Originally Posted By: acharnley
6V is the real-life voltage once you apply a load, there is no point in testing 20V, 40V, 60V which indicates there is no load attached.


You limit to 21V, so you will convert additional energy into heat (unloaded/partial loaded case). A lot of heat at faster speeds, which explains why you need the metal heat sink while most of the others work fine without. It also severely limits the lifetime of all components within the case. Elkos lose half of their lifetime with every 10°K increase. Your 25-year-stuff can and will die faster then expected.

Originally Posted By: acharnley
As the pdf shows, the testing begins at 500Hz. This value was chosen because the The Plug 3 cycles off at 250Hz.


The Plug was designed to work with hubdynamo, not an amplifier. I guess that triggers various safeguards. Looking at a real signal and comparing to the amplifier output might show why.

Originally Posted By: acharnley
In other words, this test is not a quantative test of a converters output at a given speed, but rather when there is enough power supplied to provide a 500mA output, what is the efficiency transfer of the converter.


My bike can accelerate with ~9.81 m/s² if i drop it of a cliff. I would not take that as a test of my ability to it accelerate on a flat road.