This is a picture of the first generation of the 6m antenna.
The first generation uses a very simple connection between the feeder and the antenna elements. As seen there is no balun. This of course works in a way but I got a lot of RF back on the outside of the cable as one would expect. (I use wet-room silicon & shrimp cable to seal and protect the soldering joints. But please note that the picture was taken before the sealing was done.)
A second attempt was done with an added cable balun. The balun consists of a 1/4 λ cable along the feeder. The cable is soldered at one end to the center conductor of the radiating elements and soldered to the braid at the other end of the outer conductor of the feeder. It works as a "voltage balun" which is probably not the best if the antenna elements are asymmetrical. There is also a ferrit just below the point where the cable is attached.
Regarding different baluns there is an interesting article how the baluns work on the following page (Eznec SW). It describe the difference between "voltage baluns" and "current baluns". It is not maybe what you would expect...
A measurement showed that the antenna was resonant at 45 MHz. This explains why the VSWR was not so good during the first tests without the balun. It also might explain why there was a lot of RF on the outside of the cable despite the cable balun during the second test.
After re-tuning the antenna has suddenly it's resonance a little too high, about 500 kHz. But the worst thing is that there is still some RF on the outside of the cable. It is probably so that the cable balun is not working in an optimal fashion i.e. as expected. The best way to fix it is to make a toroid balun. Unfortunately the balun I am using for the 15m band antenna is not so good at the 6m band. (Amidon T200-2)
After using the antenna during a Sporadic E in mid May 2021 it is obvious that it performs very well despite the RF leakage back on the cable.