Part 4 of the W4KAZ filter project series comments on the process leading up to the integrated box full of NVARC Ugly filters for use in the KazShack. The quest continues.
Notes: Link to photos of the project at bottom of this page. If you want to read about the project from the beginning, go to the”Band Pass Filter Fever” series page.
Part of the project goal is to put all of the NVARC filters into a switched box to allow for SO2R and use at Field Day and on IOTA expedition. The original idea was to use a simple rotary switch. Somewhere along the way the idea morphed into using relays set up to allow control from a band decoder.Toying with the relay switching idea brought up a couple of issues that I chose to avoid. Instead, the individual filters were tied together with a two pole ten position switch.
A previous project resulted in a seven position remote antenna switch.That switch is lossy on 15m and 10m because of the point to point dead-bug style wiring. I didn’t see an easy way to avoid this problem, and I’m not set up for PCB design/manufacture. Using PCB’s and strip line runs would solve the issue. I have an idea for making strip lines that may work, but it is a bit Rube Goldberg-ish, so I chose to shelve that temporarily.
So, back to the rotary switch. I had a 2-pole 10 position switch in the junk box. The contacts are silvered brass, and seem beefy enough for the job, so I tested it out by wiring up the input/output to a bypass position.
Ick. Needless to say, it is not an ideal solution. The loss through the switch alone on 10m is about 0.6db. Losses are lower on 40m and 80m, just barely measurable.
So, WTF. I used it anyway. More suitable switches are a bit pricy if bought new, and this one was already in my sweaty little hands. Impatience, “good enough”, and zero cost won out over quality. Engineer the possible.
After all of that hand wringing was done, some other practical construction choices needed to be thought out. The end goal of constructing a switched filter box could have been implemented in several ways.
Method 1: Use the existing set of filters, switching them in via an external switch box, all connected via a rat’s nest of coax jumpers.
Method Zwei: Build another set of filters into a larger enclosure, and incorporate the switch into the new design.
Method III: Use some less aesthetically pleasing choice that will also benefit from poorly conceived and hacked together engineering practices.
Well the choice was clear – use Method III!
The rationale unfolded as a matter of “least inconvenient compromise” rather than “optimal design”. I was limited in the number of parts available. That was the primary limiting factor for method 2, not enough capacitors of the proper values on hand for a full second set of filters. Keeping the individual filters available was desirable for the sake of future flexibility, so ripping them apart and re-assembling was not considered.
Parts count also played a part in ruling out method 1, as it would use up 20-plus pl-259’s, plus the coax.
My compromised solution was to use the individual filters with a slight modification. Rather than remove the so-239 connectors, I merely tacked on a pigtail of coax for the runs to and from each switch. It is a compromise in every way, electrically, mechanically, and aesthetically. But it sure was simple.
It also seemed to work electrically better than I expected, as none of the loss figures vary substantially from the losses I would expect from the switch plus those of the original individual filters. In other words, the db losses through the filter added to the db losses from the switch in bypass add up to the total loss, when each band is measured separately.
The completed box shows losses on all bands of approximately 1.0 db. This is a bit odd, given the losses of the switch itself vary by band. But the insertion losses of the filters are lowest on 10m, and highest on 80m. Since the losses through the switch component are high on 10m and low on 80m, they all seem to coincidentally hover in the 1.0db range.
It appears that the insertion losses on 40m and 80m filters are a bit higher than the NVARC spec. This is probably because of compromises made in the physical construction, as the coils in those filters are closer to the sides of the enclosure than they should be. They were built last and the enclosure used were more difficult to work with due to their non-standard construction.
The insertion losses in the 1.0 db range are significant enough to be a concern, but everything is a compromise. This is the compromise I’m required to make for SO2R without outlay of more ca$h. The ca$h reserves are currently at less than optimal levels, but there is a lot of that going on. It will be an even larger compromise operating with low power than it would be if the filters were followed by an amplifier, but such is life. Engineer the possible!
In the grand scheme, the finished NVARC box shows about 0.2db more losses than I could expect from the commercial Dunestar series, and maybe 0.4db more loss than is expected from the W3NQN variety. On a positve note, this one cost less than $50.00USD in materials, not to mention everything I learned during the construction. The time required for construction was an educational investment, and was well spent.
Previous in series: Band Pass Filter Fever – The Guinea Pigs – Part 3
Next in series: BPFF – Guess-timating the Filter Efficacy – Part 5.