Worklog Ashida Worklog

[Orako]

6/13/22 - Finally started work on my first portable the other day, and man was it fun. I got so wrapped up I almost forgot to write a worklog, so this is going to be a lot of stuff on the first post.

First things first, the prep. I purchased a black Wii off of Ebay for around $30 usd total. It turned out to be an RVK-CPU-02, neato. All my orders are currently in except for an Ashida case since I'm still waiting on a reply from the guy I'm ordering from. I don't have a printer, so I went with the cheapest option I could find and it took some digging but I found a service off of Etsy with plenty of positive feedback singing praises, who claimed to quite literally beat 3d printing service prices by 50%. We'll have to see how that goes. Any tips on info to provide him to ensure a successful print would be nice. I'm planning either black ABS or clear resin if it's not too expensive (long shot I know). Otherwise, my first digikey order, my 4layer order, and my amazon orders have all arrived. I'm waiting on backordered speakers from digikey and the variable resistor from mouser which should arrive tomorrow.

For this build I'm going to be following the Ashida build to a T with a couple exceptions. I would like to have the mx chip and the bluetooth module since I'm planning to make this a gc/wii portable. I'm comfortable using a wireless sensor bar or even the double candle trick if I need to. I don't currently plan on modifying the Ashida's case in any way, but in a future project I may try to add gamecube ports and/or an HDMI out. Finally, since it was sold out at the time, I'm using a pms-lite instead of a pms-2 so we'll have to do a little more work with the 1.8v line.

Aaaaaand we're off! I started by softmodding the Wii. This happens to also be my first experience with any homebrew as well. The Wii I purchased happened to come with the Homebrew channel preinstalled, but for the authentic experience I wiped the console and reinstalled it myself.

(Apologies for the blurry photos. I'm planning on editing this all into a video at the end and I realized literally today that my phone was recording in 720p30. Quality should greatly improve in the coming edits/replies)
Got letterbomb working pretty easily, installed the Homebrew channel and later went back and installed Bootmii and made a backup. Not much to really say here, it's a straightforward process if you follow the guides.

RVLoader was a bit more of an involved process. I didn't even know about the WADs the first couple times I tried to install it.

Eventually got it though, so that's cool. Patched out WiFi, patched in VGA, and all games worked well. Even played through a bit of Super Paper Mario while I was waiting to trim.

Now begins the hard part. A bricked Wii is maybe recoverable. A shorted or improperly cut Wii is rarely recoverable. I think, anyways, it's still my first time. Anyways, trimming the Wii was daunting but i could always cut far outside the lines. What really freaked me out was the u10 relocation. After I opened up the Wii and saw just how small the solder joints were, it definitely made me a little nervous. You know what that means: practice board time.

I really like taking things apart, so I happened to have an old laptop mobo on hand. I found a component that looked like u10, with three legs on one side and two on the other, and practiced the blob method there as I didn't have a hot air station. I tried to use a heat gun until people told me not to... and the cheap $30 Amazon junk one I bought started smoking. Yikes. When I felt I had the technique down I moved on to u10 and u9. I started with trying to remove those since if I burned up the traces on one, I could retry on the other, whereas burning the traces on u5 would be mostly unsalvageable for my skill set. With my iron set at 500f, I attempted my first micro-soldering experience that really mattered.

It's unfocused as heck, but I got it! After fiddling with u9 until I was worried I would burn the traces off, I tried again with u10 and managed to get it pretty easily. Quite the rush of adrenaline when it finally starts sliding around, though I did pry it a little with my tweezers. Thankfully I didn't tear any traces up. Next up was the relocation.

u5 actually came off way easier than u10 or u9, no prying needed. Guess I started to really get the technique down. I removed the surrounding capacitors, cut the trace and moved on to soldering.

It went really well! Using some flux as adhesive, I aligned the u10 under a cheap digital microscope I bought and began soldering. The burnt-looking stuff next to u10 is just flux, I actually haven't burnt anything on the board yet which is really cool to see. The one ripped trace underneath is from when I was removing components for the trim, I didn't really care since that trace is getting trimmed off anyways. Next up we have the jumper wire.

Honestly? The hardest part was keeping track of the via on the back. I had a little mishap where I didn't realize the via had solder mask on it, but under that microscope scratching it off and looking for the gold was quite easy. Really proud of myself so far. All that's left before the trim was to hook up the RVL-NTC.

Still alright, but it couldn't be perfect forever. When I eventually attempt this again, I see myself trying to find a different via to thread the u10 jumper through. As it stands the RVL-NTC covers the hole the jumper threads through, meaning if the u10 relo was a failure, which I can't check since I had already removed components for the trim, then I'll have to undo the NTC as well. This also meant the solder for the NTC came dangerously close to the via it was threaded through. Definitely all good notes to take for future projects. Otherwise, actually soldering on the NTC was tougher than the other joints, believe it or not. Maybe I wasn't using enough flux, but no matter what I tried the solder only seemed to blob up on the NTC, not on the capacitor. Eventually I got it and confirmed continuity, so I'm satisfied even if the joints aren't that great.

Which just leaves trimming.

Here's the trim line.

Slipped with the dremel, but I was told the board wasn't dead so I kept going.

Here's the rough board trim. It was around here I was told to cover the chips with masking tape to prevent dust getting under and causing shorts.

Here's the final, sanded, and cleaned trim. Initially wanted to keep that fifth screw hole, but eventually just decided to take it off to minimize the number of gouges I had in the board. All that's left now is to test resistances. My final resistances were:
com/+
gnd/1v: 60.1Ω
gnd/1.15v: 29.4Ω
gnd/1.8v: 36.7Ω
gnd/3v3: 6.03-6.14kΩ depending on gnd plane
1v/1.15v: 110Ω
1v/1.8v: 117Ω
1v/3v3: 6.36kΩ
1.15v/1.8v: 94.3Ω
1.15v/3v3: 6.16kΩ
1.8v/3v3: 6.19kΩ
They seem a little low compared to the other RVK-CPU-02 measurements I've seen so if anyone thinks this board is dead let me know!

That's about all I have for now. Excited for the rest even if I have to redo the trim. Can't wait to see the finished product.

 

[Valanduin]

I know it is too late for you, but you don't actually have to do the U10 relocation if you are using an RVL-PMS-LITE or RVL-PMS-2, both can emulate the U10 chip, so all you need to do is run a wire from the pad on the RVL-PMS labeld U10 to the pad/via where you need the U10 pulse. Hopefully your trim works but if it doesn't that might save you a bit of time.

 

[Orako]

I know it is too late for you, but you don't actually have to do the U10 relocation if you are using an RVL-PMS-LITE or RVL-PMS-2, both can emulate the U10 chip, so all you need to do is run a wire from the pad on the RVL-PMS labeld U10 to the pad/via where you need the U10 pulse. Hopefully your trim works but if it doesn't that might save you a bit of time.

This is good to know for the future! Not sure which I'll do in future builds. I like the idea of the extra simplicity, but at the same time I also like the ability to de-clutter my wiring job as much as possible, and keeping the u10 stuff all on one board is nice as well for just that tiny bit of extra neatness.

In other news, my Ashida PCBs are all prepped! Here's the final results:

My setup has improved since the last time, but it's a little reflective, working on that. There's a few things I want to point out about these, starting with the mistakes. These are things I want to note both for myself and for others. Even if this seems really long, I recommend checking out my note about Z-button switches near the bottom.

Mistake 1: I tried to attach the ribbon cable in the exact same way it was attached on the original GameCube controller, across the holes instead of in them. Wesk advised me in the future to just strip the wire enough to put it through the holes and solder it on the other side. It makes a lot of sense; the ribbon cable can be shaped to a degree so the length lost doesn't really matter, and routing it through the holes would have made soldering so much easier.

Doing it without routing through the holes saw me holding the ribbon cable in place with one hand and trying to press the cable on with my iron and a big chisel tip. Already this is not a good idea, as you should not be pressing hard on anything with the iron. Somehow I didn't even know this until Jeff told me. The solder also flowed through the big, empty holes, making me worried it was going to drip out. Cleaning and reflowing after was incredibly difficult, because since the leads just sat on the pads, they constantly popped off whenever heated either accidentally by the chisel tip or purposely when trying to reflow. In the end the result was messy, and I believe I may have been close to burning pads, too.

Mistake 2: Here's the huge one. Check your orientations guys! The Z buttons go on the side of the PCB with the text that says Z and Z2, soldered on the other side. The analog sliders should cover the white outline and be soldered on the other side. I had my second big oopsie of the journey after the dremel slip and put a Z-Switch on backwards, almost incorrectly soldering a slider, too. The Z-switch was the big one though, since I soldered all four joints before I realized.

Don't be like me. Because my switches were modified, they sat much tighter in the slots. I had to snip what legs I could, heat all four joints back and forth with a chisel tip, let the switch drop out, lightly wipe as much solder as I could off the half-pads on the PCB edge with the iron, and push a different switch back in carefully while heating the two inner pads on the other side. In the end, something is causing the two ground points to bridge ever so slightly and won't come undone. I believe it may be a leftover lead stuck in the solder from when I snipped off what I could. I probably shouldn't have cut stuff further in, since i couldn't wipe off leads left in the solder as easily as on the edge half-pads. I also may have risked ripping pads if the flush cutters applied upward pressure to the switch when cutting the legs. It's also possible I scraped off the solder mask while moving my iron back and forth between the joints. I wasn't really applying any pressure at all, but who knows since I rarely lifted the Iron when moving back and forth between joints. Because (I think based on continuity tests) it doesn't really matter if those two points are bridged, I'm not going to touch it anymore.

Finally, with the big mistakes out of the way, we can get into some smaller stuff and more technique-based advice. Firstly, a small mistake:

Components that sit on the pads rather than running through a hole, like the headphone jack and FFC connectors, should not have their pads tinned beforehand, otherwise the component will not sit flat.. I asked and was told this after already doing this first FFC connector, and if you hold it you can tell it's not as level as it could be. Fortunately, it doesn't really matter for this since the FFC cable can bend the little angle without any problems. I honestly recommend a cheap digital microscope for the purpose of aligning your tiny components. I used one of the rubber-padded alligator clips from my soldering station to hold things in place. Make sure it's padded in some way so you don't scrape things off moving it around while aligning the components. Start by soldering the edge pads to hold it in place, and then move on to adding a tiny bit of solder to the little legs and guiding it to the pads. Once again a microscope helps since the more of the pad is revealed, the easier it is to guide the solder on, and it's a lot easier to see that half the pad is covered under a microscope versus with your eyes.

Testing continuity between the holes on either side tells me the connections were all successful.

Another smaller mistake was that I soldered the wrong jumpers on the U-AMP, soldering J3 and J2, enabling both the analog controls and analog audio. The Ashida uses digital audio with analog controls, so only J3 is needed. This is a mistake I saw a few other people make as well so I wanted to put it out there. I think it's been said that you could easily get the other Wii-compatible audio settings working on the Ashida, but for one, don't quote me on that, and two, I have no clue if any different components are required or if it's just a matter of soldering the proper jumpers.

Thankfully it's just a matter of undoing the jumper with some wick. I didn't need to worry too much about getting all the solder, since you just need to undo enough to lose continuity between the two sides.

For desoldering components on your GameCube controller, what I found most effective was using a huge chisel tip. All of the components I needed to desolder had multiple joints in a line. I harvested the stickboxes, the analog sliders, and the ribbon cable. If you relocated U10 with the blob method, it's similar to what I did, which was to turn the board upside down, add some solder to the joints, and heat each joint quickly until the component falls out on its own. Same motions as the blob method but the component didn't come off stuck to my iron. For that reason I'm glad I relocated U10 at least this time, since I feel like I got good early practice with that desoldering technique before I was in too deep.

Once my component was out, there was some solder left on the legs that made it difficult to insert the component back into the Ashida PCB. I cleaned my iron's tip and tinned the legs as best as I could with what was there, making an even coating of solder on the leg that made it much easier to insert back into the holes.

Finally, for the Z-switches, I found that the ones on the Ashida BOM were not that great for the PCBs, even when modified. Instead I dug up the FIRES Z-button modding guide, which explains what DigiKey switches to use and how to mod them to fit the PCB. Based on what I've felt with the 100g buttons so far, I would recommend the 100g switches they provide if you have springs in your Z-buttons, and the 150g or maybe even higher if you don't. Regardless, it's a good guide I haven't seen much mention of that prevents you from needing to harvest two controllers. And one more thing, FIRES recommends buying 10 switches so you have some wiggle room for failure, and I fully agree, I messed up my first few switches and even after all 10 I still didn't have the method fully down, but I ended up with around six that could theoretically work.

And that's that for now! All I need before I can start wiring everything up is to 3d print. The local university my sibling goes to is offering 3d print services for students that we're taking advantage of, and auto quoted $40 for a full PLA set of Ashida prints on a Prusa MK3s, so we'll see how that goes and if the price skyrockets during the manual quote. All that's left is to figure out how to dig or dissolve the flux out from my tactile switches and slide switch since the tactile switches are getting stuck down and the slide switch is very lubricated, so if anyone has any tips for that it would be appreciated.

Until then, thanks to everyone who offered advice on the forums and on discord, and I will post again next time!