As to images, for me, the image found here: http://www.cqcqde.com/shop/88_108.html is most informative.
Referring to Martin's blog, it seems that the "matching" device actually is a 6:1 UnUn ferrit toroid transformer with a 600Ohm (120W) termination resistor. The aerial being rated 250W, this means that up to 50% or 3dB of power is burnt in said resistor. The bandwidth of the aerials is mentioned to be 2Mhz to 30MHz.
Martin mentions the inferior performance of the 6:1 UnUn the why it was wound in the original BB6W. To my believe there is actually some thought in this. Japan, that where the aerial comes from, is a very noisy, QRM that is, country. Hence, Diamond seems to have decided to compromise power coupling versus common mode noise by reducing the capacitive coupling between the windings.
My interpretation of this aerial comes close to a short end-fed random wire, which is broadened by a sort of termination known from aerials like the T2FD, with a twist however.
Here is what I believe is interesting about the concept. Usually end-fed random wires are used in combination with 9:1 UnUn transformers, which result in a feed point impedances of 450Ohms. The BB6W matching unit however transforms to 300Ohms only, where the termination is twice that. Obviously, one does not want 300Ohms of termination, that would just form a dummy load...
So, what we got here is a short broadband aerial, horizontal or vertical in which half the power is absorbed.
In QRO, the tricky bit is the non-inductive resistor. It is doable, and certainly to a lower price than Diamond sells the aerials.
In QRSS, which essentially is QRPP, we are not troubled by the problem mentioned above. In fact, we could use a regular 1/4W carbon resistor and still we would be fine.
Two things came to my mind
- use a 9:1 UnUn with a 900Ohms resistor
- use an air-core (auto-) transformer
It would also not be too difficult to build a termination resistor being good for 10W or so....
Last thing, what way of winding the UnUn would be most appropriate? You decide... Diamond's way would provide you with less RX-QRM, Martin's method would pass more TX-power to the wire.