But Vegas is a fun way to challenge myself and learn new skills, while pushing the hobby a little further. It feels like the inevitable endgame of Wii portablizing.

 

[thedrew]

This is wild, honestly a bit speechless. Everything is just so... perfect. Your hand soldering is a work of art, beautiful.

 

[honigbrai]

If you would put the voltage regulators on the board it might be smaller than omga + voltage regulators

Yep, a single custom PCB will never be able to beat the OMEGA when it comes to size (especially since OMEGAs "stack" the NAND and AVE under the RAM and CPU.) But it'll open the door to more integrated builds, like a single-PCB portable with regulation/charging, GC+, and direct drive all on one board.

I'll admit it's not super practical-- making an OMEGA is a one-evening job with no reballing required. But Vegas is a fun way to challenge myself and learn new skills, while pushing the hobby a little further. It feels like the inevitable endgame of Wii portablizing.

 

[RegDogg]

Is this project open-source? If not I have a bunch of Wii motherboards I can donate, the only problem with them is the AVE is damaged beyond repair, but I can confirm it boots

 

[Kirbo]

Yeah I would gladly donate a couple wiis to the cause if this is open source because this is amazing.

 

[loopj]

I recently watched a video of a reballing approach using solder paste. In this approach, instead of reballing the chip, they reball the board itself. I feel like this approach might make it easier to keep things flat and level.


Maybe helpful here?

 

[steve_m]

Yep, a single custom PCB will never be able to beat the OMEGA when it comes to size (especially since OMEGAs "stack" the NAND and AVE under the RAM and CPU.) But it'll open the door to more integrated builds, like a single-PCB portable with regulation/charging, GC+, and direct drive all on one board.

I'll admit it's not super practical-- making an OMEGA is a one-evening job with no reballing required. But Vegas is a fun way to challenge myself and learn new skills, while pushing the hobby a little further. It feels like the inevitable endgame of Wii portablizing.

are you still building the own pcb ?
i planned to do that too have some time next month and i am looking for a lot of information about it.
currently i try to build the package and later route the connections in eagle.

 

[thedrew]

I recently watched a video of a reballing approach using solder paste. In this approach, instead of reballing the chip, they reball the board itself. I feel like this approach might make it easier to keep things flat and level.


Maybe helpful here?

I’m not sure if Aurelio or Yveltal tried this approach in their board designs, but I have tried this approach many times still with no success. Voltages and resistances are all good, chips warm up like they should, but no actual boot. Still working on this though, and I do believe ultimately this is the right approach! One day we will have one of our redesigns working!

 

[steve_m]

another reballing aproach is, to "reball " the chips with a solder stencil like : solderpaste through the stencil on the chip, lift stencil , heat the chip with a heat gun to crate the balls, add a little flux. The board itself is first tinned and then cleaned of the solder and then let a thin flux coat on it. Then place the reballed chip on the board and heat it up to solder.

i think he had too much solder on it from the reballed balls and the solder from the stencil of the board , maybe the balls where too big (like he said) or the stancil for the pads was too big too. I am sure there is a application note that explains the ration of ball and pad solder

 

[YveltalGriffin]

I was wrong in my initial post when I said the shorts were from "solder blobbing" under the Hollywood. What's actually happening is damage to the Hollywood during the reflow process. We're not sure if it's popcorning due to moisture or something else, like the die attach melting.

Drew, Aurelio and I have discussed ways around this at length, and they've both tried many exotic approaches for removing the GPU without stressing it too much (cutting around it, sanding through the mobo, lasers, etc). In the end, it boils down to none of us having a proper BGA rework machine with the correct profile dialed in. According to the Chinese modders who have gotten Wii motherboard redesigns working, even with a proper rework machine (see below) and an optimized rework profile, their yield is still not 100%. It's totally possible, but I'm not interested in buying a rework machine for >$1k USD for a single project

My post from RedEnergy's thread with my thoughts on the best way forward:

Working on my own redesign isn't a priority for me right now, but IMO here's what needs to happen for any of the redesigns to move forward:

• Full characterization of a softmodded, stripped down stock mobo or omega trim:
  • current traces of all voltage rails at boot, measured with sense resistors
  • logic analyzer captures of the NAND lines, and digital video lines at boot
  • logic analyzer capture of boot0/boot1 debug port at boot
• Rework of redesigned mobo to include:
  • current sense resistors on all voltage rails
  • test pads for NAND lines, digital video lines, maybe EXI, and boot0/boot1 debug
• Proper temp-controlled reball and rework of the GPU onto new PCB (I think RedHerring's attempt is the best/most controlled yet)
• Full characterization of the redesigned mobo, just like with the original mobo:
  • current traces of all voltage rails at boot, measured with sense resistors
  • logic analyzer captures of the NAND lines and digital video lines at boot
  • logic analyzer capture of boot0/boot1 debug pads at boot

This test plan would allow for a 1-to-1 comparison between a stock mobo and the redesigned one, and should make it pretty obvious if there are hardware issues, if boot0/boot1 are succeeding or not, etc. Hopefully someone has the bandwidth to tackle this

 

[Stitches]

Unfortunate

 

[steve_m]

old chips are often tempered at 100 deg C for a while to get moisture away. There are heating pads where you desolder them with heat from under the board, that way less stress on the chip itself. Then use the heating plate again for soldering the chips on the new board

 

[blueapple]

This is some amazing work YveltalGriffin. Just curious, have you tried desoldering the CPU+GPU from a working Wii motherboard, reballing them both, and then soldering it back to that same motherboard to see if your soldering process is good?

I have also heard about what steve_m mentioned, where you want to heat the board beforehand to remove moisture. Have you tried dehydrating the chip, preheating it in a reflow oven, and then desoldering it with the BGA reflow tool?