Hacking Hardware Picofly - a HWFLY switch modchip

[cgtchy0412]

Yes it is.

All mosfet have capacitance, so need a resistor to discharge it. This to make sure the gate always low first time its powered on, so the DS are not bridged.

For this mod, the scenario happened when the battery are completely discharge, then recharge it using usbc, then the initial state of the mosfet are unknown, the DS could be bridged or not. If its bridged then the cpu 'not working', since the power are 'stolen' by those mosfet. Then the mosfet might burn, because of continuous high current flows on it.

How much probable this scenario will result in CPU being fried? even its kind of hard to completely discharge the battery if we are talking about normal usage.
Also this is not specific to any physical resistor being present, it should be any "resistance to ground" and the pico board itself already has that.
As myself experience this only once, found that the switch is completely dead no power whatsoever, after diassembly found that i need to touch the mosfet with bare hand to discharge it and it coming back to life. Lucky that this happened at the time of instalation.
Also that case happened when still using OnSemi, never again with AON, dont know if mosfet brand matters.

Post automatically merged: Aug 24, 2023

I tried the new position of the mosfet in an OLED, works well, a bit fun but not super useful for Oleds models.

And I dunno why but on the next 2 switch I had to add 47ohms for dat0 otherwise slow mode or no boot on ofw and syscfw, black screen. Everything is ok after adding +47ohms (total 94ohms)

Thats kind of mosfet has been done on many pages back.
For resistor now the forum recomends 100, 47, 100 (d0, clk, cmd).

 

[abal1000x]

How much probable this scenario will result in CPU being fried? even its kind of hard to completely discharge the battery if we are talking about normal usage.
Also this is not specific to any physical resistor being present, it should be any "resistance to ground" and the pico board itself already has that.
As myself experience this only once, found that the switch is completely dead no power whatsoever, after diassembly found that i need to touch the mosfet with bare hand to discharge it and it coming back to life. Lucky that this happened at the time of instalation.
Also that case happened when still using OnSemi, never again with AON, dont know if mosfet brand matters.

The cpu wont fried, no worry. If the mosfet turn on, the cpu are turn off. Simply like that. The bridged continuous current will flow to mosfet, than the probable to broke is the mosfet, or the power ic that supply the power.

The chance should be small enough to ignored. But theres still possibility, and if you track from the post, someone experience the mosfet smoking.

This should only happened in the installation process, since the picofly firmware will take care the pulldown resistor when its powered on. The only exception is if the Gate line somehow disconnected from the picofly, than the user will get that problem.

 

[LogicalMadness]

For all the 'PFly is bad, gave black screen, HWFly fixed it' posts, I just had the opposite happen. An Instinct NX died, which made the switch 'black screen' w\o it installed and installing a PFly fixed it.

Even more proof that PFlys are problem free, it's all on the installer to not fk up. In my case, not realizing clk was shorted to ground, and then trying to boot, killed the Instinct (pretty sure).

Also, more proof that a digital multimeter is not an "optional tool", and another reason to not rush the installs. I skipped my usual DMM double check on this one and that one small 'skip' cost me days of troubleshooting.

 

[snaker]

For all the 'PFly is bad, gave black screen, HWFly fixed it' posts, I just had the opposite happen. An Instinct NX died, which made the switch 'black screen' w\o it installed and installing a PFly fixed it.

Even more proof that PFlys are problem free, it's all on the installer to not fk up. In my case, not realizing clk was shorted to ground, and then trying to boot, killed the Instinct (pretty sure).

Also, more proof that a digital multimeter is not an "optional tool", and another reason to not rush the installs. I skipped my usual DMM double check on this one and that one small 'skip' cost me days of troubleshooting.

I have made a lot of installs of instinct-nx, picofly clones and original rp2040 chips and i can tell with certainty that picofly is faster than instinct. The reliability i can say its almost the same. All my installs work flawlessly and almost all the times they work first try but the thing with the flex cables is starting to become very annoying.
At 3 of my installs, i don't remember the chips exactly, i had infinite blue blinking light with no yellow error codes. All 3 of these times it was the cpu flex the reason of the infinite glitching progress. So i started investigating the reasons of this. And this is also why i asked to the previous post about the pull down resistor values. I prefer the flexes because the installation time is much faster. From your and Abals reply I assume that the pull down resistor values on these flexes are two low so i ordered some different higher values in order to replace those 0.5 to 1.1k pull down resistors. I still don't understand the reason why they use so low value pull down resistors on these flexes. Maybe the mosfets they use don't require higher values?? Who knows. So from my experience be very aware of your cpu flex cables and that they are not faulty!

 

[LogicalMadness]

I have made a lot of installs of instinct-nx, picofly clones and original rp2040 chips and i can tell with certainty that picofly is faster than instinct. The reliability i can say its almost the same. All my installs work flawlessly and almost all the times they work first try but the thing with the flex cables is starting to become very annoying.
At 3 of my installs, i don't remember the chips exactly, i had infinite blue blinking light with no yellow error codes. All 3 of these times it was the cpu flex the reason of the infinite glitching progress. So i started investigating the reasons of this. And this is also why i asked to the previous post about the pull down resistor values. I prefer the flexes because the installation time is much faster. From your and Abals reply I assume that the pull down resistor values on these flexes are two low so i ordered some different higher values in order to replace those 0.5 to 1.1k pull down resistors. I still don't understand the reason why they use so low value pull down resistors on these flexes. Maybe the mosfets they use don't require higher values?? Who knows. So from my experience be very aware of you flex cables and that they are not faulty!

After this Instinct experience (it came back to me because clk disconnected from the flex) I'm never using ANY flex\adaptor again. This Instinct's APU-flex was replaced by a single irh8342 on the back, using a 0201 33k resistor across S and G.

 

[snaker]

After this Instinct experience (it came back to me because clk disconnected from the flex) I'm never using ANY flex\adaptor again. This Instinct's APU-flex was replaced by a single irh8342 on the back, using a 0201 33k resistor across S and G.

Did you used solder to the clk pad under the flex and then solder it on top of clk via point with lot of no clean flux? I never had any returns and i always used the chip points included flex (not the cpu flex we mentioned above) on the instinct-nx chips before picofly released.

 

[LogicalMadness]

Did you used solder to the clk pad under the flex and then solder it on top of clk via point with lot of no clean flux? I never had any returns and i always used the chip points included flex (not the cpu flex we mentioned above) on the instinct-nx chips before picofly released.

I find flexes go on easier if you solder after it's laid down. So, no, I did not add solder to clk and then solder the flex to it (I don't think I did at least). The more I think about it though, I actually MAY have JUST for clk, due to it's 'scratch the surface' nature, I might have added solder to it w\ the thought of 'I want to be sure it can take solder' or 'do I need to scratch more?'. It was months ago. Then it came back a couple weeks ago, I quick resoldered clk and now also uv cured it down. It came back again the same day... Which is why I tried rushing a full reinstall for him in the 1st place.

Regardless, I've had too many issues w\ flexes\adaptors as a whole to use them at all anymore. I bought some wire I really enjoy using (PTFE Single Copper Wire 34 AWG Silver Plated 1 Core Micro Fine UL1423 High Temperature Electrical Cable
https://a.aliexpress.com/_mNmkDAA) and I don't care how long it takes anymore (quality over quantity, I allowed this forum to influence me and started wanting the 'top time score'). I only install HWFly-ish chips if the client REALLY wants it anyway (he did in the beginning, but is now completely onboard with PFly) and if it happens again I'll just skip the flexes and adapters.

 

[FreeLander]

How much probable this scenario will result in CPU being fried? even its kind of hard to completely discharge the battery if we are talking about normal usage.
Also this is not specific to any physical resistor being present, it should be any "resistance to ground" and the pico board itself already has that.
As myself experience this only once, found that the switch is completely dead no power whatsoever, after diassembly found that i need to touch the mosfet with bare hand to discharge it and it coming back to life. Lucky that this happened at the time of instalation.
Also that case happened when still using OnSemi, never again with AON, dont know if mosfet brand matters.

Post automatically merged: Aug 24, 2023

Thats kind of mosfet has been done on many pages back.
For resistor now the forum recomends 100, 47, 100 (d0, clk, cmd).

I'd like to report that I also got the slow emmc error when using 47s. This should be the standard.
I was excited to try out RP-tiny since it has these pre-soldered, but seems that we still need to solder 2.
What is a good approach for adding 2 resistors to cmd and data on tiny? Should I just add a single to each pad as if it was a regular?

 

[QuiTim]

I'd like to report that I also got the slow emmc error when using 47s. This should be the standard.
I was excited to try out RP-tiny since it has these pre-soldered, but seems that we still need to solder 2.
What is a good approach for adding 2 resistors to cmd and data on tiny? Should I just add a single to each pad as if it was a regular?

That should work, no problem. But if you want to make it look good i think that replacing the resistors on board would be much better

 

[Dee87]

I'd like to report that I also got the slow emmc error when using 47s. This should be the standard.
I was excited to try out RP-tiny since it has these pre-soldered, but seems that we still need to solder 2.
What is a good approach for adding 2 resistors to cmd and data on tiny? Should I just add a single to each pad as if it was a regular?

i think the best thing to do is take 47ohm of and set a 100ohm resistor instead of adding another 47 resistor to it .

if u really want to take the route with 2 47ohm i would suggest cut the line that goes from the 47 ohm to the solder pad and add another resistor behind the original one .

just got to make sure u cut enough of the via away that u can set the second resistor properly with out any bridging underneath

 

[AlicatFit]

Hi could someone give me some advice please?I opened up my Mariko OLED Picofly to replace the thermal paste.Got it all back together and now it only boots into OFW can't get into hekate.Thanks.

 

[Dee87]

Hi could someone give me some advice please?I opened up my Mariko OLED Picofly to replace the thermal paste.Got it all back together and now it only boots into OFW can't get into hekate.Thanks.

well theres not much we can tell u ,

u need to give us the morse code and some pictures of the install maybe u disconnected something on accident

 

[AlicatFit]

well theres not much we can tell u ,

u need to give us the morse code and some pictures of the install maybe u disconnected something on accident

I'm a noob on the subject, is the morse code the lights that flash at boot?If yes it flashes blue once and then green once.

 

[QuiTim]

I'd like to report that I also got the slow emmc error when using 47s. This should be the standard.
I was excited to try out RP-tiny since it has these pre-soldered, but seems that we still need to solder 2.
What is a good approach for adding 2 resistors to cmd and data on tiny? Should I just add a single to each pad as if it was a regular?

So I got off my a§ and gave it a try because sometimes I don't feel good giving out advice for things I've not tried myself
Since I did not have any 0201 resistors at hand I used 0402.
And since the solder pads on board are not made for this size I had to place the resistors slightly off place in order to be sure that they are making good contact.
Everything is measured and everything works as intended.
I did this using a 0.1 solder tip but I think that if one is to use 0201 resistors it would be much easier to use kapton, solder paste and hot air.

 

[Dee87]

I'm a noob on the subject, is the morse code the lights that flash at boot?If yes it flashes blue once and then green once.

well if it flashes blue once and then green once it should glitch and boot the no sdcard or to hekate , so in ur case give of some pictures of your install please

Post automatically merged: Aug 24, 2023

So I got off my a§ and gave it a try because sometimes I don't feel good giving out advice for things I've not tried myself
Since I did not have any 0201 resistors at hand I used 0402.
And since the solder pads on board are not made for this size I had to place the resistors slightly off place in order to be sure that they are making good contact.
Everything is measured and everything works as intended.
I did this using a 0.1 solder tip but I think that if one is to use 0201 resistors it would be much easier to use kapton, solder paste and hot air.View attachment 390148

i just add a bit more solder to the pads and use hot air and flux to set them and push a bit down on the resistors.

i also use 0402 resistor size

 

[QuiTim]

well if it flashes blue once and then green once it should glitch and boot the no sdcard or to hekate , so in ur case give of some pictures of your install please

Post automatically merged: Aug 24, 2023

i just add a bit more solder to the pads and use hot air and flux to set them and push a bit down on the resistors.

i also use 0402 resistor size

View attachment 390150

Yep, this definetly looks better. I'll give it a shot tomorrow. Thanks for the tip

 

[Dee87]

Yep, this definetly looks better. I'll give it a shot tomorrow. Thanks for the tip

look isnt everything mostly it just needs to work :-)

 

[Purple_Heart]

I have a question about hwfly.... Can I ask it here?

 

[QuiTim]

I have a question about hwfly.... Can I ask it here?

I think here would be better

 

[Eds05]

How much probable this scenario will result in CPU being fried? even its kind of hard to completely discharge the battery if we are talking about normal usage.
Also this is not specific to any physical resistor being present, it should be any "resistance to ground" and the pico board itself already has that.
As myself experience this only once, found that the switch is completely dead no power whatsoever, after diassembly found that i need to touch the mosfet with bare hand to discharge it and it coming back to life. Lucky that this happened at the time of instalation.
Also that case happened when still using OnSemi, never again with AON, dont know if mosfet brand matters.

Post automatically merged: Aug 24, 2023

Thats kind of mosfet has been done on many pages back.
For resistor now the forum recomends 100, 47, 100 (d0, clk, cmd).

For the moment I put 100ohms only on DAT0, why CMD need it too ?