evnchn1. (Further understanding of QC3.0 protocol, contents of old Post may not be valid) Took that power bank apart and looked up the chip number. The chip was little buggy the manufacturer admits, as it outputs 7.3V upon entering QC3.0 continuous mode. It is still in the specs though. I can charge other device because those device observe the voltage and request higher / lower voltage accordingly. It seems like Zenfone 6 does not observe the voltage given by the source supply and requests higher voltage rather blindly (requests 20 times +0.2V from 5V to 9V). Try and change that. This can have extended benefits as some cables which have high resistance lose some voltage, and observing the input voltage at the phones basically compensates for that.This power bank does not obey Quick Charge standards and outputs 12V, knowing that this phone only have 9V input, you know the consequences. Once killed my phone, luckily it was within 10 days of purchase and I was able to return it after it self-recovered for 2 days and refuse to fast charge anymore. Using my phone I can verify it bugs out all the time. If you have this power bank, stop using IMMEDIATELY / only use slow charge port.
Meanwhile, ASUS, can you do some software safety to guard the device from problematic chargers? Maybe you can stop the the charging when voltage requested by QC does not match power bank real output voltage and does not trigger hardware safety.
Warning, Killer power bank watch out, awaiting fix
2. The phone cranks down the charging current if CPU/GPU working and heat dissipation is not enough. Perhaps allow the user to set charging to take priority in lighter processor loads, so that the CPU and GPU clocks gets cranked down instead, maintaining the fast charging.
If you don't understand what is happening from my observations, you can comment below
I would like to add to the points as it seems that 18W QC3 is bugged/capped in terms of max power. My power meter always recorded 13W at max in stock QC3.0, whereas I was able to pump 17W to the device by first tricking the device into 10W 5V2A mode first, then giving it 9V to get 9V2A ≈18W.
Refer to yellow line for voltage.
The first images shows the faulty powebank giving out 7.3V when entering continuous mode (sloped yellow line). I had to unplug the device to prevent further damage.
While the second image the slope starts at 5v properly
The benefits of implementing the suggestion of mine does not stop with increased charging speed, better compatibility and compensation for loss of voltage in cables. Since adjusting voltage by step takes 2ms in optimum scenario, and that the current implementation waits 2000ms for every step, it would mean theoretical 1000 times faster for getting 9V from charger.
Asus battery department, if you are listening, please get working. What I'm afraid is that the same kernel code gets reused time after time, and every device will be like this.
Thanks very much.
Refer to yellow line for voltage.
The first images shows the faulty powebank giving out 7.3V when entering continuous mode (sloped yellow line). I had to unplug the device to prevent further damage.
While the second image the slope starts at 5v properly
The benefits of implementing the suggestion of mine does not stop with increased charging speed, better compatibility and compensation for loss of voltage in cables. Since adjusting voltage by step takes 2ms in optimum scenario, and that the current implementation waits 2000ms for every step, it would mean theoretical 1000 times faster for getting 9V from charger.
Asus battery department, if you are listening, please get working. What I'm afraid is that the same kernel code gets reused time after time, and every device will be like this.
Thanks very much.
