Hacking Hardware Picofly - a HWFLY switch modchip

[jmmc]

So,this situation is normal?

I don’t know if it’s normal but when I only had one resistor speed of 16mb was around 200mb/s and when it finished the benchmark it said that emmc was in slow mode. I added one more resistor and the speed increased to more or less your results and the message was gone.

 

[cgtchy0412]

Hello guys, I wanted to order the resistors to use as pulldown, which value and size is recommended? What are there on hwfly flats?

I personally use 10K for pulldown now. As i once use 1K but then some of the unit sometimes boot directly to ofw (glitch failed) and sometime boot hekate. Maybe its specific by mosfet type as i use AON.

 

[abal1000x]

.

 

[sergiochendry]

I got wierd switch v2
Yesterday i instal picofly
All perfect, try to reboot many time
And today, something wierd happening
After turn off, swith cannot turn on again
I left about 2 hour, and switch turn on normally
So i remove pico and every cable
Still the same, cannot turn on
Plug in usb charger only draw 0.2-.03 ma
I left switch while charging maybe like 1 hour
Turn on normally to ofw and usb draw 1,6a
What happen with this switch?

 

[cicci084]

https://gbatemp.net/threads/picofly-a-hwfly-switch-modchip.622701/post-10176878

1k for each mosfet, did I get it right?

 

[abal1000x]

.

 

[Danook28]

I got wierd switch v2
Yesterday i instal picofly
All perfect, try to reboot many time
And today, something wierd happening
After turn off, swith cannot turn on again
I left about 2 hour, and switch turn on normally
So i remove pico and every cable
Still the same, cannot turn on
Plug in usb charger only draw 0.2-.03 ma
I left switch while charging maybe like 1 hour
Turn on normally to ofw and usb draw 1,6a
What happen with this switch?

Some time you need disconect battery charge soket from the conector then conect and some time press 15 sec and then perss one on power. Me be your pico fly glitch is or ps1 sp2.

 

[sergiochendry]

Some time you need disconect battery charge soket from the conector then conect and some time press 15 sec and then perss one on power. Me be your pico fly glitch is or ps1 sp2.

I already did that, still the same
Even i press restart
Switch can’t turn on

Post automatically merged: Jun 6, 2023

Seems the power problematic.
Might check this chip
View attachment 376210

Datasheet:
https://www.alldatasheet.com/datasheet-pdf/pdf/879766/ROHM/BM92T30MWV.html

Like short or just ic nearly dead?

 

[Danook28]

I already did that, still the same
Even i press restart
Switch can’t turn on

Post automatically merged: Jun 6, 2023

Like short or just ic nearly dead?

Is charge cable charge when you flip it???
Turn the charging cord over and see if it charges from both sides, from the top and bottom.

 

[Ragna666]

Hi guys!
Can someone recommend me a a good soldering wire that doesn't have rosin core in it?
And a cheap but good soldering iron.

I done a v2 on saturday but because of the shiiit quality rosin core tin and weak soldering iron instead of max 1hour, it became 5hours of misery.

 

[Kiuxn]

Lets analyze logically.

When the data is big, the rate is high. But when the data is small, the rate is dropped. The difference between the two is, when the data is big, the request to the CMD is not busy compare to when the data is small. The request is only once or twice, then the data stream on to the Dat line. So the DAT line is okay, the CLK is okay, because when its transfer big data the rate is fast.

So i suspect the CMD line is noise, so there will be checksum, error, retry, checksum error, retry until it succeed, then when success the small packet goes to DAT line. And repeat again and again. This reduce the rate of data read from DAT line.

Here is my 1 cent suggestion:
1. Change the cable of the CMD with new one. Maybe the old cable problematic.
2. Try to clean the flux or any dirt around the CMD point.
3. In lite, the CMD point is directly goes to the emmc and cpu with some resistor paralleled to GND, you might check that resistor, it might broke.
View attachment 376193
If you use other device track it yourself in here
https://balika011.hu/switch/
4. Try to reduce little bit the resistance value. Use 33-40 ohms.

I checked the resistor, rewelded and replaced the cable, and the problem is still the same.

 

[abal1000x]

.

 

[jmmc]

Theres two left:
1. Try lowering the resistor CMD in pico to 30-40 ohm.
View attachment 376220

2. Reflow the emmc.
Maybe theres copper/tin debris goes under the emmc that short circuiting some pin.

Which results do you have on your benchmark? Can you post an image?

 

[cicci084]

Yes 1k for each mosfet, parallel.

By calculation it should be around hundreds kilo. I calculate around 300K, but i still not yet confirmed it.

When the pico run, the firmware will configured the gpio as pulled down, read from rp2040 datasheet its around 80k ohms.

Post automatically merged: Jun 6, 2023

so what should i order?
"confused"

 

[abal1000x]

.

 

[Adran_Marit]

Do we have any confirmation if the HWFLY rp2040 chips work with the firmware or if they have used their own firmware?

 

[Danook28]

Do we have any confirmation if the HWFLY rp2040 chips work with the firmware or if they have used their own firmware?

You can use it after buy it install one in my friend console lite with out add any firmware and it is working.

 

[Kiuxn]

Theres two left:
1. Try lowering the resistor CMD in pico to 30-40 ohm.
View attachment 376220

2. Reflow the emmc.
Maybe theres copper/tin debris goes under the emmc that short circuiting some pin.

Switching to 33ohm, the emmc shield has never been opened, and the same result is obtained.

 

[Adran_Marit]

You can use it after buy it install one in my friend console lite with out add any firmware and it is working.

Yes I know they have firmware on them, what I'm wanting to know is if it is the firmware from here of if they changed the pinout so it doesn't work

 

[abal1000x]

.