Hacking Hardware Picofly - a HWFLY switch modchip

[abal1000x]

.

 

[QuiTim]

Well, this is my preliminary investigation results: a dat0 ripoff. @abal1000x @QuiTim
@Dee87

View attachment 377775

View attachment 377777

Well, my switch lite worked for four days after a "successful" pico install, full hekate's nand backup and etc.

In two different episodes I was some OS freezing... In the last freeze, I forced a power off, and since, black screen and the Led code was "*=* No eMMC block 0 read error" I checked that the wires had continuity, there was no grounding. So I removed the chip, the wires and now under the microscope I noticed that my dat0 was ripped out. I don't know if this happened during chip removal, after the black screen or if it was already like that. Could this be related to my black screen?

Thx for the help.

Well, this is good news, in a way
I would try a couple of things here to fix that trace.
1. Easy - Cover the surrounding points with kapton/solder mask and put a blob of solder there in hope that it will create a short between the 2 points (use good flux and sn/pb solder)
2. Medium - use this specific dat0 adapter (as in Oled) as it has no anchor points and then create a bridge from dat0 to one of the points where it need to be (check picture "a5" the blue X on top right)
3. Hard - use a 0.02 jump wire to try to fix the broken trace between two points.
Picture legend:
a1=dat0 in place in oled to give you an idea of size;
a2=dat0 that im talking about;
a3=how it should be placed;
a4=whats going on beneath;
a5=Lite board traces
This is pretty much what @abal1000x is saying but with pictures

 

[Uberfish]

https://gbatemp.net/threads/picofly-a-hwfly-switch-modchip.622701/post-10171602

4.7K and 47K

" Mine showed no value when tested"
If what you mean is
you test the resistance value and it is not measured,
than surely its damaged, need to replace.

Yep! Theres no value when measured. I tested diode to check it just wasn't a bad multimeter probe contact and there's a voltage drop. Does anyone have the size of these resistors?

Out of interest, since I'm not experienced with smd testing, why is there no resistance value but still a voltage drop? I guess it's because the resistors are connected to the board and I'm measuring the drop across ground to the other component?

 

[abal1000x]

.

 

[floxcap]

DragonFly

Just got my dragonFly boards in, I'll let you guys know how they go
Also I made some changes and have ordered some to be assembled. If everything goes well you guys can have some to test.

• Dual MOSFET onboard in single package
• CMOS oscillator
• All components are virtually the smallest available package size
• Ideally would be made of 0.1mm flex, but 0.8mm/1mm FR4 is cheap and good enough for this revision
• 22mmx13mm(7/8"x1/2"), final version will be 15mmx10mm(5/8"x3/8")
• I think I'll sell them for $15 AUD ($10 USD)
• Most importantly, rounded traces

View attachment 377641View attachment 377639

Nice job. Do you think there are any risks with putting the mosfet(s) on-board?
Personally I'd be concerned with current/ground/capacitance issues - I think it's better to have the mosfets as close as possible to the CPU to mitigate these problems.
It'll be interesting to see how you go with these - especially over the longer term - good luck

 

[CodyRoseman]

Nice job. Do you think there are any risks with putting the mosfet(s) on-board?
Personally I'd be concerned with current/ground/capacitance issues - I think it's better to have the mosfets as close as possible to the CPU to mitigate these problems.
It'll be interesting to see how you go with these - especially over the longer term - good luck

Only one way to find out. I think it should be ok. the apu only ever sees 2amps during full load, and that current isnt going near that capacitor and especially not 2amps during boot.

 

[floxcap]

Well, this is good news, in a way
I would try a couple of things here to fix that trace.
1. Easy - Cover the surrounding points with kapton/solder mask and put a blob of solder there in hope that it will create a short between the 2 points (use good flux and sn/pb solder)
2. Medium - use this specific dat0 adapter (as in Oled) as it has no anchor points and then create a bridge from dat0 to one of the points where it need to be (check picture "a5" the blue X on top right)
3. Hard - use a 0.02 jump wire to try to fix the broken trace between two points.
Picture legend:
a1=dat0 in place in oled to give you an idea of size;
a2=dat0 that im talking about;
a3=how it should be placed;
a4=whats going on beneath;
a5=Lite board traces
This is pretty much what @abal1000x is saying but with pictures

Point #3 - Haha - yep - I ended up with ripped pads when trying to figure out why my first lite install didn't work (solder, remove, re-solder etc)... except in my case it was both pads - so I had to dig down a layer to get to the copper below for both pads and re-join them
It worked - but not at fun thing to have to do.

 

[CodyRoseman]

Nice job. Do you think there are any risks with putting the mosfet(s) on-board?
Personally I'd be concerned with current/ground/capacitance issues - I think it's better to have the mosfets as close as possible to the CPU to mitigate these problems.
It'll be interesting to see how you go with these - especially over the longer term - good luck

the parasitic capacitance from the ground planes should be completely negligable for this application, but the resistance and current capabilities of the trace width might. Ill find out

 

[chronoss]

a3=how it should be placed;

On the pic, a3 is a Switch Lite model, that i know, no need a DAT0 adaptator on switch lite...

 

[Switxh]

I've got a Lite doing the same and it doesn't seem like *you've gotten an answer, correct?

Yeah unfortunately I never found the cause or a solution to the problem. Every other Switch I have done had a Samsung EMMC. That OLED with the slow / random glitch times was the only SK Hynix I have done. I am inclined to put it down to the SK Hynix chip just sucking. Every other Picofly install I have done the glitch times are almost instant. 1-2 seconds at most.

 

[Uberfish]

Reference to your uploaded picture, it is the left resistor or the right resistor which damaged?

Both, actually I initially thought I maybe overheated them during the soldering but I only tapped with the iron, not held it there. So I'm wondering what could cause them to blow...

 

[QuiTim]

On the pic, a3 is a Switch Lite model, that i know, no need a DAT0 adaptator on switch lite...

you are right but we have a specific situation here. the dat0 on pcb has been damaged and we are going through ideas on how to fix it. read the posts above that one and you'll get the full picture

 

[abal1000x]

.

 

[RiotRetroGaming]

yes, but cutting that side fits better under emmc?

Post automatically merged: Jun 13, 2023

I've heard that those adapters give problems

I've not had any issues, I'm not sure what all the fuss is about. I get 400-690 on diode tests no problem?

 

[Switxh]

I've not had any issues, I'm not sure what all the fuss is about. I get 400-690 on diode tests no problem?

The diode reading on DAT1 and DAT0 is the same. Even if you bridge them you will still get the same diode reading. Bridging DAT1 and DAT0 is no joke and will cause corruption.

 

[RiotRetroGaming]

The diode reading on DAT1 and DAT0 is the same. Even if you bridge them you will still get the same diode reading. Bridging DAT1 and DAT0 is no joke and will cause corruption.

Ah ok, that gives a bit more clarity

Post automatically merged: Jun 14, 2023

Ah ok, that gives a bit more clarity

I've just been given this to do... this should be fun!

 

[NandieeHU]

What this could be? Readings fine, RST is good.

 

[cicci084]

1) Hi guys, today they deliver this console to me, I will update you on the solution.

2) In the meantime, I have a solution with this console which side are the EMUMMC? (click for help me, more info)

3). I also installed a multicolor picofly (cables), in my opinion the cabling (especially CPU) lengthens unnecessarily.

 

[Uberfish]

When you asked why that could happened.

I assume your soldering is okay, no short circuit, then one possible scenario is...
You inserted a conducting debris to the emmc when inserting the adapter. The right resistor is connected to the Data Strobe pin of the emmc, which near the vss
View attachment 377828

I'd be super surprised if any debris got that far under the emmc, I'm not using any sprays or inserting the adaptor with force. I haven't inspected the CPU flex cable yet, it was suggested that could cause issues if there's a short somewhere around that. After I installed it I did continuity testing, and it passed. Any idea on that?

 

[NandieeHU]

Just come back from lunch and works perfectly fine.