Hacking Hardware Picofly - a HWFLY switch modchip

[Helldrigo]

Hello fellow modders!
I have an OLED switch modded with the 2040 zero chip.
It always has a problem of inconsistently not booting into hekate, instead going straight to OFW, which was solved by rebooting several times until it finally starts hekate.
But now I've tried too many times to boot without success.
Upon inspection, my chip blinks the Long-Short error, indicating that D0 is not connected. I checked the wire with a multimeter and it looks okay.
Can anybody help me find the problem? Thanks

 

[psycho-neon]

Hello fellow modders!
I have an OLED switch modded with the 2040 zero chip.
It always has a problem of inconsistently not booting into hekate, instead going straight to OFW, which was solved by rebooting several times until it finally starts hekate.
But now I've tried too many times to boot without success.
Upon inspection, my chip blinks the Long-Short error, indicating that D0 is not connected. I checked the wire with a multimeter and it looks okay.
Can anybody help me find the problem? Thanks

Its the dat0 adapter issue. It may give you a reading, but it could be shorted to dat1. I would suggest to re-set the adapter and give it a better anchorage.

Otherwise, if you have the tools, go for reballing or kamikaze.

 

[Viktorsilva]

Hi Helldrigo



That´s a common error on the Oled. Dat0 adapters are crappy. You really need to push it underside EMMC after you get shure it stands in the right place. At the same time you need to anchor it with solder.Only then you can stop pushing it underside. In you case, i bet that multimeter gives you the right value because when you touch the dat0 adapter, you press it just a litle bit but enought to make contact. Im shure your RP2040 will start booting again.

You can try Kamikaze or Reflowing the nand and solder dat0 directly in the bga point. Those are solid connections!


Boa sorte para essa reparação

 

[dehydrated_lemur]

Has anyone heard of a bug in picofly where fuses don't burn when updating via daybreak?

 

[impeeza]

Has anyone heard of a bug in picofly where fuses don't burn when updating via daybreak?

Is not a bug THAT'S the desired behavior, daybreak prevent the fuse burnt

 

[QuiTim]

Has anyone heard of a bug in picofly where fuses don't burn when updating via daybreak?

Do you mean this: https://gbatemp.net/threads/picofly-a-hwfly-switch-modchip.622701/post-10137417
If so, the 'bug" you are refering to was fixed in v2.66

 

[dehydrated_lemur]

Do you mean this: https://gbatemp.net/threads/picofly-a-hwfly-switch-modchip.622701/post-10137417
If so, the 'bug" you are refering to was fixed in v2.66

yup that's what I was questioning. thanks for the info.

 

[rull_bull]

Dat1 via is gone ? At leat it looks like yes. All ready reply to you PM. You can bring your switch to me and maybe i can save it. I advise you not to do anything else. I live next to Parque das Nações .


i did successfully rebuild 2 vias on a nintendo switch oled in a very similar disaster. You can check it on this thread :

https://gbatemp.net/threads/picofly-a-hwfly-switch-modchip.622701/post-10349793

Post automatically merged: Feb 5, 2024

Dont forget to isolate the cpu solder points with kapton tape. Clean all the area with IPA solution first.


Kamikaze mod is risky but provides a solid connection to Dat0. Yes, you can reball the EMMC and make it solid too, but you will create thermal stress on solder joint and if i can remember problems like RSOD on PS3, PS4, Xbox 360 , etc....they all where created by the thermal stress (composition of lead-rich tin with low temperature melt point) even with professional welding done on a factory production line.That´s why i prefer the kamikaze too. Dat0 adapter could bend after some weeks (caused by heat).

Nah, RROD and YLOD was caused by lead-free solder that was cracking due to thermal stress.

if they had used solder with lead the problem would not have happened.

But due to the regulations, you cant have lead in your products and thats why they used a lead-free solder alloy that was not tested enough regarding thermal stress.

Lead solder is always better, except for health issues.

 

[terminalshort]

Hi all, I have been suggested to post my issue in this thread so here we go:

I have a mariko Switch with a HWfly chip (Picofly FW). It worked just fine until one day after many hours of use the application freezed, in Atmosphere. After reseting it, it booted normally, but it freezed again after a shorter period. After that it inmediately freezed while trying to turn it on again, sometimes not even booting and giving the error 2162-0002, amongst others.
I let it stay for the night and the next day it worked fine for many hours until the same happened again. After that the frequency of the crashes increased again.

First I though it could be corruption in the emuMMC or an SD fault, but the freeze also happened booting Stock firmware. I've tried different SD cards, factory reseted SysNAND and reinstalled everything but no luck. I've made it freeze in Atmosphere, Stock and Hekate. Even removing the modchip and booting OFW with no SD card.

The time it takes to freeze the first time from a cold boot is random, but it is usually long. Sometimes I get to play for an hour. After that it is much faster, a few minutes usually. This is consistent.
This has led me to think it is a temperature related issue instead of a problem with the eMMC. I have alsoreplaced the MT92 IC in case it was faulty, but the problem persists.

When the console freezes while in Hekate, CPU temperature readings are slightly above 30Cº (no heatsink). Current is in between 200-300 mA depending on backlight strength, so it seems normal. I checked for shorts in all supply rails but everything seems fine, just measuring some low impedance in the CPU supplies but it should be normal.

Finally, I looked at the board with a thermal camera after booting Hekate and I see a hotspot right above the MT92 IC. It inmediately reaches 42Cº with no activity. While in a game it gets hotter, I've measured up to 60Cº after a few minutes into a game.
This IC that's getting hot has the markings PP5 951. As I understand this is a MOSFET. Is it normal it gets this hot?

I don’t have pictures at hand ATM, but I can take some if needed.

P.d.: Also, in very rare ocassions the application crashed but I could get back to the home menu. Then it would crash shortly afterwards but this is definitely a different behavior. eMMC corruption? If the eMMC is broken why does temperature have an effect on the crash frequency?

 

[Viktorsilva]

Hi all, I have been suggested to post my issue in this thread so here we go:

I have a mariko Switch with a Picofly chip. It worked just fine until one day after many hours of use the application freezed, in Atmosphere. After reseting it, it booted normally, but it freezed again after a shorter period. After that it inmediately freezed while trying to turn it on again, sometimes not even booting and giving the error 2162-0002, amongst others.
I let it stay for the night and the next day it worked fine for many hours until the same happened again. After that the frequency of the crashes increased again.

First I though it could be corruption in the emuMMC or an SD fault, but the freeze also happened booting Stock firmware. I've tried different SD cards, factory reseted SysNAND and reinstalled everything but no luck. I've made it freeze in Atmosphere, Stock and Hekate. Even removing the modchip and booting OFW with no SD card.

The time it takes to freeze the first time from a cold boot is random, but it is usually long. Sometimes I get to play for an hour. After that it is much faster, a few minutes usually. This is consistent.
This has led me to think it is a temperature related issue instead of a problem with the eMMC. I have alsoreplaced the MT92 IC in case it was faulty, but the problem persists.

When the console freezes while in Hekate, CPU temperature readings are slightly above 30Cº (no heatsink). Current is in between 200-300 mA depending on backlight strength, so it seems normal. I checked for shorts in all supply rails but everything seems fine, just measuring some low impedance in the CPU supplies but it should be normal.

Finally, I looked at the board with a thermal camera after booting Hekate and I see a hotspot right above the MT92 IC. It inmediately reaches 42Cº with no activity. While in a game it gets hotter, I've measured up to 60Cº after a few minutes into a game.
This IC that's getting hot has the markings PP5 951. As I understand this is a MOSFET. Is it normal it gets this hot?

I don’t have pictures at hand ATM, but I can take some if needed.

P.d.: Also, in very rare ocassions the application crashed but I could get back to the home menu. Then it would crash shortly afterwards but this is definitely a different behavior. eMMC corruption? If the eMMC is broken why does temperature have an effect on the crash frequency?

Hi

Mosfets tend to get hot during operation. Did you check if fan is working ? Does it spin after a while you start your console ? Did you change Thermal Paste after installing the CPU Mosfet? What resistors are you using on Dat0, Clk and CMD ?

 

[terminalshort]

Hi

Mosfets tend to get hot during operation. Did you check if fan is working ? Does it spin after a while you start your console ? Did you change Thermal Paste after installing the CPU Mosfet? What resistors are you using on Dat0, Clk and CMD ?

Yeah the fan is working just fine, it starts spinning very fast when I remove the heat sink and start a game.
I changed the paste twice, last time after I replaced the MT92.
BTW the modchip is the HWfly so I just had to solder the capacitors to the ribbon cable and nothing else. The firmware inside is picofly, not the chip. I explained myself wrong.

 

[Endracion]

Here's two easier/safer mosfet positions on the back of the motherboard for the Switch Lite. They're both working fine. I moved the rp2040-zero LED to the home button led while i was at it just to test. Looks pretty neat, but not very useful.
I will personally not be using these points since the added time it takes to remove the motherboard makes it not really worth it for me, but I guess they might be helpful for people with less experience who are afraid of the apu caps.

I believe the big pad super easy position on the back for the OLED that I posted about before should be the standard recommendation though; it is far superior to all other locations (the pads are even pre-tinned!) and the motherboard has to be removed anyway.

Been browsing the last 200 pages to try and find these references lol! Thanks for this image - I'm also trying to find the same kind of point on the V1/V2, to save the images for reference. Do you happen to have one for those as well?

 

[Viktorsilva]

Yeah the fan is working just fine, it starts spinning very fast when I remove the heat sink and start a game.
I changed the paste twice, last time after I replaced the MT92.
BTW the modchip is the HWfly so I just had to solder the capacitors to the ribbon cable and nothing else. The firmware inside is picofly, not the chip. I explained myself wrong.

If i were you i try to remove the Modchip, clean everything with IPA and an old tooth brush and test the Switch .

 

[terminalshort]

If i were you i try to remove the Modchip, clean everything with IPA and an old tooth brush and test the Switch .

I will give it a try. I tried removing the chip but not the flex cable.

Update: Issue still there. I removed the flex cable and with it came a capacitor. Since these are decoupling capacitors and I have read of people having no problems without them I gave it a shot. If anything, it appeared to me that crashes were even moto frequent, but it might be my imagination.

After that I soldered back the cable and put two new 1uF capacitors. A messy job since I had to use 0603 capacitors, but managed to put everything back in place. Switch behaves the same way.

 

[Endracion]

Been browsing the last 200 pages to try and find these references lol! Thanks for this image - I'm also trying to find the same kind of point on the V1/V2, to save the images for reference. Do you happen to have one for those as well?

I poked around with the multimeter, here are the points on V1/V2:

Confirmed working as well, I just finished installing the picofly in a V1.

Or my shitty drawing of it next to the original back mosfet location from Abal.

 

[Viktorsilva]

I will give it a try. I tried removing the chip but not the flex cable.

Update: Issue still there. I removed the flex cable and with it came a capacitor. Since these are decoupling capacitors and I have read of people having no problems without them I gave it a shot. If anything, it appeared to me that crashes were even moto frequent, but it might be my imagination.

After that I soldered back the cable and put two new 1uF capacitors. A messy job since I had to use 0603 capacitors, but managed to put everything back in place. Switch behaves the same way.

you can get those capacitors under the picofly Zero. That is a very strage sinthom. Are you sure it wasn´t all ready like that before you installed the modchip ?' did you try to clean the cache memory or do a factory reset ? My faith in these methods are not great, but who knows....

Turn off the console. With the console turned off, hold down both volume buttons ((+) and (-)), and press the POWER Button to turn the console on. Be sure to keep the volume buttons held down until the Recovery Mode menu is displayed. Select “Restore Factoy Settings

 

[Gray-Venom]

Hi all! Please help me with the problem. I installed the chip, then launched the console, the chip blinked blue (a couple of times white), then 3 times green, then 3 more times green and 3 more times and... nothing. As I understand it, this is a "===" error. What could I have done wrong?

Spoiler: Chip image

Diode mode measurements:
3.3-0.923
С - 0.676
А - 0.742
D- 0.634
B- 1.590
Sorry for my english! Thank you!

 

[snaker]

I poked around with the multimeter, here are the points on V1/V2:View attachment 418831

Confirmed working as well, I just finished installing the picofly in a V1.

Or my shitty drawing of it next to the original back mosfet location from Abal.
View attachment 418832

you can also use 30 awg single core wire to fix it better.

 

[Vigintiduo]

Guys I'm having a nightmare with an attempted oled install. I have now removed every wire and mosfet, but as soon as I attach the battery, the charging icon shows up and then I get this artifacts with buzzing sounds.
Some times I can make it to the menu, but it still stops working.
Any idea on how to solve this mess?

 

[psycho-neon]

Tried to modify RP2040 Tiny to easier cable routing. Let's see how it goes once they arrived from jlcpcb.