Hacking Hardware Picofly - a HWFLY switch modchip

[Takezo-San]

Sooooooo did a Lite the other day and I can confirm that it worked great. My first Lite got bricked because i used 2.72fw and this time used a working firmware (2.67 in this case). Used all the points from the guide so the ground point and all other points aren't an issue. Some redemption for me here.

And used the IRFHS8342PbF with a pull down 1k on it using the back of the motherboard for my points as @abal1000x discovered. It initially worked 3-4 times but it was strange. It held the blue led for a second or two and then blinked blue and then green. Great. But then after 3-4 times, it did the infinite blue glitch. I'm not sure why. Maybe the pulldown resistor was too high, who knows. @floxcap might be onto something with using a lower value pull down for this mod. So I used another IRFHS8342PbF and removed the pull down on the other one too. Paired them, connected the relevant points and success. Normal super fast glitching. Thought I 'd share the finding but 1k might be too much for a pull down.

 

[Blakejansen]

How fucked am I?

I accidentally scratched some of the solder mask near the chip when removing the rail.

I am worried that I accidentally severed the tracks.

Anyone know if my switch will still work?

 

[Dee87]

How fucked am I?

I accidentally scratched some of the solder mask near the chip when removing the rail.

I am worried that I accidentally severed the tracks.

Anyone know if my switch will still work?

The only way to know if u damaged the trackes is to use a multimeter or take the risk and try to turn it on

 

[Takezo-San]

How fucked am I?

I accidentally scratched some of the solder mask near the chip when removing the rail.

I am worried that I accidentally severed the tracks.

Anyone know if my switch will still work?

I've done that. Still works. Just solder mask cover it up and look the other way.

 

[Blakejansen]

I've done that. Still works. Just solder mask cover it up and look the other way.

Points D,A,C all measure 350 instead of 500-800. Not sure what is wrong there.

3.3 measures at 750

In diode mode, Point B is showing OL not sure if I am supposed to switch it to to a different mode. I have spent 3 hours soldering point B on and off.

 

[abal1000x]

Sooooooo did a Lite the other day and I can confirm that it worked great. My first Lite got bricked because i used 2.72fw and this time used a working firmware (2.67 in this case). Used all the points from the guide so the ground point and all other points aren't an issue. Some redemption for me here.

And used the IRFHS8342PbF with a pull down 1k on it using the back of the motherboard for my points as @abal1000x discovered. It initially worked 3-4 times but it was strange. It held the blue led for a second or two and then blinked blue and then green. Great. But then after 3-4 times, it did the infinite blue glitch. I'm not sure why. Maybe the pulldown resistor was too high, who knows. @floxcap might be onto something with using a lower value pull down for this mod. So I used another IRFHS8342PbF and removed the pull down on the other one too. Paired them, connected the relevant points and success. Normal super fast glitching. Thought I 'd share the finding but 1k might be too much for a pull down.

1k is too low, you might change it to 10k, or 50k.

 

[RatchetRussian]

Been having a blast modding OLEDs lately. Still haven't had a chance to try the back mosfet though.

 

[Blakejansen]

DAT0 adapter reading was at 650, but when I check point C on the instinct, I am getting a reading of 350. A,C,D all give me readings of 350. No way that all I soldered all 3 points wrong and they are all at 350. I don't want to try turning my switch on until someone with more experience chimes in.

 

[lightninjay]

DAT0 adapter reading was at 650, but when I check point C on the instinct, I am getting a reading of 350. A,C,D all give me readings of 350. No way that all I soldered all 3 points wrong and they are all at 350. I don't want to try turning my switch on until someone with more experience chimes in.

My experience says that you're talking about an instinct which has no business being in this thread about Picofly.

 

[Blakejansen]

My experience says that you're talking about an instinct which has no business being in this thread about Picofly.

I thought the information would be universal since maybe someone has encountered this problem before on this thread, thus being relevant to the conversation.

Maybe someone has had this problem before with the pico install and could give me some suggestions.

 

[abal1000x]

DAT0 adapter reading was at 650, but when I check point C on the instinct, I am getting a reading of 350. A,C,D all give me readings of 350. No way that all I soldered all 3 points wrong and they are all at 350. I don't want to try turning my switch on until someone with more experience chimes in.

You're talking about reading the value. I am asking what value you are measuring actually.

If you mean diode mode, than its the drop voltage when the multimeter give a source current. I am assuming you use gnd as the black probe and the point such as dat0 as the red probe. Of course when you connecting to the pico, it will change, since the current flows not only to the emmc but also to the pico part. Hence difference read of the drop voltage. That is what i am understanding.

 

[Blakejansen]

You're talking about reading the value. I am asking what value you are measuring actually.

If you mean diode mode, than its the drop voltage when the multimeter give a source current. I am assuming you use gnd as the black probe and the point such as dat0 as the red probe. Of course when you connecting to the pico, it will change, since the current flows not only to the emmc but also to the pico part. Hence difference read of the drop voltage. That is what i am understanding.

I have my multimeter set to diode mode.

This is the multimeter that I am using. Points A,C,D are all between 350-380. Not sure why I am not getting to 500-800.

Wires look good and the DAT adapter point was reading at 650. ( Wire connected to adapter =650 Wire connected to chip = 350)

3.3 is at 750

I can't find the cause of lower than usual diode values.

Black probe touches the metal usb shield, red probe touches points(A,C,D)

What is causing my A,C,D points to read at 350 instead of the normal 500?

 

[abal1000x]

Wires look good and the DAT adapter point was reading at 650. ( Wire connected to adapter =650 Wire connected to chip = 350)

I am sure you're mistaken. Wire simply conductor. You will get the same value on each point in the line, except your wire problematic.

 

[Takezo-San]

I thought the information would be universal since maybe someone has encountered this problem before on this thread, thus being relevant to the conversation.

Maybe someone has had this problem before with the pico install and could give me some suggestions.

Look here. I thought you had a pico mod, why on earth are you using a filthy dirty nx chip? Anyway your fix would be as follows. First check if those torn traces give continuinity in diode mode. If they short then you must repair those traces.

In future, create a new thread asking for help about your specific chip you're using and the problem you're having. This thread does not support or help hwfly/nx installs or anyone looking for debugs/troubleshooting with those chip installs.

 

[twins333]

Sooooooo did a Lite the other day and I can confirm that it worked great. My first Lite got bricked because i used 2.72fw and this time used a working firmware (2.67 in this case). Used all the points from the guide so the ground point and all other points aren't an issue. Some redemption for me here.

And used the IRFHS8342PbF with a pull down 1k on it using the back of the motherboard for my points as @abal1000x discovered. It initially worked 3-4 times but it was strange. It held the blue led for a second or two and then blinked blue and then green. Great. But then after 3-4 times, it did the infinite blue glitch. I'm not sure why. Maybe the pulldown resistor was too high, who knows. @floxcap might be onto something with using a lower value pull down for this mod. So I used another IRFHS8342PbF and removed the pull down on the other one too. Paired them, connected the relevant points and success. Normal super fast glitching. Thought I 'd share the finding but 1k might be too much for a pull down.

Why not use 2.73? That's the officially supported and recommended version. Plus the only change between 2.72 and 2.73 is: "add LED signal on success" , as per: https://github.com/Ansem-SoD/Picofly/blob/main/Firmwares/changelog.md
Also there aren't any breaking changes between 2.67 and 2.72 that could cause your switch to brick.
I feel that people using different versions than the recommended one are just trolling at this point, since 2.73 has been recommended multiple times on this thread.

 

[Takezo-San]

Why not use 2.73? That's the officially supported and recommended version. Plus the only change between 2.72 and 2.73 is: "add LED signal on success" , as per: https://github.com/Ansem-SoD/Picofly/blob/main/Firmwares/changelog.md
Also there aren't any breaking changes between 2.67 and 2.72 that could cause your switch to brick.
I feel that people using different versions than the recommended one are just trolling at this point, since 2.73 has been recommended multiple times on this thread.

I downloaded and used 2.72fw upon its hourly release. At that point there was a programming problem with clk of the fw. 6hours or so later, the fw bug was fixed but it was too late for me. The emmc got corrupted during the initial glitch and incurred the bsod. Emmc was corrupt. Hope that brings clarity to the problem I had with 2.72fw. Nothing wrong with it now.
I prefer the colour array of 2.67fw. Not a fan of the yellow and white led system. Norhing personal, just my preference. I will use 2.73fw down the line eventually.
BTW, what's the difference from 2.67 and 2.73 other than the colour codes and colour array.

v2.67 + Don't bypass to OFW on first install
v2.70 + new LED indication, i2c undervoltage hack
v2.71 + support for SQc open-source board
v2.72 + disable CLK check, it's unstable
V2.73 + add LED signal on success

Hmmm upon reviewing the above info, I will try 2.73fw next. @twins333 thanks for the reminder.

 

[twins333]

Have you tried 2.72?

Yes and it worked just fine. That was the latest version available when I got my v2 modded. I upgraded to 2.73 . Same with my brother's v2. Worked just fine, but I upgraded to 2.73 with the picofly_toolbox.

Post automatically merged: Jul 21, 2023

I downloaded and used 2.72fw upon its hourly release. At that point there was a programming problem with clk of the fw. 6hours or so later, the fw bug was fixed but it was too late for me. The emmc got corrupted during the initial glitch and incurred the bsod. Emmc was corrupt. Hope that brings clarity to the problem I had with 2.72fw. Nothing wrong with it now.
I prefer the colour array of 2.67fw. Not a fan of the yellow and white led system. Norhing personal, just my preference. I will use 2.73fw down the line eventually.
BTW, what's the difference from 2.67 and 2.73 other than the colour codes and colour array.

v2.67 + Don't bypass to OFW on first install
v2.70 + new LED indication, i2c undervoltage hack
v2.71 + support for SQc open-source board
v2.72 + disable CLK check, it's unstable
V2.73 + add LED signal on success

Hmmm upon reviewing the above info, I might just try 2.73fw. @twins333 thanks for the reminder.

Sorry, I forgot that 2.72 had two releases and the first one had major problems. You are correct about it ruining switches. My bad here.
My point about 2.73 stands. It's still a good version to upgrade to(and the latest stable one at that, at least for me).

 

[Dee87]

I downloaded and used 2.72fw upon its hourly release. At that point there was a programming problem with clk of the fw. 6hours or so later, the fw bug was fixed but it was too late for me. The emmc got corrupted during the initial glitch and incurred the bsod. Emmc was corrupt. Hope that brings clarity to the problem I had with 2.72fw. Nothing wrong with it now.
I prefer the colour array of 2.67fw. Not a fan of the yellow and white led system. Norhing personal, just my preference. I will use 2.73fw down the line eventually.
BTW, what's the difference from 2.67 and 2.73 other than the colour codes and colour array.

v2.67 + Don't bypass to OFW on first install
v2.70 + new LED indication, i2c undervoltage hack
v2.71 + support for SQc open-source board
v2.72 + disable CLK check, it's unstable
V2.73 + add LED signal on success

Hmmm upon reviewing the above info, I will try 2.73fw next. @twins333 thanks for the reminder.

always use the latest release bro :-)

yes the first 2.72 had a bug, but u can fix a corrupted emmc :-)

Post automatically merged: Jul 21, 2023

Been having a blast modding OLEDs lately. Still haven't had a chance to try the back mosfet though.

i love do see another persone modding an Oled the only Correcr way with a permanent dat0 adapter !!!!!!!

Keep on going

 

[Takezo-San]

but u can fix a corrupted emmc

How? I thought if the emmc is corrupt, the switch is done!?

 

[twins333]

How? I thought if the emmc is corrupt, the switch is done!?

Check this: https://gbatemp.net/threads/switch-unbricking-guide-for-dead-or-replaced-emmc-consoles.609891/
It might have some info for the restore/fix of the emmc.
Here as well: https://gbatemp.net/threads/switch-unbricking-guide-prodinfo-fix-pikafix.578026/