Hacking Switch OLED HWFLY v3 - stuck on training

[Akinci]

Hi everyone,

I have a oled switch with a hwfly v3 mod fitted by a friend. We cant get it to train, no matter what we try. It stays on flashing yellow, I have left it for 30 minutes+ a couple times. All the points have been measured during the install and values are within tolerance following the sthetix install guide.

I have tried flashing the newer hwfly firmware so currently have 0.7.1 flashed, have also tried 0.7.0 and 0.6.2 with same result. Strangely 0.6.2 gets stuck at flashing orange light. Also done several hwfly chip resets in between these attempts.

Please note the switch still boots OFW.

Does anyone have any ideas? Or is there anyone on the forum I can send the switch too to have it checked over?

Please see a quick video of where it gets stuck on training and close up image of the install.

Thanks,

 

[Bonovox40]

This sounds like something I had happen recently w an OLED mod, however was using a v4 chip. After the initial install, chip was rapidly flashing yellow a few times, with a yellow pulse at the end of the pattern. Then it would almost be too fast to see red flash, the repeat that yellow pattern. Screen was black also, would never turn on.

I ended up flashing the chip to 7.1 fw, and no change.
Took it to a second installer, he noticed that a capacitor near one of the solder points seem to be damaged, and tried replacing it with a spare one but no change. He also noticed that the CLK pad that you scratch off to get access to appear to have been scratch down too far. He ended up exposing some of the trace lines, and soldered directly to that spot. Still no change.
He ended up removing the chip and the switch won't boot at all, even though it had a charge before he put on the chip. So either the capacitor or board or emmc damage ruined the switch.
A few days later I got a new Switch and brought it to the second installer, since he knew what he was doing. He connected the chip (without that new 'h' shaped flex cable helper piece) and the adapter, and it works fine. When I turn it on it flashes purple then goes to Green before the Hekate screen shows up.

I would check the dat0 adapter, and also the clk point again. If you're getting values for everything that should have diode multimeter reading, then I would disconnect the chip from all the solder points, and make sure the switch can actually boots into stock firmware. If not, you may have damaged it in a similar way like mine, and may need more serious repair or replacement.

 

[Akinci]

This sounds like something I had happen recently w an OLED mod, however was using a v4 chip. After the initial install, chip was rapidly flashing yellow a few times, with a yellow pulse at the end of the pattern. Then it would almost be too fast to see red flash, the repeat that yellow pattern. Screen was black also, would never turn on.

I ended up flashing the chip to 7.1 fw, and no change.
Took it to a second installer, he noticed that a capacitor near one of the solder points seem to be damaged, and tried replacing it with a spare one but no change. He also noticed that the CLK pad that you scratch off to get access to appear to have been scratch down too far. He ended up exposing some of the trace lines, and soldered directly to that spot. Still no change.
He ended up removing the chip and the switch won't boot at all, even though it had a charge before he put on the chip. So either the capacitor or board or emmc damage ruined the switch.
A few days later I got a new Switch and brought it to the second installer, since he knew what he was doing. He connected the chip (without that new 'h' shaped flex cable helper piece) and the adapter, and it works fine. When I turn it on it flashes purple then goes to Green before the Hekate screen shows up.

I would check the dat0 adapter, and also the clk point again. If you're getting values for everything that should have diode multimeter reading, then I would disconnect the chip from all the solder points, and make sure the switch can actually boots into stock firmware. If not, you may have damaged it in a similar way like mine, and may need more serious repair or replacement.

Thanks for the quick reply mate, I'll pass this onto my friend to see if we can check what you mentioned. I updated the OP as the switch still boots OFW.

 

[l7777]

Two things.

That photo looks like a v4 chip.

Training issues are usually a DAT0 issue with the OLED Switch.

 

[Akinci]

Two things.

That photo looks like a v4 chip.

Training issues are usually a DAT0 issue with the OLED Switch.

Two things.

That photo looks like a v4 chip.

Training issues are usually a DAT0 issue with the OLED Switch.

Hi, yes sorry think your right it's v4. Is there anything I can check or do on the DAT0 to get it working?

I will likely need to try and find someone to check it over if I can depending on the level of work needed, so will pass the information on. I can barely solder myself so haven't got much hope lol.

Thanks,

 

[randy_w]

Is there anything I can check or do on the DAT0 to get it working?

Get a multimeter and measure the voltage drop of C point of the modchip. Set multimeter to diode mode, black probe on ground and red probe on point C. You should get a reading about 0.6-0.8V. If it's open circuit then DAT0 adapter is loose. Just take it to your friend or a repair shop and ask them to reflow the emmc chip, so the adapter's contact point is soldered to the solder ball and it will never come loose again.

 

[Akinci]

Thanks for help with this, just an update: checked the DAT0 values and they were ok. Thought I'd double check all the other values while I was there and found the CLK cable has come off, it must have come loose when putting it back together but all other diode values were within tolerances. Just need practice soldering before I re-solder it back in place as I have next to no experience.

So just a tip for anyone with the infinite training pulsing yellow light, check your diode values. I'm going to solder this back in place and hopefully it will work now.

Encountered another issue though using the hwfly 0.7.1 firmware, which didn't happen on 0.5 or might have been an earlier version as it was updated around March 2022. On the earlier hwfly FW versions the training will fail about 30 seconds in for me at least and boot OFW but on 0.6.1 and above it stays on training forever even if you disconnect the CPU ribbon cable but led changes a flashing orange.

 

[OhItsSlick]

Thanks for help with this, just an update: checked the DAT0 values and they were ok. Thought I'd double check all the other values while I was there and found the CLK cable has come off, it must have come loose when putting it back together but all other diode values were within tolerances. Just need practice soldering before I re-solder it back in place as I have next to no experience.

So just a tip for anyone with the infinite training pulsing yellow light, check your diode values. I'm going to solder this back in place and hopefully it will work now.

Encountered another issue though using the hwfly 0.7.1 firmware, which didn't happen on 0.5 or might have been an earlier version as it was updated around March 2022. On the earlier hwfly FW versions the training will fail about 30 seconds in for me at least and boot OFW but on 0.6.1 and above it stays on training forever even if you disconnect the CPU ribbon cable but led changes a flashing orange.

Were you able to solve the issue?