Hacking Hardware Picofly - a HWFLY switch modchip

[abal1000x]

.

 

[Switxh]

Doesn't use Sda and Scl right?

Yeah I'm not using SDA SCL. 2.74 firmware. It's SK HYNIX emmc with 4 anchor dat0 adapter. (Reseating that was one of the first things I tried)

 

[QuiTim]

Yeah I'm not using SDA SCL. 2.74 firmware. It's SK HYNIX emmc with 4 anchor dat0 adapter. (Reseating that was one of the first things I tried)

Did you try 2.67fw, for some people this version seems to resolve some unusual issues

 

[LogicalMadness]

An update: Unfortunately installing mosfets in this way made no difference for my glitch times Still randomly glitches super slow and other times glitches in 1-2 seconds. So far I have tried two different HWFLY flex cables, single mosfet and dual mosfet and there is no real difference between any of them for me. I was resetting training data and letting the chip train upto 20 times for each configuration I tried. It took ages lol. Thanks for help anyway! I wonder if my CPU is just hard to glitch?

What size wires?

 

[Switxh]

Did you try 2.67fw, for some people this version seems to resolve some unusual issues

No I don't think I tried 2.67. I can't find any evidence of it in my downloads. I know I did try 2.64 cus i still have it. I have found 2.67 I will flash it and see what happens. Thanks!

What size wires?

for the mosfets? I used 30 AWG that I removed the insulation from. For the rest I used 30 AWG and 0.1 enamelled wire for CLK and CMD

 

[POPOLO]

An update: Unfortunately installing mosfets in this way made no difference for my glitch times Still randomly glitches super slow and other times glitches in 1-2 seconds. So far I have tried two different HWFLY flex cables, single mosfet and dual mosfet and there is no real difference between any of them for me. I was resetting training data and letting the chip train upto 20 times for each configuration I tried. It took ages lol. Thanks for help anyway! I wonder if my CPU is just hard to glitch?

Personally try OLED installation (the number of samples is only four)
NO1 100%
FW 2.74 (light indicator blue~~white) + HWFLY flex cables, dual mosfet + I2C SDA SCL

NO2 90% success 10% chance of original factory startup
FW 2.73 (light indicator blue~~yellow) + HWFLY flex cables, dual mosfet + I2C SDA SCL

Install HWFLY flex cables, dual mosfet is sure to install the power-on current will be very low
Insufficient welding Leakage Current will be very high Abnormal power consumption
====================================================================================
It is said that using a self-made cable single mosfet +FW 2.74 (light indicator blue~~white) without connecting SDA + SCL can be very stable to boot

 

[jkyoho]

Personally try OLED installation (the number of samples is only four)
NO1 100%
FW 2.74 (light indicator blue~~white) + HWFLY flex cables, dual mosfet + I2C SDA SCL

NO2 90% success 10% chance of original factory startup
FW 2.73 (light indicator blue~~yellow) + HWFLY flex cables, dual mosfet + I2C SDA SCL

Install HWFLY flex cables, dual mosfet is sure to install the power-on current will be very low
Insufficient welding Leakage Current will be very high Abnormal power consumption
====================================================================================
It is said that using a self-made cable single mosfet +FW 2.74 (light indicator blue~~white) without connecting SDA + SCL can be very stable to boot

This is 3.3v line current or the gate?

 

[abal1000x]

.

 

[Pickle_Rick]

Yes, in the AIO thread he stated a theory that not all pico boards can handle the 300 MHz speed. This might be why older versions like 2.67 and 2.73 work better for some people. So he now has 2.73 as the latest on his firmware post in this thread.

Hmm, should we downgrade if ours is already working on 2.74? Seems kinda dangerous to run it at 300mhz at all times if a lot of units can't even handle the overclock. Like it'll burn it out a lot faster.

 

[POPOLO]

Battery to Device Simple Ammeter
It is convenient to check whether the working current is normal
abnormal power consumption
Like handmade can create a

 

[jkyoho]

Battery to Device Simple Ammeter
It is convenient to check whether the working current is normal
abnormal power consumption
Like handmade can create a

You should use a 1A range ampmeter then. For the Vsys/battery current draw, I remember OFW only draws around 150-->280-->4**(v2)or 6** mA, from power button press to Home page.

 

[LogicalMadness]

No I don't think I tried 2.67. I can't find any evidence of it in my downloads. I know I did try 2.64 cus i still have it. I have found 2.67 I will flash it and see what happens. Thanks!

for the mosfets? I used 30 AWG that I removed the insulation from. For the rest I used 30 AWG and 0.1 enamelled wire for CLK and CMD

Did you install a pulldown resistor across gate and source?

 

[Switxh]

Did you install a pulldown resistor across gate and source?

I could only find 100ohm in the correct size all my 1k ohm and higher were a size too big so I didn't do the pulldown resistor no

On a side note, I've noticed sometimes when you go to turn the Switch on it is unresponsive until you hold the power button for 20s and then let go and press power button again. I have had this happen on multiple different Switches and I'm wondering what causes it.

 

[QuiTim]

Inspired by @Takezo-San recent achievement and being bored while waiting for my new Oled to arrive I thought I'd give it a try a put the Ball in Reball a.k.a trying to do a reball without stencil or paste just by arranging the solder balls manually.
And what do you know, turned out much better than I expected.
This was an emmc 153 pulled from an old phone so there is still some way to go until I decide to do this on Switch but I’m getting there
Edit:
The process: spread a thin layer of flux on emmc (with finger); arrange the balls; heat them on 300°C lowest air flow until they are attached to surface; add more flux on top of them and reheat with 300°C medium airflow until nice and shiny; clean the flux as if your life depended on it.

 

[meeebro]

Ok bad news, tried the yellow dat0 adapter instead of the OAT0 I was using before and I still get the failed to init emmc error. Reading is once again 0.7v. Anything else that I could be doing wrong?

 

[Takezo-San]

Inspired by @Takezo-San recent achievement and being bored while waiting for my new Oled to arrive I thought I'd give it a try a put the Ball in Reball a.k.a trying to do a reball without stencil or paste just by arranging the solder balls manually.
And what do you know, turned out much better than I expected.
This was an emmc 153 pulled from an old phone so there is still some way to go until I decide to do this on Switch but I’m getting there
Edit:
The process: spread a thin layer of flux on emmc (with finger); arrange the balls; heat them on 300°C lowest air flow until they are attached to surface; add more flux on top of them and reheat with 300°C medium airflow until nice and shiny; clean the flux as your life depended on it.

Amazing stuff. Did the old phone work with reball reintall and how did you manage to place those fiddly balls on top of the pads. I tried to use solder balls but there was no divet in the chip pads for them to settle in. Hence I opted to stick with the paste. But yes! We inspire and motivate each other. I just hope we can inspire others to do the same and discard the thought of dat0 adaptors that are hit or miss. Great work!

Edit - Just saw the video. So you slide the balls into place. Got it. But how did you stop them from flying away with your hot air station? Low air flow/high heat?

 

[Dee87]

Amazing stuff. Did the old phone work with reball reintall and how did you manage to place those fiddly balls on top of the pads. I tried to use solder balls but there was no divet in the chip pads for them to settle in. Hence I opted to stick with the paste. But yes! We inspire and motivate each other. I just hope we can inspire others to do the same and discard the thought of dat0 adaptors that are hit or miss. Great work!

Edit - Just saw the video. So you slide the balls into place. Got it. But how did you stop them from flying away with your hot air station? Low air flow/high heat?

Just low are flow and heat till they beginn to melt then u can use more airflow.

Balls are cool but also a pain in the Ass that's why I prefer the past works faster and easier

 

[QuiTim]

Amazing stuff. Did the old phone work with reball reintall and how did you manage to place those fiddly balls on top of the pads. I tried to use solder balls but there was no divet in the chip pads for them to settle in. Hence I opted to stick with the paste. But yes! We inspire and motivate each other. I just hope we can inspire others to do the same and discard the thought of dat0 adaptors that are hit or miss. Great work!

Edit - Just saw the video. So you slide the balls into place. Got it. But how did you stop them from flying away with your hot air station? Low air flow/high heat?

The phone was dead long ago so no way to know.
As for the process, I edited my original post with more info but just briefly, once you put some flux on the emmc the balls are easy to arange, once that's done 300° lowest airflow until the balls stick to emmc, afterwards add more flux all over the emmc and heat again 300° medium airflow until everything is set nicely. 30min total.
The balls have a melting point of around 190° so that makes everything easier.
My only concern with the whole reball process is that I have no reall feel of the amount of heat emmc can take without sustaining damage since I don't have a "real" rework station, the actual temperature from my heatgun can vary :/
This is what i'm currently using:
https://a.aliexpress.com/_mqrLI16

 

[HelloShitty]

Would this modchip belong to this thread?

If so, I'll ask mods to move one post of mine to here...

 

[Moromoromoro]

I guess this has been answered before but have a 2 switches boards one had-cpu-01 and had-cpu-10:
-worked fine for a few days able to boot into cfw and ofw no problems, rp2040 zero board(2.67fw) from waveshare 94ohm resistor on data 47ohm cmd resistor on CLK. Used hwfly CPU flex
-started with 0dat issues, eventually resulted in BSOD - no sd card ok hekate black screen (i can see separate thread on picofly aio read over 40 pages but have not found any conclusive problem)

-later found that due to the fact that the CPU heatsink is close to cmd it got shorted. which has been resolved but did not fixed the issue.
-Switch takes 0.08a steady draw on idle, apu capacitors shorted capacitors near ram shorted capacitors near 77812 shorted.

What i have noticed that once i shut down from the CFW (power>shut) the console shut and the picofly go through another sequence (i think it is blue to green on shutdown)? I think that first BSOD resulted once I have disconnected the SD card adapter before the sequence occurred then I was getting the BSOD after that.

Now the second board does go into NO SD - hekate result in black screen booting to ofw result in pico flashing dat0 fault with black screen

I have scraped the first switch for test purposes and noticed that the right row (looking where the flex is installed) all capacitors are 100ohm where working board has 2.5k ohm (even after all capacitors has been removed around apu and apu desolered from the board

Any help would be much appreciated - IMHO its apu short and its doner board now but would love to hear that i am wrong and it can be somehow saved