So, a long hiatus between band pass filter musings. In Part 1, I laid out several reference sources for band pass filter projects. Part 2 details the decision process, plus some notes on what happened with initial attempts at reproducing the K4VX Filter and NVARC filter projects.
Here I have a bit more detail on each project. After obtaining a small supply of the capacitors specified by the NVARC design, I’ve completed the 40m and 80m filters.
Measurements on the 40m NVARC filter show about 0.8db loss through the filter, with the SWR pass band covering the entire band easily. Outside the band, the SWR rises rapidly above 7.370 Mc, and below about 6.775Mc. That would seem to indicate the filter is resonant lower in the band, but it actually shows about 2 watts more attenuation at the bottom of the CW segment than it does at top of the SSB area.
The 80m NVARC filter also shows about 0.8db of loss through the filter. The SWR is about 1.5:1 across the entire band, and the filter seens to have its sweet spot right near the SSB DX window at about 3.775 Mc. That should prove fortuitous, since it is also where my 80m folded dipole resonates, but it is completely by chance.
These last two filters probably have slightly higher losses than they should due to construction techniques. The cases I had available for their enclosures were not ideal. Their assembly did not allow easy construction by the NVARC guidelines, and the coils are probably mounted less than optimally inside their cases. Through experimentation I found that slight variations in coil positioning had an effect on their insertion losses.
Additionally, the 15m filter began acting up, showing terrible losses. It turns out that in slinging it around the shield had become dislodged and was either in contact with or too close to one of the coils. Re-securing the shield solved the problem, and put the filter back very close to the NVARC spec’d performance.
By comparison, the K4VX set I have show less attenuation. I have K4VX versions for 20m, 40m, and 80m. Both the 40m and 80m filters show very low losses through the filters, both at about 0.3db of loss, only a couple of watts. The 20m version shows losses similar toits NVARC counterpart, in the vicinity of 0.7db. I expect to rebuild the 20m filter from this series using the ceramic caps rather than silver mica caps. It will be interesting to see if the loss figures change.
I have not yet built the 10m and 15m versions of this design, and may not. The attenuation specified by K4VX on these bandsis not as good as the NVARC spec. It might be worth trying the NVARC filter designwith toroids rather than air wound coils. An excellent experiment idea, and the 10m and 15m NVARC design seems to work well as described and reproduced here.
Given the low losses through the K4VX design, I may use that set for the run station, and the NVARC design for the mult station in an SO2R setting. The K4VX design is also physically much smaller, another practical advantage. The NVARC design has a better set of attenuation figures specified, but it will be nice to get actual measurements on the filters before declaring them a better choice. Some actual on-air testing can’t hurt either.
The coax stub project has been placed on the shelf for the moment. It is worth noting that coaxial stubs are probably better described as notch filters rather than band pass filters, as they are designed to place a notch on the harmonic or sub-harmonic frequencies. The book by W2VJN, “Managing Interstation Interference – Coaxial Stubs and Filters”is a treasure trove of useful information. For anyone with an interest in the subject of coaxial stubs, the book is worth every red penny of its price. Add it to your library and you won’t be displeased.
Previous in series: Band Pass Filter Fever – Untangling The Web – Part 2
Next in Series:Band Pass Filter Fever – The Kludgy Switch Box – Part 4.
Leave a Reply