Hacking Hardware Picofly - a HWFLY switch modchip

[MegaDeKay]

A rp2040 chip in hwfly format with flex cables. Latest firmware 2.74

No. 2.74 was pulled. See below.

(v2.74 + 300 mhz precision rp2040 may be not stable at 300mhz.from @rehius post)

Edit: I think the rp2040 is not stable that's why switch boots to OFW if the glitch didn't work.

 

[jmmc]

Does anyone have the 2.74, it’s so much faster than 2.73 at least in some of my previous installations. I want to install them on newers

 

[MegaDeKay]

The TDP of the cpu is around 10Watt. So for 1.8V it will be use something like 5Amps maxed.

So the design of the pcb must make sure the power line could handle 10 watt of electricity goes to the cpu. The area and the thickness play a role part in the design.

The trace of power line usually different. Its bigger than the data line. If you know the trace, please direct me the position.

To add to this, power is often distributed on planes rather than individual traces and the TDP is spread across different voltages. This is the picture I found doing a quick search on the Tegra X1 and it is probably the reference platform layout. One more thing... the average TDP of the 210 used in the Switch is 10W but it can peak at 15W. Link.

 

[bilalhassan341]

Someone can explain me this error :

==* CPU always reach BCT check (no glitch reaction, check mosfet)

I faced this error very often. This issue is also related to your install also or weak mosfet or mosfet not working. it is mostly fixed by the SDA And SCL wires because I'm too lazy to check. But it is also related to mosfet. Just first place the SDA and SCL to see if it post or not, or you can press the reset button on RP2040 when joycon logo shows to see. If SDA and SCL wires and reset button method don't work then check mosfet.

 

[JaRocker]

perhaps; maybe by mounting the status LED further away, perpendicular to the motherboard, such that it is casting its light into the side of the translucent cover, instead of resting under it and casting upward. hopefully that would still diffuse it enough to be evenly lit. I had done something similar with the status LED of my SAMD21 installs on the original switch using the light sensor, where it cast into the side of the sensor's light pipe. if I do another install I will play around with other solutions

thanks, looking forward to your next post

 

[w0nd3rb0yfx]

Does anyone have the 2.74, it’s so much faster than 2.73 at least in some of my previous installations. I want to install them on newers

For me it only worked fine with fw 2.74.

 

[Dee87]

Someone can explain me this error :

==* CPU always reach BCT check (no glitch reaction, check mosfet)

1 or 2 MOSFETs ? If one install a second
What wires from MOSFET to apu if 0.1 use 0.2
Otherwise check ur soldering

 

[tSmoker]

Here's a story for you.
I've been fighting with OLED all evening. There was an error: *== No eMMC CMD1 request (poor wiring, or dead CPU)
At the same time, the console was loaded into OFW
The soldering is good.
Mosfet suspected (Becouse of OFW running). Rewired. (I always use 1 and always have enough)
And when I was putting it back together after soldering mosfet (only one), I found an error... Mixed up CMD and CLK lines!
Just a little bit of knowledge for you in the knowledge base.

 

[detilmalala]

HELP, I was formatting a nintendo switch and it turned off during the process and now it won't turn on, does anyone know if the console can be recovered?

 

[RatchetRussian]

i've never heard someone have problem with cpu flex cable.

The common problem is usually short circuit.
Theres lot of thing you might check, for example on picofly board.
1. Check resistance between 'cpu pin' on pico with GND.
If its around 20-60 ohms its okay, if its around 0-5ohms, its shorted.
2. Check continuity between RST and other pin. Do the same thing on CLK, CMD, Dat0.
3. Use microscope and examine the board, especially the place that you modded. Search for something that looks shorted, or something that might ripped.

Resistance to cpu pin is 800 ohms.

I think it is just a bad picofly chip. It's the only conclusion I was able to come to after testing.

When I desolder the rst wire from the picofly, the rst voltage goes to the correct 1.9v so I think something on the picofly is causing it to be 0.9v.

I tried testing by desoldering 1 wire at a time. The switch will boot to OFW successfully unless either dat0 or cmd are connected to the picofly.

All of it makes no sense to me. I will just try a new picofly and see if it works. I am just glad the switch is okay.

edit: IT WORKS! It was a bad picofly indeed. Maybe I killed it when desoldering the type c connector. I did not see any shorts on the picofly itself though.

 

[jkyoho]

HELP, I was formatting a nintendo switch and it turned off during the process and now it won't turn on, does anyone know if the console can be recovered?

recovery mode
or hekate mount raw partition and NxNandManager wipe user partiton

 

[detilmalala]

recovery mode
or hekate mount raw partition and NxNandManager wipe user partiton

Do you know any tutorial?

 

[cgtchy0412]

Does anyone have the 2.74, it’s so much faster than 2.73 at least in some of my previous installations. I want to install them on newers

You can find it here
https://github.com/rehius/usk

 

[axlcat]

Greetings, after many installations I have a problem with bluetooth and wifi, the joycons and the internet don't connect, the same thing happens with or without picofly, I changed the wifi module and it's still the same, any idea what's wrong? It's a nintendo Oled

 

[BlueBeans]

Need some help guys. Just got this message from someone who picked up from me yesterday.

 

[lightninjay]

Greetings, after many installations I have a problem with bluetooth and wifi, the joycons and the internet don't connect, the same thing happens with or without picofly, I changed the wifi module and it's still the same, any idea what's wrong? It's a nintendo Oled

I'd be willing to bet you accidentally broke two components loose near the APU shield. They are small 0201 SMD components and are connected to the WiFi/Bluetooth chip. With these traces broken, you have no wifi or bluetooth, which also keeps joycons from working wirelessly.

It should be these two, and you likely broke them off when removing the shield to solder the flex cable/Mosfets. They are 100nf capacitors and can be replaced using the same ones that can be robbed off the RP2040, as noted in the definitive guide.

 

[axlcat]

I'd be willing to bet you accidentally broke two components loose near the APU shield. They are small 0201 SMD components and are connected to the WiFi/Bluetooth chip. With these traces broken, you have no wifi or bluetooth, which also keeps joycons from working wirelessly.

It should be these two, and you likely broke them off when removing the shield to solder the flex cable/Mosfets. They are 100nf capacitors and can be replaced using the same ones that can be robbed off the RP2040, as noted in the definitive guide.
View attachment 380110

Thanks for repply, I'll try to fix that tomorrow (it's almost midnight here)

 

[lightninjay]

Thanks for repply, I'll try to fix that tomorrow (it's midnight here)

Same here, woop woop East Coast!

 

[Kiuxn]

Use a single mos for at least 40 switches（V1or V2), Instead of using SDA/SCL,it works well.

 

[abal1000x]

Use a single mos for at least 40 switches（V1or V2), Instead of using SDA/SCL,it works well.

Thats beautifully perfect.