One of the big issues around the Pi supply problems, was that an assumption was made that the Phone chargers met the USB standards.
If you dropped below 4.8v on the 5v rail, the video stopped.
The reality was these were a nominal 5v supply used to provide a charging source for a 3.6v battery, and many were below the USB spec.
Coupled with the additional voltage drop across the Polyfuse, some had issues.
The relatively low rating of the Polyfuse that supplied the USB, also resulted in non compliant USB outlets.
Despite the Open Source the actual hardware design wasn't released, but it was confirmed that the GPIO and USB, had insufficent track widths to support higher currents direct to the 5v supply rail, so there were the odd parallel supply designs bandied about.
The other problem was that many people assumed that once the load was removed, the polyfuse was happiness filled and reset itself ....WRONG
It uses heat to trigger the high resistance and needs to cool down over a period of time (hours in some cases).
Hence after triggering it, people didn't wait, and the problems were compounded.
This is not their fault, and the instruction book was ....
A polyfuse is going to provide protection, but the silicon devices are still usually quicker.
(Just like the dropped drink about to land on the carpet ...gravity is nearly always faster ..especially if its red, and the carpet is light coloured)
Inline tool.
I personally wonder if having an optional USB connector board, that includes the zeners, resistors, pullup and a protection device (along with a real USB plug) might be an alternative construction approach. I know that there are a few designs that use slightly offset pins on the programmer to allow programming without fitting a socket on the 'host'.
This design may allow the use of thinner PCB's rather than the thick version required for the USB connection, and removes the need for gold ...
mark
