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

[Electric Rain]

Thanks Marshall. I didn't expect you to take the time to look it over.

I took my time with the footprints... The USB connector was the hardest one, since it has winged feet, two posts (needs holes in the PCB), and an oddly-shaped keepout area (because the bottom of the connector is metal). The mechanical drawing of it on the datasheet was a little overwhelming. When I got the parts from Digikey, I placed the important ones (not the resistors or capacitors... too small) on the board to make sure all my footprints lined up. Somehow, I actually got it all right! Everything was exactly where it was supposed to be, and I never even took geometry (homeschooled)!

Like I said, if Sparkfun/BatchPCB wasn't charging per square inch, (most fab houses charge XXX for 0-10 square inches, I've noticed) I would definitely make the board larger. But, it shouldn't be too bad. Based on the printout of my board, it seems perfectly workable.

Thanks for the suggestion about the decoupling caps. But, I don't think I can do that with my current design. This is my bad for not clarifying how everything's laid out, but I have to keep components off the back-side of the board. The IDE connector is right-angle through-hole, and is the only component mounted on the back side of the board. But since it's the only through-hole component on the entire board, this means that all solder-points are still on the top side. The IDE connector faces "up" when looking at the board in 2D. This design will allow the board to connect directly to a 2.5" hard drive without the need for any ribbon cable. This cuts down on cost and also saves space. However, it's a tight fit... because of the thickness of the hard drive and the short length of the IDE connector's pins, the board will end up being almost completely flush to the back side of the hard drive, so I won't have any room for parts there. The "standoffs" on the hard drive protrude a fraction of a millimeter above the bottom of the HDD's exposed PCB, and this should be enough to provide separation of traces and vias, but possibly not enough to squeeze parts on the back of the board. I will explain this better and with visual aids in my next blog. Just know that none of the decoupling caps are further than a couple pins away from the ground inputs of the chips (should I have put them closer to the V+ inputs, or doesn't it matter?). You can see this yourself if you open the Eagle PCB. This was a top priority in my new design of the board.

Thanks a lot for looking it over, Marshall. Do you think you can get anyone else to look at it, or don't you think it's necessary? Oh, and what do you think of my overly-complicated daughterboard method?

[Dark-Aries]

I tried to look it over...but got lost....

But this means again..maybe we can all buy ourselves xmas presents...and you can get a nice big fat check and internet stardom and fandom...


With great power...comes no responibilty on the interwebs.

[Electric Rain]

NO CHECK!!!

Okay, I don't see this going much further... I guess I've overestimated everyone's abilities here. But, that sounds rude. Let's instead say that I've underestimated my own. Then, I guess I'm just waiting on Marshall and Splode.

Edit: ...I have a bad feeling about Splode. I think the NNN may have gotten him. He hasn't posted for a while, but I sent him a PM last night, so we'll see how that goes...

[Ness]

That was n64, not gamecube. I wanna get my hands on one....

Um...it was definitely Gamecube, not N64. Unless Nintendo made a copy of Twilight Princess for the N64 and modded a Gamecube controller to play on the N64 as well...it was definitely a Gamecube.

[Electric Rain]

That was n64, not gamecube. I wanna get my hands on one....

Um...it was definitely Gamecube, not N64. Unless Nintendo made a copy of Twilight Princess for the N64 and modded a Gamecube controller to play on the N64 as well...it was definitely a Gamecube.

I would be curious to know how those units work... But, it won't matter soon, I suppose.

Oh, and I'm withdrawing my request for assistance, by the way. Nobody seems to want to help, so I'm not going to ask anymore. I just finished comparing schematics again, and everything looks perfect, as it always has. So, none of you need to feel obligated to help anymore. Splode is MIA, so I'm ignoring his offer as well. I'm too impatient, so I'm going to go ahead and queue up my board with Sparkfun sometime in the next few days. Still don't have my cameras with me, so the second video blog is still on hold. I'll keep you guys posted.

[collinE]

Dude, I know you have the best intentions and didn't mean to make it sound bad, but you did. It may be true that you underestimated your electronic prowess or overestimated our abilities, but you didn't have to say it. That was insulting

Most of us, if not all of us, could not do from scratch what you just did. It's amazing to me. As of my post you have 47,423 views on this topic. We care and we want to see it happen, but we can't help you.

But seriously, congrats on your crazy circuitry. It would be a lot of hard work before I got to that point.

[Electric Rain]

...Sorry... I've tried to involve everyone here as much as I could, but I am only now realizing that Marshall was right when he stated I would be better off to approach this solo. Sorry if I insulted anyone... I hope you'll all still buy from me when I get to that point.

I don't think I can understand why anyone couldn't compare schematics, regardless of their electronic knowledge... but, it doesn't matter anymore. I checked it over very carefully this last time, so I'm 100% positive now that my schematic is consistent with the original.

And you're right, this has become one of the most successful (in terms of post and view counts) threads in the history of these forums, and I appreciate everyone that has helped make it so. Thanks for sticking with me through all my ups and downs, guys.

[Biolriz]

Hello there!
I've been following this thread a little while now and i was hoping to see some good results soon

I think i can help you out a bit with your PCB..

Is the pics you posted a true pic of the current pcb?
if so, how are you thinking to etch it?
To etch so small traces, you might need access to special machines wich probably nobody on this forum have access to..

If you are going to etch it "hobby-wise", you'll most likely need to aim for a bigger pcb surface.

I see a handfull of "errors" on those pictures.
with "errors" i mean that much of it can and will lead to shorts.

What program did you use to make that pcb?
Did you use Auto-route function?

What experience do you have in electronics, soldering, smd soldering, and etching?

Sorry for all those questions, but if i'm going to help you out, i'll need answers on those :p

And sorry if my english is hard to read (it's not my 1th language )

[Electric Rain]

What program did you use to make that pcb?
Did you use Auto-route function?

What experience do you have in electronics, soldering, smd soldering, and etching?

Eagle, yes (I used ELECTRA), basic, moderate, little, none... respectively.

I've looked into DIY etching a lot, but I'm not going to be etching the prototype myself. I'm using BatchPCB, and I was actually about to put through my order today... but, what are these "errors" you're speaking of?

[ShockSlayer]

What program did you use to make that pcb?
Did you use Auto-route function?

What experience do you have in electronics, soldering, smd soldering, and etching?

Eagle, yes (I used ELECTRA), basic, moderate, little, none... respectively.

I've looked into DIY etching a lot, but I'm not going to be etching the prototype myself. I'm using BatchPCB, and I was actually about to put through my order today... but, what are these "errors" you're speaking of?

Ya know, I actually looked at it for a while, and I couldnt find any horts. But then again, I dont understand half of it, but I do understand what a short would look like! Hey, I TRIED.

Hah, ER, we all have faith in ya!

SS

[marshallh]

If you are going to etch it "hobby-wise", you'll most likely need to aim for a bigger pcb surface.

I see a handfull of "errors" on those pictures.
with "errors" i mean that much of it can and will lead to shorts.

He's using a commercial board-making services with CNC equipment. No issues with production there.

The shorts you are seeing are because of the raster space the images are mapped out onto. Think of the lines touching each other as a rounding error in math.

Rain, if you want to be 100% sure there's no problems, look at the top of the first board image, look at the top right corner of the PROM. Zoom in with Eagle and make sure that's within tolerances. Also check the top right corner of the FPGA. That's the only things I can see. And I doubt they're actually problems.

When I post board layouts, I export the image at twice the resolution and resample in my image editor, so there's less of the lines touching stuff.

[Biolriz]

Ok, good to hear!

Well, i was looking over it again today but i didn't find what i thought i saw yesterday.. maybe i was just too tired. :p

Anyway, i have marked some stuff that i got annoyed of and some traces that i think is a bit too close and one small error.
And what's up with the thing i've marked with a question mark ?



Are you going to solder the smd's yourself or use some service there too?

I'm just trying to help, don't run me down please
I shure will buy a couple from you when you are done!

[Electric Rain]

Marshall is correct. The oddities you're seeing are due to the rasterization of the original vector layout. The clearance and such were all set in the design rules before running ELECTRA, so there are no errors there, whether it looks like there are or not. You can see this a lot better (in vector form) if you open the Eagle file... but, I'll make it a bit easier, now that I actually have someone that wants to look it over.

I'm re-designing it, yet again... the traces are just too funny looking, even though it checks out electrically and mechanically (spacing). I'll upload the new design as a PDF when I'm done, which preserves the vector format.

Edit: Done. It's not perfect, but it looks a little better. The craziest of the crazy traces have been taken care of, but ELECTRA still has its quirks. Nevertheless, everything is still making good contact with 8mil clearance and (nearly) as short as possible traces. PDF (221KB)

The traces are still vectors, so you can zoom in as far as you want (try about 600%) and they'll only get clearer. You'll notice I didn't add silkscreen yet... I had to keep re-doing it every time I re-route because the vias end up in different spots, so I'm holding off on adding it back in until I'm sure I have the final design. What do you guys think? Is it good enough?

[Biolriz]

Now That's looks Very good!
Can you give me a link to the datasheets of the components you are going to use?
That is so that i can verify that everything goes where it's supposed to go.

I can see the light in the end of the tunnel is brighter now

[Electric Rain]

I've updated my schematic a bit. PDF (42KB) I did this mostly just to show the actual values of the passive components.

These are shortcuts to the Digikey pages for every part I ordered for this project. ZIP (6KB)

I should have noted these earlier, but there are some minor differences between my schematic and Destop's that should be taken into account when comparing.

1. Destop's board was designed with a standard IDE connector that was meant to be connected to a 3.5" HDD with an IDE cable. These are 40 pin connectors. I designed mine with the smaller 44 pin IDE connectors that are used for 2.5" (laptop) hard drives. The four additional pins, 41, 42, 43 and 44, are for 5V, 5V, GND, and Type, respectively. Ignore "type." Nobody knows what it even does.

2. Related to #1, because Destop's was designed for use with a 3.5" HDD, a separate power connection was also needed. But, with 2.5" HDDs, there is no separate power connection; that's part of what the additional 4 pins are for. So, the power for the HDD that will connect to MY board is supplied THOUGH the board, via the PWR connector. This is also the reason for the larger trace that runs from the PWR connector to the IDE connector. (You'll see this in the south-east corner of the PCB.) Because the power is supplied via the board, there is a small addition as well. A third 100uF smoothing capacitor has been added next to the 5V input on the IDE connector. It's probably not needed, but I figured I'd throw it in for good measure. Everything else on the board had a smoothing cap, you know?

3. Obviously there's the whole flat flex connector daughterboard thing, but you guys already know what's up with that.

4. On the far east end of Destop's schematic, there's an LED with a current limiting resistor... this is totally unnecessary. I don't know why Destop needed this, but I left it out.

5. There are a few jumpers on the board, labeled "JMP_EXTP", "JMP_USBP" and "JMP_GC_POWER." These are here to provide three different options of power to the board. I have taken them off for several reasons, but they will take far too long to explain now. Future boards will probably handle power a bit differently, but this prototype will use a separate power connector, seen in the south-east corner of the board and the north-east corner of the schematic.

6. There are two headers in the original schematic... a debug header and a JTAG header. These have been replaced with simple "test points" seen on the north end of the board, encompassed in a rectangle which is on Eagle's vrestrict layer. (I put the vrestrict layer there to prevent the autorouter from placing vias in that area, which may otherwise be confusing when intermingled with the test points.) I did this to save space and money. Which of these test points are actually needed remains to be seen. For example, the JTAG points may not be needed at all, since Destop wrote a program that allows the PROM to be programmed through the FX2 via "virtual JTAG" software. But, some people on Dextrose had problems with this, so the JTAG points may still be needed as a backup, thus their existence on the prototype and hopefully the prototype only. Most of these points will not likely be on future revisions of the board.

7. There are a few unlabeled (without values) parts in Destop's schematic. This includes two caps just north-west of the FX2, and two resistors to the east of the FX2. Common electronic sense says that the caps are just for smoothing, like every other cap on the board. So, I'm assuming a value of .1uF for these. The resistors were a little bit trickier. I had no idea what these were supposed to be at first, so I resorted to the FX2's datasheet. Luckily, I found out that these were supposed to be 2.2K, so, they are now labeled as such on my schematic.

Other than some minor differences in the way I chose to route the wires in my schematic, that's about it. I figured it might be a bit confusing if I didn't detail these changes... so...


I want to order the board ASAP, but now I want to wait for you to check over everything you wanted to check over. Let me know when you're satisfied, and thanks a lot!