Worklog SparkleBear's First Portable Nintendo 64 Build Project Worklog.

[SparkleBear]

Hello all.

It feels very auspicious for me to be starting this project with the portable n64 build. My interest in building one of these is the first thing that drew me to the BitBuilt community but went down the wii rabbit hole for a while before finding myself here; much more well equipped to take on a project like the portable N64. There are so many more resources and open source projects for the wii, which is convienent and fun but I am looking forward to the necessity of figuring a lot of things out on my own here.

Here are some of the design goals I want to incorporate into this build:

- N64 trimmed motherboard with PIF relocation. (Thanks to @Gman 's guide on that)
-N64 Advance VGA board.
-No cartridge slot - Internal everdrive x7 (potentially a sd switcher that allows mounting as a drive on computer and mounting on everdrive)
-FPX usb c power delivery charge and play.
- digital audio with gman's n64amp
- powered by modified pms1
- 2x 4mb ram swap from 2x expansion paks to rev05 n64 mobo.
- 4x 18640 batteries for long play sessions.
- custom designed pcb for n64 controller chip and joystick attiny24 micro controller for encoder emulation. (never used kicad before)
- design my first case in fusion 360 (never designed one before)
- Potentially purchase a 3d printer and print my own custom case.
- Cooling fan? or potentially passively cooled. Not sure yet.
- resin cast buttons.

This will be the most ambitions portable project i've taken on yet... very excited for the journey.

Purchased some n64 mobo off of ebay and have been studying the guide... Here we go!

This lot below is untested and likely wont all be working... But I'm sure I can piece together one working unit

A bunch of stuff already came in the mail before starting this log -- Already have the working n64 mobo, the 2 expansion paks for ram swap, the n64 advance vga board, n64 controller mobo, jumper paks...

Lets go!

 

[SparkleBear]

So after many hours of reading and studying about this trim, I finally started the process. Got a rev 05 mobo working and started testing... didnt take any photos of the un-altered board but you know what it looks like already! ;P

First step was make it naked and use my new hot air station to remove the 2x 2mb ram chips from the rev 05 board. @Gman recommended to use 350c temp setting on the hot air and not to force the chips at all... to take my time and they will slide off like hot knife in butter. It was quite satisfying.

Next, same process with the expansion paks. Remove the board from the enclosure and hot air desolder the chips. Slow and steady...

I cleaned up the traces and did some drag soldering with the amazing amtech flux that was recommended by so many. (really great flux)
And got the chips soldered in... Time to test it with a jumper pak. Removed the case for the jumper and plugged it in.

At first the n64 WOULDNT BOOT! My heart dropped a bit.. but I went back with some more flux and touched up the pins on the traces with a bit of fresh solder...
And gratefully it booted right up! Woot!

Alright! Progress...

 

[SparkleBear]

Next here, I decided to do some jumper pak trimming shenanigans. Unfortunately, I took it a bit too fast with the trim of my first jumper pak and destroyed it... woops!
But luckily I had a second one as a backup.

At first i was going to try to just use a trimmed jumper and design around that height. This was Gman's advice... but I decided to risk it and do some jumper pak relocation attempts. On this thread, I thought of wiring it right to the vias to save space... so I mapped them out per pin here:

But after thinking about it more, I was not sure if the different length wires would mess with some delicate impedance dance that was important for the ram to work properly... So, I ultimately decided to wire straight from pins to the traces instead....... two hours and 36 tiny wires later:

I quadruple checked that there were no shorts and touched up any potential ugly joints and prayed that the board would boot! SUCCESS!

Now to think about how to cool the chips before folding the jumper over and moving forward.

So thankful its been going smoothly so far.

 

[SparkleBear]

Next day, I added heat sinks to the big chips and a cut a copper plate for the ram chips. Learned about ShockSlayer's method of 4 drops of super glue in corners and one drop of thermal paste in the middle. This is an alternative to the more permanent thermal adhesive... Thanks @YveltalGriffin for that tip, and for being a stellar dude. You are generous with your knowledge and that is awesome.

With the heat sinks and copper plate installed, i decided that if i was going to fold over the jumper pak to lay flat on the ram I would want to treat the copper as to protect from shorts. I used some conformal coating circuit sealer stuff that I got at the local electronics store. Laid a thick layer down and let it dry before moving on.

I carefully folded over the jumper pak making sure the wires made a smooth curvy bend. Tested and it still booted up. Now, secured it down to the coated copper plate with hot glue. I also drew in some trim preliminary lines.

Tidy'd up the magnet wire, tested to ensure booting then protected and reinforced joints with carefully placed hot glue around wires and joints. Used hot air station after laying down glue with glue gun to make uniform and look nicer. I also scraped off the conformal coating on the top of the copper plate as i imagine this would help with head dissipation more. I may also add a slice of heat sink to that section of the copper plate.

Did one final boot test before trim... all good. Now prepped final trim lines and removed components from the trim path. Also removed DD port from bottom and cartridge slot. Alright, here we go!

Looks good to me but I don't have a way of testing the trim until the PIF boards arrive from JLCPCB and get the modified PMS1 that @Ethan built for me up there in ol' Canadia (thanks buddy)... Once those come in, I can build the pif breakout board, and run the relocation wires... THEN I can test this bad boy out. So for now, here it rests.

So far, big thanks to @Militaryman_20 for helping me understand ram swap logistics, @YveltalGriffin for help with understanding the ram bus, @Gman for advice on trimming and desoldering ram, @Ethan for general support in the process and making me the power supply board I will use...

Really stoked on this project... long time coming! Stay tuned. Thanks y'all for checkin out the log.

CHEERS!

 

[Gman]

I'd reccomend 32awg wire for the cartridge slot those wires look like it will be awfully stiff

 

[SparkleBear]

Thanks @Gman ... 32ga Magnet wire? or some sort of minuscule stranded?

Still considering a pcb bend or rcp wiring depending on my case design.

 

[RoseDagger]

32ga Magnet wire? or some sort of minuscule stranded?

Magnet wire is the wire used for electromagnets and such, thin flexible wire with only a thin enameled layer of insulation so they can be spoled close together

Magnetic Wire | Electronic Components Distributor DigiKey

Magnetic Wire are in stock at DigiKey. Order Now! Magnetics - Transformer, Inductor Components ship same day

www.digikey.com

 

[SparkleBear]

I have many spools of magnet wire… I was just curious if I should use Magnet wire or a different type of wire (something stranded) for the cartridge slot relocation... thanks though @RoseDagger ...

I remember @Downing talking about how it’s important to use a more flexible wire for that relocation in one of his videos. I could see potentially using magnet wire at the end when I’m about to button it up… But in testing when everything is loose I feel like they’ll just break

 

[wiiplayer]

very cool

 

[SparkleBear]

Today, I picked up my PIF Breakout boards from JLCPCB, and the FPX / PMS1 / 64AMP from Mr Pickles (@Ethan )... So I will finally be able to start the PIF relocation and attempt to boot the trim.

Here is a snapshot of the BOM for the pif relocation:

I had to harvest the capacitors from a scrap wii because I couldn't find my reels of SMD caps. *sigh* ... Turns out C80 is .1uf (100nf) and C290 is 1uf. Might be helpful info for anyone else who might be trying to populate the pif board in the future.

Resistors and Capacitors to the breakout board first... Didn't have P1-4 750R pull up resistors, @Gman recommended to go with 1kohm as a substitute.

Then the chips next. I would tack down one corner, line up the rest of the pins, then use flux and drag solder those puppies down. Then I would reflow everything with hot air station just to make it pretty and make all the joints are shiny and proper.

Hot glued the PIF Breakout board over the X1&2 portion of motherboard. Wanted to get it nice and secure before I start the relocation wiring...

Patient wiring of all the points for PIF relocation from the Advanced N64 Trim guide. Before wiring up voltages I tested the PMS1 that Pickles (@Ethan ) made for me. Button from BTN to ground, wired in a single 18650 to B+ and GND, Clicked button... voltages tested as expected; All checked out with 3v3 and 5v. Wired up LED and FPX charge board to CHG and GND. Charges at 12v with PD charger. (I might trim off the unneeded/unpopulated regulator section at that blue line I drew... but might not if not needed...) I also added a few rows of the heat sink to the copper plate for ram cooling.

I disconnected the battery ground wire before moving on. Now to wire up the voltage pads from PMS to N64. When I glued down the breakout board, I covered up the chip where I would have to wire the third 3v3 point... woops. I scraped the mask off of the plane that originally would've carried the 3v3 to make a solder point for the final voltage 3v3 point (you can see it below the breakout board. )

Definitely almost forgot the special notes gman left us to make sure PIF 6 is connected to three points (rcp, cart, CPU) and pif 7 to CPU and cartridge pin. I opted to wire directly to CPU rather than run the wire on the back of the board. Wired up 5v too. I had already soldered wires for a cartridge slot that are way too thick for the final build but wanted to have an interim way to test cartridge before soldering to the everdrive PCB. I will eventually use 32awg magnet wire for this job. Yahhh... these are WAYY too thick.
Didn't take another photo until after I wired up the screen and a controller... TRIPLE CHECKED for any dead shorts between voltage rails and ground. Ok, the moment of truth........... aaaaaannnnndddddd....

It wouldn't boot... Checked voltages coming out of PMS and no voltages were present. I thought I either fried the board OR it went into shipping mode accidentally somehow. (that happened to me before in a build) And luckily it was just the shipping mode... ( to get out of shipping mode either disconnect the battery ground and reconnect OR connect unit to charger.) However, after getting the PMS outputting proper voltages and tested the voltages at the board the N64 would still not boot... I did see the screen flicker but no signal would come thorugh. I will be going over the wiring tomorrow in the daylight and troubleshooting this stage. Hopefully I can get it to boot tomorrow. My hunch is sketchy cartridge slot wiring. I double checked all the wiring with PIF relocation and I was sure there were no shorts on the voltage lines. I'll just check all my work tomorrow and probably rewire the cart slot...

Here is a close up of the PIF Relocation wiring.

(I also accidentally got a little thermal paste on the ram pins, would that mess up the boot process?)

If anyone sees something fishy in the setup, please let me know. I'll report back soon after some investigation.

 

[SparkleBear]

Upon waking up this morning I immediately began my investigation. I discovered that I accidentally wired the cartridge with the top and bottom row swapped!!! That is atleast one reason why it was not booting. I thought I may have blown my board and my everdrive. I tested my ED on my home console and thankfully it was working. That gave me hope that my trim was not busted either. I inspected the pinout of the cartridge slot and am so thankful for the horizontal symmetry. Even if I swapped top and bottom row, the 3v and ground pins are the same. PHEW!

I also discovered a flat damaged pin and replaced it with one of the unused pins (24/49). I desoldered all the wires from the cart slot and wired it up properly. When I next powered it up I thought it was not booting but I had a hunch it was actually just the screen/driver board that was messed up... I stripped an unused RCA cable and soldered it to the AV chip at the composite pad. To my excitement it was actually booting... I put that screen away for further study for now.

Hurray! I then wired up one of those 5" screens I have saved for a gwii build. Looks good!

I caught mario's wink of encouragement... he knows I dodged a bullet there. Lucky me!

Now, I am planning out where to attach the everdrive, the 64amp, and the n64 advance vga board... I think right above the jumper pak would be a tidy place for the 64amp. The everdrive pcb will be glued to the back of the n64 and rewired with 32ga magnet wire for a permanent installation. I think there is just enough room for the vga board to be secured underneath it. My goal is to have everything as a tidy little bundle so I can take measurements and start to design my case.

The backside of the everdrive is bare enough for some good hot glue anchors!

To-do list:
-- Joystick converter solution... potentially a custom PCB for that and the controller chip.
-- buy speakers and headphone jack to work with 64 amp.
-- Learn fusion 360 enough to build the case. (potentially buy a 3d printer too)
-- decide if I want to relocate the SD card for external access or just leave it inside... or do the SD switcher with access through the USB C port.
-- potentially do one of those ashen board bends to make the portable internals smaller...


Anyway, thanks for checking out the thread. Thanks for all the support and guidance everyone. Another thanks to @Ethan the Pickle lord... helping discover the secrets of the PMS. Special thanks to @Gman as always, supportive and informative, like a G. I'm learning so much and its very exciting.

 

[SparkleBear]

The Joystick PCBs came in, surprisingly. As did my mouser order. However, I ordered the wrong ATTINY24 format... got the little square chip but I need the SOIC-14 format... whoops! Also, seems Micro's v3 joystick converter boards need something a little different size than the gamecube box. I ordered a few different ones that look similar in hopes one of them will fit. I found a european company that would sell me the exact one but it was like $35 in just shipping... Not worth it for a few joystick boxes. Ordered the correct attiny24a from mouser too.

Now that the trim is confirmed working from yesterday's tests, I desoldered all the voltage and video lines to plan out some placement work. I laid the everdrive pcb onto the back of the n64 board and planned out the cartridge soldering relocation. I plan to trim off a bit of the pins and a few angles to match the shapes of the n64 board. Its just too late at night to run the rotary tool in the neighborhood. Also, this will allow the VGA board to fit perfectly on the back of the board too! Puzzle Snuggles PCBs

Decided that I will go ahead and mount the 64AMP right above the bottom of the jumper pak. Fits nicely and allows a short run from power, ground and the 3 digital lines coming off of the RCP. Tidy'd up the wires the way I like 'em. Satisfaction...

I test fit the long advanced flex cable and decided the short little breakout will fit this build better... I think those are designed more for home console installations. The little breakout will do just fine. I soldered the flex down to the RCP... shiny! Ran the appropriate lines to the PIF board... ran the CTRL line directly to the pin so it wouldn't be a pain in my butt when I solder in the controller line down the road.

I didn't double stick tape the flex down incase I need to troubleshoot the Digital Audio lines... Also, I may desolder the AV chip after I confirm the VGA is working just to make it cleaner. But I also may want to just keep it so I can A/B composite and VGA... for funzies.

Here is the overview of the board in its current state:

Gonna hit up the electronics store tomorrow and grab some little speakers or a headphone jack to test out the digital audio. I will wire up the VGA and test out the whole system. Trim the everdrive down a smidge and wire it up to this bad boy... Definitely going to take extra measures to be sure there are no shorts between the pcbs with either conformal coating and/or kapton...

FUN!

 

[Gman]

Nice work and great worklog too. A joy to read.

 

[SparkleBear]

Trimmed down the EverDrive... Made a sweet little angle on the side and cut off the development usb port on the battery side. I also trimmed down half of length the pins on the everdrive. I really wanted it to fit behind the n64 in a sleek way; so I followed the edges and angles of the trim. This also allowed more space for the VGA board bach there. Taped the ED down to a cartridge half and the n64 mobo down to even height to allow for even and secure wiring. 32awg magnet wire for this job then i'll fold it over. Wired up the top row to the back of the cartridge pins. Soldered the bottom row of n64 cartridge pins using wires slightly longer than the top row and fed them inbetween each space... folded the board over, curled the wires around and soldered the front of the cartridge.

After finishing the cart slot I forgot to check for shorts between voltage lines before powering up (big oversight). I tried booting but nothing happened. Thought it might be in shipping mode again... tried reset. Still no boot... THEN, I checked for shorts. and DEAD SHORT with 3v3 and ground. Tracked it down and Ended up accidentally leaving a short between the layers of the everdrive.. Should've learned better from my Wii lessons. I thought I fried something. Turns out Luckily, the PMS's battery was dead so that is why it wouldn't turn on. I thought I actually blew up the PMS... but no, it was a dead battery. (However, in trying to figure this out i actually DID fry the BQ chip... lol) But thankfully, the everdrive and the n64 were un-harmed in the debacle. Ended up kinda freaking out for a moment there. Thanks to @Gman for talking me off the ledge. ;P
In the moment he spoke words of wisdom... "Take it easy. Fix it right and no need to worry until then" Its in those moments when you think things are going so wrong that I need to be extra sharp and calm. (Thanks G. I learn a ton from you dude.)

After tracking down the short was in the ED, I desoldered all the connections and sanded it until the short went away. I didnt want to spend another 30 minutes soldering a potentially busted ED back into a potentially busted 64; so I ended up re-attaching a cart slot for testing. SM64 booted to a huge sigh of relief; the n64 is OK! With some fiddly half in half out insertion of the ED I ended up finding the sweetspot where it would boot too. Wow, beyond relieved that the two boards are OK!! RIP to the BQ, gave its life for my education and liberation. Thanks little chip...

Later that night, I got the everdrive soldered properly to the board and tested the build out with a spare pms I have lying around... Tested for shorts. Booted it up... Working! YES!

I was so hyped I stayed up til 3am installing the VGA board. Decided to run all the wires from the flex back towards the rcp and down the wire high way... through the mounting hole and up to the other side where I hot glued the n64 Advance VGA board. I also ended up deciding to use the output flex cable for general strain relief and i guess it has filters on it... It also makes it really clean. Still determining how to route it for the final build. so this will actually work out for the best... Flexible choices, flexible cable.

I wired the RGB and V/H Sync lines right from the end of this FPC... there were no dedicated outs for H and V at the end of the FPC so i tracked them down to the pins on that IC. Made twisted pairs for the RGB lines and ran them to the driver board. I also used a wii wifi wire to hook up 5v for the VGA board, this goes from the 10th pad to the 5v voltage line, I happened to use the driver board's 5v line but any would've worked. I forgot this line at first, but it wouldn't display video until I connected this line.

So alas, in the late late hours, a VGA enabled n64 trimmed motherboard boots into the silence of a September night...

Didn't get to test out audio yet... the troubleshooting marathon was enough to keep me up til 3:30am... but I could rest easy knowing the project is on track.

Cheers all!

 

[SparkleBear]

Hey all,

I've been running into issues with inconsistent p1 controller data problems. Sometimes it works, sometimes it doesn't. Right now it DOESN'T... sigh...

I've re-soldered different sets of wires and tried different controllers and controller ports. I am measuring perfect continuity to the PIF chip on data line as well as measuring 3v between the voltage and ground pins on the controller port.

Anyone have ideas about what to check to troubleshoot this issue? Thanks so much!

 

[ShockSlayer]

Try disconnecting the RGB board(and probably leave it disconnected until you can get the controller stable again.) If it's still being troublesome shorten any wires that you can, and clean up the affected areas on the off chance there's some loose/lodged junk. Lastly, you could try swapping the PIF or the lower chip itself if you have another one handy.

 

[SparkleBear]

Try disconnecting the RGB board(and probably leave it disconnected until you can get the controller stable again.) If it's still being troublesome shorten any wires that you can, and clean up the affected areas on the off chance there's some loose/lodged junk. Lastly, you could try swapping the PIF or the lower chip itself if you have another one handy.

Thanks mate! It was the loose junk/flux... I was starting to prep another pif breakout board... but discovered we could just clean it to make it work! A good ipa wash down got it sorted. Both you and @Gman had the "clean that up" message. That turned out to be the trick!

Before and after:

Thanks boys. All is working as it should again... Moral of the story. Keep it clean!

 

[SparkleBear]

Alright fellas, the ghost in the machine came back... I spoke too soon. The controller behavior is very strange indeed... It seems that when I cleaned the pif board that first time and it worked I jumped to the conclusion that it was the flux... then a few minutes later the controller stopped working. Gunnar said it could potentially because the controller wire was too long... so I took out my spare controller motherboard, added a header left over from the BB Screen board and attached it to the controller board so it would be easy to take on and off for testing. (I also added a header to the original female controller port.) These shorter wire runs did not work either... After lots of other unsuccesful troubleshooting steps, including adding an extra ground wire between pms and n64 and touching up all the solder points, I discovered something strange and potentially useful for solving this riddle; the controller works when the area around p1 is moist with the IPA... don't ask me why, but when is still evaporating, the controller data works and I can successfully control the n64... when it dries off completely -- it stops working again. After this discovery, I also tried soldering data wire directly to the pin also to no avail...

The header solutions:

Within the red circle is the magic wet zone:

I thought maybe, because of the moisture, there was a very high impedance connection being made between PIF 16 another pin in the circuit to allow the controller to work; so I found a 1.5Mohm resistor and probed between those pins and alas, the controller started working again.. I would remove the resistor and it would work for another few seconds and stop... I tested this a few more times with the same results. I read the pinout for the PIF and it goes to a pin on u8... this is not a way to fix the circuit but shows me that p1 data CAN work... so I will keep tracking down the issue... This gives me hope ;P

So curious...

 

[SparkleBear]

A few weeks ago I fried my PMS1. Woops.. I finally fixed it last night. Learned to program PIC chip (thanks to tip by @Aurelio : https://bitbuilt.net/forums/index.php?threads/how-to-program-your-gc-board.2541/)... Replaced the BQ and bam... working pms1 again. This unit was only populated for 3v3 and 5v. So, I trimmed down the board after studying the gerber... There is a voltage bus that basically splits the board in half (well... thirds) on the backside. As long as I stayed away from cutting this line it should work fine!

Taped it down and cut it with the back side showing so I was clear where my limits were... Cleaned up the edge with fine grade file and some sand paper. I will also seal that edge with conformal coating.

Tested for shorts, wired up battery and checked voltage lines... All good. Wired up n64 and powered right up

I lost the pads for the LED but i followed the traces to the Resistors and wired to the end of those for identical functionality of LED. However, the led is always too bright for me so I stuck a 7.5k ohm resistor between the vsys power line and the led. That made it the right brightness for me.. atleast for now while in testing mode.

Wrapped the whole deal in kapton until the day when it can be mounted in a case... don't want any more fried boards. ;P

I had been using a spare PMS2, so Im glad I finished the dedicated power supply for this build. One day the 64pms will come out again... but this will do for me for now.

 

[Gman]

you wired the leds to wrong side of the resistors effectively full voltage across the leds