After finally getting the power supply of the existing miner back on line, we found that the FOR-A
experimental cards had gone dead.
In fact Two FOR-A cards had gone dead.
Both cards have exactly the SAME fault the 3V3 rail is shorted to the majority of the 5V line
It looks like the Shitty Japanese PSU design is back with a vengeance (and everything was working beautifully together).
Since both a modified and un-modified card went defective, it seems logical that there is some sort of underlying PSU design fault in the FOR.A product.
We will need to spend a bit of time resolving this.
Ok 1 hour later……
SURPRISE!!!,The root cause is the USB cable…. (yep for all those people throwing hubs and cables randomly at their computer in an attempt to maximize the connections)
So how does a USB cable take down a PSU
If you take a look at the way a USB plug /socet is constructed, the external shield wraps ROUND the plug, but because like everything else it is made from the cheapest crap metal available.
The wrapping process leaves a visible & physical break in one side of the plug.
now here is the pinch……
IF the join is on the BACK of teh plug BEHIND the white insulator holding teh connections, then there is no problem
but IF the join is in the FRONT directly inline with the connections, this allows the join to move releasing pressure on the parts that hold the USB connection together.
Now since USB was NEVER designed to operate with a direct face to face connect/disconnect (take a look at a USB connection they are staggered so as they make contact in a certain sequence), this causes the USB devices to flip in & out of connection and in doing so , it reverse bias some of the hubs leading to a direct short of power into the connected RS232 devices.
So the motto of the story is…. Don’t go building massive trees of USB devices., all it takes is one shitty cable design too give you a bad day.