Official Gamecube IDE/USB interface thread (PROJECT BEGUN!)

Electric Rain · Post by **Electric Rain** » Sat Sep 27, 2008 5:29 pm

palmertech wrote:The gamecube scene reminds me of the PSX homebrew scene... Fragmented, badly documented, and all the authors drop off the face of the earth after a few years.

Yeah, that's a good point. I wonder why that is...?

tom61 · Post by **tom61** » Sat Sep 27, 2008 8:19 pm

Electric Rain wrote: You're right, though. Hot-swapping isn't really necessary I suppose. I was mostly thinking about "hack" disks like SDload or GCLinux... don't those need to be hot-swapped? Eh, I really don't know how that works, I guess... The GC scene was never very well documented IMO.

You should be able to just load GCLinux or SDLoad directly. The only protection on the GC is disc protection, and this would bypass that, or need a mod chip anyways.

Electric Rain · Post by **Electric Rain** » Sat Sep 27, 2008 9:59 pm

No, I mean after they're booted up, I thought you could swap out the disks for a disk with other stuff on it... like multimedia, in the case of GCLinux. But now that I think about it, yeah... I'm wrong, actually.

The media has to be included... Same with emulators. For example, SNES9xGC requires the ROMs and the emulator to be on the disk together. You can't boot up a disk with the emulator on it, and then swap it out for the ROM disk. I guess I was thinking along the lines of the PS2 and swap disks (I.E. Swap Magic).

Okay, then that settles it; hot-swapping isn't necessary anyway.

extremesonic · Post by **extremesonic** » Tue Sep 30, 2008 8:37 am

palmertech wrote:The gamecube scene reminds me of the PSX homebrew scene... Fragmented, badly documented, and all the authors drop off the face of the earth after a few years.

Not really, I think the cube homebrew scene is one of the better documented, hell it is still worked on simultaneously with Wii homebrew(in a few cases, mostly stuff that started on the cube).

R3ason · Post by **R3ason** » Thu Oct 02, 2008 3:58 pm

Any luck with Digikey yet?

Electric Rain · Post by **Electric Rain** » Thu Oct 02, 2008 4:59 pm

The order should be shipping out on Monday if something stupid doesn't happen again.

*crosses fingers*

Electric Rain · Post by **Electric Rain** » Mon Oct 06, 2008 10:33 pm

Okay, something slightly stupid happened, but it's all good. I may be a day late, but I'm NOT a dollar short this time.

Everything has been sorted out, the order has been re-submitted, and it's shipping out tomorrow. W00t. I should have the parts by the end of the week, and... we'll go from there. I don't want to make any promises I can't keep, but I'll try to have the second video blog up ASAP. Around the same time, I'll be posting the final board design up here for everyone to check over. Once I get the okay from as many people as I can that everything looks good, I'll be sending the board design off to either Sparkfun or hopefully Splode if he can get his friend to make up a couple prototypes for me instead.

Edit: UPS tracking says it will be here by Monday.

Electric Rain · Post by **Electric Rain** » Fri Oct 10, 2008 11:42 am

Yeah, I'm triple-posting... what are you gonna do about it?

Just wanted to let everyone know that I got my parts early! Ub3r w00tage!

My cameras won't be with me for a few days though, so my next blog with have to be slightly delayed... I'll keep you guys posted with this. Right now, I have to go test-fit the parts and make sure everything lines up the way its supposed to on my PCB design. Well then, ja mata ne.

Dark-Aries · Post by **Dark-Aries** » Fri Oct 10, 2008 1:02 pm

That's awesome! That means there might be some x-mas presents for us to buy for ourselves...yay!

Electric Rain · Post by **Electric Rain** » Sat Oct 11, 2008 4:39 am

NOTE: This is really really long, but it's the most important part of this entire project; checking the design. I need all the help I can get! Anyone who has a few spare hours, please look this over. I'm asking a lot, I know, but we're FINALLY there! We're at the building stage, and we MUST ensure that everything checks out so I don't fry something stupidly!

Okay, so after test fitting the parts, I saw that I needed to reverse the IDE connector, resulting in the need to re-work the entire design. I originally did it like this on purpose, so it wasn't a massive mistake or anything... but, after getting the IDE connector and realizing the pins were too short to attach to the drive the way I intended, I had to flip it such that the drive would be attached upside-down. It's too hard to explain exactly what this means without visual aid, but you'll all see soon enough.

I like the new design much better, as it seems to be a bit roomier somehow and I was able to easily fit two holes for M3 screws, allowing the board to be mounted to the HDD. The HDD can, in turn, be mounted to the "main" enclosure, whatever that will be.

The parts are still pretty crammed though, allowing little room for silkscreen within the clusters of parts. Only what NEEDS to be labeled is labeled. Resistors are labeled simply with "R". However, based on their location on the board, it is easy to tell them apart. (They're spread out well. The two below the FX2 are 2.2K, the one above it is 10K, and the two up by the test points are 4.7K. Easy.) This was needed because the resistors and ceramic capacitors are all 603 packages, so they can't normally be told apart. Unlabeled 603 packages are all .1uF capacitors. The test points are labeled as well, with DB1-8 (debug points) and JT1-6 (JTAG points). A legend will still be needed, but there's enough silk screen to keep things from getting too confusing. Credit to destop, marshallh, gannon, R3ason, Spode (under the assumption that he can get his friend to make up a couple prototype boards for me) and myself are given in silkscreen on the bottom of the board.

I will NOT be selling this version of the board. This is a prototype, and there are things about this board that need to be revised before I sell them. For example, the mode switch debug point is being pulled up by a 4.7K resistor, but I think whether the device is in HDD or USB mode depends on whether or not this point is high or low. However, I may be wrong. Further experimentation is required, as it may be one of the debug pins that determines this instead. The cover switch debug pin is also being pulled up; it may need to be low at times as well. Not to mention the rest of the debug pins... what are they for, exactly? Facts are mixed in the Dextrose thread. I may also be able to cut several things out, making the board smaller and cheaper. But, I wanted to lay out the board as close to the original design as possible to prevent problems. Getting a working prototype is the top priority; improvements and cost-cutting solutions can be implemented later.

Board dimensions are still 70mm x 45mm, as they were with my original design. Yes, it's tiny. Making it a little bigger and giving the parts some room to breath wouldn't hurt anything, but I'm really cheap, and I didn't want to pay another $2.50 for a 6th square inch. Sue me.

So yeah, the current design is a little under 5 square inches, which will cost $12.50, plus the base fee and shipping fee (~$15-20). Once I perfect the design and get some of you to buy a few of these, I can order in bulk (ish), and get them much cheaper through someone else that doesn't charge extra for every square inch, so a couple more square inches probably won't cost any extra. I truly would prefer to make the board a bit bigger, actually. It's pretty cramped now.

I used an ELECTRA to route the board, which resulted in a few funky looking traces here and there. I could have fixed them, but I figured I wouldn't bother since it's just a prototype and they're still making good contact, which is all that really matters. Well, that and their length. You wouldn't want a 10 foot long trace spiraled up forming an antenna, I suppose.

With all of that in mind... take a look, everyone!

I need as many eyes on this thing as possible to make sure everything makes sense and I didn't miss anything vital. Ask questions, please. Anything you don't understand is probably because you don't know my exact design...

For those of you that can work in Eagle, here are the PCBs, schematics and parts library. I figure it will be easier to look everything over in Eagle. Go nuts.

Here's the schematic in PDF form... I didn't do this very well. I really only needed it for the Netlists, so while things are labeled to an extent, it isn't as clear as I'd like. For example, most of the parts don't have values, I had to design all of the parts except for the FX2 and the regulators, and I didn't lay them out quite how I should have... just compare them to the FX2 and you'll know what I mean. But, it's acceptable and should be easy enough to reference to the original schematic. THIS IS THE MOST IMPORTANT PART. This is what I need everyone to do if nothing else! Compare my schematic to destops. If the schematic is correct, the PCB is correct, since they're "synced" with each other. The following four images are of the PCB. Top with silk, top without silk, bottom with silk, and then bottom without silk.

NOTE: This is important... and confusing. As I've mentioned before, I'm using a daughterboard to connect the mainboard to the Gamecube mobo. The daughterboard will be soldered in place of the disk drive connector on the Gamecube mobo via pins that will protrude from the bottom of the board. These are routed to an 18 pin flat flex connector on the top of the board. There is another flat flex connector on the mainboard, and the two will mate via a 4" flat flex cable. This is how to check these connections: Look at Destop's schematic while you're looking at the image of the daughterboard below. In the south-east area of destop's schematic, you'll see the CN1 connector, which says CN-101_BOARD at the bottom. You need to make a note of what a connection is called, and what pin it's connected to. Then, refer to the daughterboard, find that pin number on the CN1 connector (the box with the green pads inside it), trace it to the flat flex connector, and find out which pin on that it's connected to. Then, refer to "FF18" in the south-east of my mainboard schematic. Make sure that the pin number you traced back to on the daughterboard matches with the NET NAME of the original signal you were tracing from destop's schematic. For example, let's say we're checking the DID6 signal. If we look at destop's schematic, we'll see that DID6 is pin 19 on the CN1 connector. Keep that in mind. Now, look at the daughterboard. PAY CLOSE ATTENTION TO THE ORDER OF THE PINS ON THE CN1 CONNECTOR! As you should be able to figure out from the silkscreen, the TOP pin in the FAR LEFT column is pin one. The one below that is Pin 2. This means pin 3 is the TOP pin in the second column, and pin 4 is below that... etc. So, if we count over to pin 19 (top row, 5th pin from the left) and trace the connection back to the flat flex connector, we'll see that it connects to pin 14. Now refer to my mainboard schematic. If you still remember at this point, the original signal we were tracing was DID6. If you look at the FF18 connector in the south-east corner, you will see that pin 14 is indeed labeled as DID6! This connection has been confirmed to be accurate!

As you can see, it's kind of complicated, which is why I'd like someone other than myself to confirm its accuracy. Is anyone up for the task?

Oh, and ignore the 5V and GND lines. They're connected by a separate connector, seen in the lower-right hand corner on the PCB. This will change somewhat in future designs. Okay, so here's the daughterboard, finally.

So, how about it? How does everything look? Marshall, do you think you'll be able to get your friends to look this over? And Splode, do you think your friend can get a couple boards made up?

I don't care if I talk a lot. It's important information, isn't it...? At least I didn't ramble this time... much...

gamecube6 · Post by **gamecube6** » Sat Oct 11, 2008 8:29 am

*twitch*

No, You didn't ramble... It's just a little scary. (Way over my head)

Good luck on this, It'd be great to see a GCp with no CD drive!

Electric Rain · Post by **Electric Rain** » Sat Oct 11, 2008 1:51 pm

Oh, come on! Anyone can follow lines, right...?

ShockSlayer · Post by **ShockSlayer** » Sat Oct 11, 2008 1:57 pm

Epic man, epic. The fact that you were able to start from shiznik and get here is incredible and fantastic. I wish I could help, because then I could be a part of what can be considered one of the greatest achievements coming here from forums.benheck.com!

Oh, and I'm still down for buying one of these later, m'kay?

SS

Electric Rain · Post by **Electric Rain** » Sat Oct 11, 2008 7:46 pm

I'd like everyone to know that I'm not "waiting" on anything anymore. The parts are here, I have my new tools... all I need now is the PCB, but I would like at least one additional pair of eyes to look over this and say "yeah, everything looks cool." Basically, I'm waiting on you guys. And Splode, I suppose. I'm not rushing, I'm just letting everyone know that it's kinda in your hands at the moment...

By the way, you'll probably notice that some of the silkscreen for the part outlines is going to get cut off because of the vias. I'll probably fix this by drawing boxes over top of the silkscreen on the vrestrict layer and running the auto-router again, but it's not like it matters much. It's just aesthetics, and it's only the prototype, after all.

Oh, and thanks for the kind words, SS.

Post by **marshallh** » Sat Oct 11, 2008 9:20 pm

Nice work. I can't see anything obviously wrong. The most important thing is whether or not you paid close attention as you were making the device footprints and pinouts.

If I were doing it, I would make a larger proto PCB. I'd seperate each part of the design into its own space to make finding errors easier (not only when you're poking around with a multimeter but also when referring to the original board design)

Also, a tip on the decoupling caps: put them on the backside of the major chips (put 4 caps behind the FPGA in a square pattern, also some behind the FX2). It's important to keep them as close as possible to the power inputs on the chip.