Rpi portable in Wii U gamepad housing / reusing original parts

Hey Marius,

     First Pi project here as well and I will also be putting an RPi into a Wii U Gamepad. I will be using Emulation Station, and I hope to finish this summer so I can take it to my dorm in the fall. Anyways, I wanted to ask if we could keep in contact with each other during these projects so we can figure out what works best. Also, as for your question, I'm not too sure. I've read something about using the wires that are already there to connect to the pins in the Pi. Personally I will just be resodering the buttons and using a Teensy micro-controller. Hope we can find some success!

Hey Michael, glad you reached out! It's a great idea, it would make this project more instructive, and more fun too! Looks to me like we have the same goal here, haven't been able to find any blogs/post/tutorials of people putting an rpi into the gamepad (nor any general hacking of the system), which i find strange. Surely someone have done this before (cause its a freaking cool idea), but alas we got to figure this one out for our selves. Just imagine the possibilities: retropie running with emulation station as a portable retro console, then add touch screen digitizer to use as a tablet with another linux distro. Further down the road you could run limelight to stream pc titles to the gamepad just like a nvidia shield. Hook it up to your tv and a bt controller to get all the greatenes on a bigger screen, or use it as a streaming box with xbmc. Even use it as a gamepad on your pc. All in the same enclosure!

But one's got to start small, so stage 1 for me will be to figure out a way to put together a handheld console with es. I've enclosed a link to a mind map of my ambitious plan with some variations. Green circle is where i want to end, and then there are some alternatives if I hit a snag or dead end. Right now I'ven't ordered all the parts yet, so this is still early stages for me. Bought a AT070TN92/-3/-4 lcd off ebay yesterday, and also a wii u gamepad housing with original buttons, joysticks and speakers. Which means I still need a battery, charging circuit, audio amp, teensy (need to read both thumb sticks analog values simultaneously, much easier with a teensy I figure) and some more before I can really start.

But please, tell me about your goals/planned features for the gamepad (shoud we just call it the Pii U from now on? ). I'm afraid my electronic skills as well as my programming skills are all but mediocre, so my help will be limited. Still, I want to help you in any way I can.

Link: https://www.dropbox.com/s/22tlhbss17rlxil/Plan.pdf?dl=0

Cheers, Marius

Attachments:

Plan.pdf

Thanks for replying!

For my project I will just be emulating games, maybe moving on to further possibilities in the future, but I will go up to a PSX emulator (which the Pi2 can run) since the gamepad has 2 analog sticks.

I am also in pretty much just the planning stages of this project (I'm 16 so I need to snag a job to buy all the parts with, but such is life) though I have a list of materials on board.

-I know for certain that I'll be using the Gamepad's original battery (or perhaps the bigger one they sell separately) with a PowerBoost 500 from Adafruit. The PowerBoost is capable of running the electricity to the Pi and also recharge the battery.

-I haven't really looked into the teensy too much, I didn't think about the simultaneous joysticks...

-I'm going to have usb cables running to the bottom of the Gamepad (where the charging dock would be) and use small flash-drives (64gb each) to store ISO and ROM files on. This will make them easily accessible and loading new games would be easy.

-I won't have the Ethernet cable, or Hdmi at all accessible.

-Haven't looked into speakers much. I might just go with headphones only.

The hardest part of this project is definitely going to be fitting everything in the Gamepad, and that's probably why it has not been done before. If you've opened one up I'm sure you've seen how the battery pack really limits space inside the case, so I'm a bit worried about that. I'm going to buy all the parts and a shell of a Gamepad before I actually purchase a functional Gamepad, just to make sure things all fit inside.

And yes, my skills are pretty limited too. I'm only 16 so my money is also quite limited, but I'll send you updates every once and a while just so we can avoid making stupid mistakes.

FYI, you can do some calculations based on battery capacity and  current draw.

With one core running the PI2 is said to draw about 280mA. (see: http://raspi.tv/2015/raspberry-pi2-power-and-performance-measurement ) But this would be at 5.2V and has to be converted from the 3.7V average battery-voltage. At 90% efficiency, you'll draw 1.56 times more current from the battery: 437mA.

With an extended capacity battery of 2550 mAh, that would mean 5.8 hours of running time. Hmm. Not too shabby. My gut feeling was that I'd get something on the order of half an hour.

Thanks for sharing Roger! I haven't given this much thought yet, so good of you to raise awareness. I think your gut feeling might be more right than you imply, as the rpi isn't the only component to be taken into the equation The screen I think is the most power hungry, and the one I ordered draws perhaps 540mA at 5v (if it will run at that low voltage). These are some numbers i picked up at a forum post. Which means that the two most power hungry components in worst case senario draws 1000mA, which make it 2,5 hours with a 2550mAh battery. And then you'll need to account for the low power stuff: teensy, speakers (if you have) and so on.

Yeah. It is the awareness and the order-of-magnitude that count.

If the first calculation comes to: "half an hour" (as I secretly expected), then you get to chose: "half an hour is fine", "whoa, lets forget about this", or "I need a bigger battery". (on a different forum someone asked: can I use this solar cell to recharge my battery? Answer: Sure.... your fridge will discharge that battery in about 3 days. Then the solar cell will recharge the battery in 3 months....)

Also, these calculations show where you can achieve significant improvements. I was asked to make a driver for certain hardware with performance requirements. First implementation: 100 times too slow! Conclusion: Suppose I manage to find that 95% of the code executed is unnecessary, then I'll be still five times too slow. Most of that code was "standard Linux". I scrapped that "route". Going around that standard Linux code got me well within the range of acceptable performance.

In your case, "attacking" the teensy for current-consumption will reduce the total consumption at most by 5% (20mA on 400mA total). So if you want to reduce current consumption: look at the display...

oh my, I hit the backspace button when I shouldn't, now I've got to write this all over again

-The batteries can be hard to fit, yes. The use of a battery that you know fits in the case is an interesting idea. I would definately go for a larger battery though, as you can see in my reply to Roger the official larger battery Nintendo sells at 2550mAh 3.6v will mabye give you 2-3h max of game time, depending on your screen (if my crude calculations are correct; they probably aren't). There exists however larger capacity batteries that fits the gamepad housing from 3. parties, e.g. Nyko's powerpak at 4000mAh 3.7v. Adafruit with their super game pie reckons 6h of game time on their system, sporting a rpi a+ and 5inch hdmi lcd. And that is with a battery capacity of 6600mAh. So safe to say, I'll need atleast 6600mAh in my gamepad to drive the Pi2 pluss the 7inch lcd. I'm thinking of splitting the battery cells and cram them seperately at the top, or at the bottom plus where your hands grab. Another solution is to mount it at the back sticking out of the gamepad; it will be out of harms way but look a bit ugly mabye. I'm also planning to use the powerboost 500c btw (i figure you meant the 500c, since the 500 doesn't have a charging circuit).

-The teensy 3.1 will be able to read two analog imputs simultaniously, though how much documentation that excists for help in this matter i don't know (see: https://www.pjrc.com/teensy/teensy31.html, under "Two analog to digital converters"). I'm also targeting the Psx emu (which means we'll have to get all 24 buttons to work in order to play). I've opened a gamepad once, and it seems to be hard to resolder the buttons to the teensy, as the "pcb" really is some bendable plastic. And the 5 cables running from the thumb sticks aren't labeled, though if we are lucky they can be recognized as Vcc, Xout, Yout, Sel and GND. Mabye if this turns out to be too advanced for me, I'll try to jam a ps controller of sorts into the gamepad, that would make it plug n play on the software side atleast.

-Your plan with the usb sticks sounds smart, mabye I'll have to do it similarly since I already need an internal usb connection for the teensy (would look pretty silly to have the teensy sticking out of the usb port outside of the case..). As for ethernet I thought I'd cut the costs in not getting a wifi dongle, but mabye that'll be easier that extending the ethernet port to an I/O front (again, space is limited). It would be nice to have internet on this thing to ssh into it.

-Luckily for us, the gamepad is buildt with some (cupertino, cover your ears) empty space, which means this build should fit, but one tricky part is to account space for cables and such. I'm trying to draw the gamepad with parts in sketchup to get a feel of scale and space issues, but I'm still figuring out how to work the program. If I manage to produce anything of use I'll post it here

haha, I'm finding myself between the "i need a bigger battery" and "whoa, lets forget about this"   sure, the display is the sinner here, I've already ordered it so idk if it is possible to "hack" it in any way to draw less power, other than turning off the backlight..

Though as this lcd draws a constant power, running it off higer voltage means less amps pass though it. Even though some amps is lost when a regulator is used (I'm totally guessing here) to ramp up the current, would it be a more power friendly solution in theory to ramp up the current passing through the display? Or have I got this completely wrong?

haha, I'm finding myself between the "i need a bigger battery" and "whoa, lets forget about this"   sure, the display is the sinner here, I've already ordered it so idk if it is possible to "hack" it in any way to draw less power, other than turning off the backlight..

Though as this lcd draws a constant power, running it off higer voltage means less amps pass though it. Even though some amps is lost when a regulator is used (I'm totally guessing here) to ramp up the current, would it be a more power friendly solution in theory to ramp up the current passing through the display? Or have I got this completely wrong?

You know a classical 7805, right?

8V in, 5V out, it just "burns" the extra energy from that 3V. A good regulator like that will pass all current to the output, and use very little current on the ground pin. All that energy is wasted.... So a buck converter uses the 3V to "charge" an inductor, and then switch off the input, and run the load off the energy in the inductor for a while.

When you look at it this way the "store and recover" is the hard part. There are conduction losses in the "switch" that connects the inductor to the input at first and then that side of the inductor needs to be connected to ground for the discharge phase. The easiest way is to do that with a diode. There you get a 0.6V voltage drop: also losses.

What all this comes down to is: you're not converting 8V to 5V with your buck converter, but only 3V. If you manage 70% efficiency on this conversion, that's pretty good: you lose 30% of the energy that is actually stored and recovered in the inductor. So you're losing only 11% of the total energy delivered to the load.

However, if your input voltage goes up to say 12V, you're now losing 30% of the 7V that actually needs conversion. Overall efficiency has dropped to 82%.

Long story short: If you're converting DOWN, the lower the input voltage is, the more efficient your conversion will be. (The display might use the nominally 12V as a power source for the backlight That is an UP conversion where more input voltage means more efficiency.....)

So I got my gamepad off Ebay, it came in today, and I have realized there is VERY little space to run things through. I may end up extending the back with another case. Have you gotten any work done on yours? I'm still waiting for some parts to come in.