Hacking Hardware Picofly - a HWFLY switch modchip

  • Thread starter Thread starter mathew77
  • Start date Start date
  • Views Views 3,669,773
  • Replies Replies 17,052
  • Likes Likes 15
Question on reballing the eMMC, what temps and air settings are you guys using? I've been using 380 @ 50% and go nice and slow. I've heard the eMMC can be damaged by heat so didn't want to go too hot.

160 for a minute, 265 for a minute, 370 until the emmc comes loose which usually happens after around 40 seconds. 20% airspeed always (but this varies depending on the station.)
 
he was refering to all switches in Generel, if u dont know what D and S point is check the PicofFly Guide

I managed to attach CMD point. I tugged a little, but the connection seems a tiny bit loose. I don't want to pull too hard, but it does appear to be attached with the moderate force I have put on it. Should I just leave it be?
 
From what I can see.

I am getting BSOD after trying picofly. I just applied solder to tge solder points and thought I will complete it in daylight.
Went through the forum, cleaned my soldered points and flux but still BSOD.
Attaching images, please someone have a look at it.
 
About your issue.
Since you could goes to hekate, then i think the dram connection should be okay.
I remember someone whose dram trace got disconnected, cannot even goes to hekate.

Have you tried to reball the emmc, or use the permanent method?
If you are not confident, dont do it.
Reball, indeed need basic skill and appropriate tool to safely executed.

No reball, as I don't have a decent BGA system.
I took the DAT0 adapter and have already exchanged it for another one. Unfortunately, the problem is still the same. The question is, why did the switch crash shortly before the end when creating the Nand ? Maybe a short circuit between DAT0 & DAT1 and now the nand is corrupted ?
 
Hey guys! Got an OLED with SKHynix, did eMMC reball and with RP2040-zero (3x47Ohm) + FW2.73 I sometimes get slow glitch time (up to 15s and other times 2-3s). Tried adding extra 47 Ohm on CLK and CMD, but the timming did not improve. Once booted, the console works well and the eMMC has the correct speed.

Is this SKHynix "stubborn" or is there some room to improve? What was "weird" as well is that usually I read between 0.450V-0.5V on eMMC (red on GND), but on this console they're around 0.55V.
 
Hey guys! Got an OLED with SKHynix, did eMMC reball and with RP2040-zero (3x47Ohm) + FW2.73 I sometimes get slow glitch time (up to 15s and other times 2-3s). Tried adding extra 47 Ohm on CLK and CMD, but the timming did not improve. Once booted, the console works well and the eMMC has the correct speed.

Is this SKHynix "stubborn" or is there some room to improve? What was "weird" as well is that usually I read between 0.450V-0.5V on eMMC (red on GND), but on this console they're around 0.55V.
Its not only skhynix, almost all mariko have that weird issue.
Average on 3s but sometimes its just out of sync and re-search the glitch point again which resulted the long glitch time.

The only differences between erista and mariko (beside the cpu ofcourse) on the 1V line is that in marista there is EM filter put on, while on mariko doesn't have it. Maybe this could result that weird random time issue, but i am not convinced enough on that reasoning.
 
  • Love
Reactions: Phantomas77
I always choose the lesser risk.

My step is first paste and reball the bottom adapter. If the bending is negligible than goes to reflow it to the mainboard pcb. Then check the continuity of cmd and vdd. Then paste and reball the emmc, then reflow to the adapter. Then check again the continuity of cmd and vdd to the main board.

I choose this order because this is the lesser risk.

If i Reflow the emmc to the adapter first. If its failed then i need to repeat the reballing emmc and reflow it to the adapter again. This might broke the emmc if i do it again and again.

I prefer reball the adapter first. If its failed, the adapter itself is cheap only $1, not much loose. And the mainboard pcb are rigid, its >5layer. So its not easily broken compare to the emmc.

I always choose lesser risk and lowest price as much as possible. Complexity is not a concern for me. As long as the risk is less.
I told a flex is better since it doesn't have that short coming of having a rigid structure, though the only problem here is the price since you have to order bulk to make it worthawhile otherwise they will charge you an arm.
 
I told a flex is better since it doesn't have that short coming of having a rigid structure, though the only problem here is the price since you have to order bulk to make it worthawhile otherwise they will charge you an arm.
Subjectively i prefer green adapter than the flex.

If you want to reduce the flex price, try to create it yourself on kicad and manufacture it on the jlcpcb.
 
  • Like
Reactions: Dee87
Subjectively i prefer green adapter than the flex.

If you want to reduce the flex price, try to create it yourself on kicad and manufacture it on the jlcpcb.

Jlcpcb butchered the deign I sent them. I went with another fab but the price was steep.
 
I am getting BSOD after trying picofly. I just applied solder to tge solder points and thought I will complete it in daylight.
Went through the forum, cleaned my soldered points and flux but still BSOD.
Attaching images, please someone have a look at it.
Missing resistor. I believe it's 100 ohm
 

Attachments

  • 1694610006008.png
    1694610006008.png
    2.3 MB · Views: 150
I changed from flex cable to backside mosfet installation which works fine. Just in case I will ruin the cap near the mosfet on an OLED. Can anyone help me what cap it is that I can order some spare sparts ?
 

Attachments

  • Oled_Mosfet.png
    Oled_Mosfet.png
    192.4 KB · Views: 125
I changed from flex cable to backside mosfet installation which works fine. Just in case I will ruin the cap near the mosfet on an OLED. Can anyone help me what cap it is that I can order some spare sparts ?

or you can do this install instead
 

Attachments

  • 2020_1007_111737_059.jpg
    2020_1007_111737_059.jpg
    3.6 MB · Views: 143
  • Like
Reactions: QuiTim and cherup
Great! Thanks for that info. This is only working for OLED or are there also safer methods for V2 / lites etc. for the mosfet installations ?
there are, but you normally wouldn't remove the motherboard on those so it adds complexity in other ways instead. Here are two alternatives for the switch lite
Great! Thanks for that info. This is only working for OLED or are there also safer methods for V2 / lites etc. for the mosfet installations ?

See this post https://gbatemp.net/threads/picofly-a-hwfly-switch-modchip.622701/page-579#post-10200708
there should be similar positions for v1/v2, I didn't check. Having to remove the motherboard makes it not worth it for me
 
  • Like
Reactions: QuiTim and snaker

Site & Scene News

Popular threads in this forum