I've been puttering around with end-fed half-wave antennas for 40 through 10 meters (concentrating on 40 for now), and have had trouble coming up with a good match-box design. I've tried different ideas from an array of online sources ranging from auto-transformers in various turns ratios to basic L-match circuits. I finally found a design which, at least in the test phase (using fixed-value non-reactive loading resistors) shows an extremely good SWR spread. I do not use any caps at all in the design, just a mix-43 torroid and a 3 to 17 turns ratio transformer. Using a vector network analyzer and a 10-inch jumper, with a test load of 3.3k ohms, the SWR is 2.1 or below from 160 meters all the way through 10 meters. With a load of 2.2K, the SWR is about 1.7 on 160 meters, and 1.5 or below on 80 through 10 meters. With an 18-foot cable connected, the readings with a 2.2k load were 2.1 on 160 meters, and under 2 on 80 through 10. With 18' and a 3.3k load, the readings are 1.6 or below across the HF band. With 18' and a 1.5K load, the readings were again 1.6 or below across the HF bands. With a 4.7K load the SWR spread was quite a bit higher, with 160 meters being around 2.9, 80 meters at about 2.5, 40 meters at around 2.5, 20 meters around 2, and the rest of the band at under 2 with 10 meters being in the 1.6 or below range. I have not put the transformer into use with a practical antenna installation yet, but of all the designs I've tried, this one shows the most promise.
I plan on posting updates.
I plan on posting updates.