Bacteria project: N64³: --abandoned--

[ShockSlayer]

From my personal experience, soldering to through holes never worked out quite well for me. Don't remember why, but it is definitely something worth mentioning, as shorter wires = less resistance = faster data transfer(which might be what you need), and nothing beats the original traces!

SS

[bacteria]

Thanks for the advice guys.

Two options I see; either wire between traces directly and keep wires very short; or use longer wires to connect component to component. Which is best?

[Tibia]

Well, to me, the problem looks like it's coming from the different resistance values of the changing distance. I'd be inclined to run the wire from component to component. You'd be able to attempt to keep the same distance, and perhaps even trim a bit more of the board itself away.

[bacteria]

Interesting thought.

Think what I will do is require component to component and see whether the system works or not. No point wasting another board if this one can be rigged.

[palmertech]

Assuming, of course, it was not fried.

I think the "cut pin-by-pin" is the best idea.

[bacteria]

I have discovered from my continued working on the TurboGrafx that it really isn't just a case of A connects to B to C so wire A to C to cut out B. These traces can sometimes be temperamental about wire length and also lead you to false paths; for example, you check one connection (eg a pin on a game cart) and find the end pin it connects to on a chip; however, sometimes the wire junctions to another connection, so whilst the contact remains to that point also, the connection hasn't gone in the right order. Sometimes it is obvious when a trace whips around a board, in and out holes, until it gets to another component; however, sometimes en-route the connection branches off - that is what I am talking about. That makes rewiring boards far more complex than it already is (especially if you have cut a trace and the system doesn't work and you have to work out why! I have have that a few times on the TurboGrafx board, especially as I only have one such board and no schematics! ). The TurboGrafx card keeps throwing up wobblers, but I am getting it to work. All pins rewired, and I am slowly dremelling out a connection at a time from the game cart pin and testing the system to make sure it is still happy. It is the only way - one connection sometimes works in two places so cutting the original trace might disrupt the second connection to a different chip, stopping the system working. Other issue I get with the TurboGrafx is the capacitors - when soldering, I get tiny sparks on the chip pins; tiny shocks when my arm is on the board when I am soldering; sometimes having to leave the board a couple of minutes before testing so the system works! There are a lot of capacitors on the board and I don't know how to discharge them between tests - ok so far though, the board seems quite tolerant to modding (so far!).

I will do this approach with trim and rewiring the N64 board. I have learnt a lot about rewiring, thanks to the TurboGrafx board, without that experience I wouldn't stand a chance with the N64 board reduction. The N64 board is more complex but I am learning the principles and tricks now, which I can use to better effect on the N64 board.

I was optimistic before on my first board trim attempt, guess that is why it failed - only needs one mistake after all for board to die or not work. I will need to be very methodical indeed on my next attempt, and really take my time. I will start this after finishing trimming the TurboGrafx board down, as mentioned before.

Using a 1mm drill bit is great in the Dremel to not just expose copper traces from a board but also to completely destroy a trace for testing, and also of course in the process cut through part of the board. It is the way forward.

Here is what I mean - you can see the cuts.



To prove it works so far, pic below. BTW - screen was a PS2 screen for a PS2 original system, it has a stretched screen, image is not great but ok; I removed half of the casing, relocated the board inside the system, put in A/V ports - did this a while ago, as I can then use system as a test system (isn't good enough to use in a project, but handy for testing)

[bacteria]

Ok, here is the proof I can do it! I learnt a lot from this, quite "an education".

TurboGrafx board is cut down, fits in my case now.



Top down view, with my case on top, to prove the board fits inside the case ok:



When cutting a board, it is a good idea to use a sanding drum on a dremel to sand the sides down, as otherwise it is easy for cut traces to touch each other and make a short. Also, use some compressed air to blow debris off the board before turning system on.

Going to put the TurboGrafx board away for the time being, got a fair bit of work left on that yet to do: tomorrow I intend to make a start on rewiring a new N64 board. Wish me luck!

I have one last unmodded N64 to play with! I guess I will have to keep my eyes peeled at car boot sales for another one as a spare! The one I have I haggled the seller down to £5; not bad considering the console had an expansion pack inside - just checked, all works ok...tomorrow it gets gutted (literally!).

[themadhacker]

wow, you never give up, you just keep trying until you get it right!

[ShockSlayer]

I recommend you try trace to trace before component to component, it just makes more sense to me. You break a trace, then you reconnect it. That's pretty much the way "we" define relocation.

Oh, and instead of a dremel, try a razor. You'll be much more precise!

SS

[bacteria]

Using an engraver pointed cutter on the dremel to cut out a line in the board, just over about half way through the board (deep score), testing the traces are severed, then wiring and testing system.

Made some progress, progress slow. Tried to solder to the very thin traces but it was going to take forever as the N64 board resists this; so I am soldering to the through holes therefore as preference. So far, connected the 3v line to a capacitor, connected ground to two severed places and also three wiring connections (through through holes). System works; I am just wiring a couple of wires at a time, retesting, etc.

This will take many hours to complete!

I am cutting the board slightly smaller still; you see the line in the pic - doesn't perfectly match the line on the other side, but quite good. This time around, the board will end up slightly smaller than the PSone screen in length and breadth.

This really is the best path in the board - only look like three choices, the one I tried before, but there were jumpers and capacitors under the U4 chip which makes it not ideal; between the two main chips, but that is masses of really fine wiring; and even if you succeed, the board will be double thick as you have a chip under the board as well as above that requires cooling and heatsinks: and the path I am taking this time.

Anyway, progress so far; may have an updated pic later:



UPDATE - about 40% done on the underside now. Works fine at the moment. Probably won't have much project time until Monday now, will finish the back of the board then and then start the front.

[jjhammerstein]

Wow!
Never give up!

[Life of Brian]

I think this approach is a much better way to go about it. Keep it up!

[ShockSlayer]

Wow!
Never give up! Trust your instincts!

Fix'd. Much more appropriate for the n64, no?

I think this approach is a much better way to go about it. Keep it up!

I agree completely. This will make any possible troubleshooting much easier!

SS

[bacteria]

Indeed, rigging up a couple of wires, testing, moving on means you know precisely where the issue is. Important thing is to try and get all the grounds connected to as going along, otherwise when working on the other side of the board, components might stop working if there are any missed areas and connection breaks. I might have to cut parts of the board at an angle, as there is no easy straight path to follow along the board that doesn't run through something on the other side - working on it though!

[Triton]

the odds of this actually working are still slim to none at best. the frequencies and timing of the RDRAM (iirc) are EXTREMELY precise and sensitive to change. think about it, only a few people have successfully relocated the expansion pack to the board and you are talking about cutting off a full one third of the board, the third that has the clock generation for the RDram. best of luck but dont get your hopes up