|
Did some further research, analysis, and testing.
One of the things that surprised me was the breadth of the beamwidth, which on a dish you'd expect to be pretty narrow. At these frequencies and in my circumstances, it was about +/- 10 degrees. Hmmmm, perhaps the antenna elements weren't in the focal point of the dish. Did some physical measurements and calculations, found that my empirical "focus" of the elements was about 3 inches off from the theoretical. Moved to the theoretical, not much help--in fact, spectrum analyzer says it's 5 dB lower. Grrrr. Here's my theory: the parabolic is designed to focus the beam for a C-band antenna...a dipole a couple of centimeters across. For my much larger UHF elements, I have to move them in toward the dish so I get the signal before it's focused too tightly. Following my theory further, that's why the back-reflector didn't help: it was bouncing the signal off at an odd angle that didn't hit my antenna elements, rather than reinforcing the beam (to really work, the reflector would also need to be a small parabolic).
After reading several articles on real-world bow-ties for reception (rather than transmission), I determined that the length of the "whiskers" was too long (16") and could be safely reduced to 12" or even 8" (particularly given the small area of focus for the dish). My first change was to cut the whiskers down to 12" and was surprised that the performance might have actually improved a bit (even down at channel 2, 56 MHz!). Next up: reshape the bow tie to form dual dipoles (phased, stacked vertically). This isn't really optimal for the C-band dish, which has an f/D ratio of about .35 -- but we'll see.
I spent time researching antenna feeds using AARL and other sources, and found that virtually every article emphasized the importance of careful tuning of the antenna elements and feed, with a focus (so to speak) on transmitting. This would make a lot of sense: when transmitting, it's relatively easy to send out a continuous power signal and then have a friend with a receiving antenna more than a mile away to take measurements-- over the course of a phone call, all kinds of quick optimizations could be made and compared against the received signal strength. Once you've optimized the transmit, the receive is already done. But in my situation -- terrestrial DTV, broadband reception, deep fringe -- you don't get the luxury of easily-compared tweaks. As a corollary to this, all the articles are focused on single-frequency optimization (unless, of course, there are two feeds...) -- none of the articles mention anything about bow-tie antennas or broadband reception. So with a DIY dish receiving antenna, YOYO....and it seems that a broadband dish is inherently sub-optimal (because a lot of its gain comes from the tuning).
As I live in a deep-fringe area with essentially no local interference, the spectrum analyzer numbers don't really tell the story on marginal channels (i.e., pixelation)--the signal strengths vary over the duration of the spectrum-analyzer sweep, so they're hard to compare. Further, the issue isn't raw signal strength: it's singal-to-noise at the demodulator that is trying to fish the digital streams out of the signal. If there's any multipath (likely in a refracted signal from way-over-the-horizon source) or competing transmitters on the same wavelength (definitely a problem on one of my frequencies), the signal analyzer may show that I've got 5 or even 10 db of "signal" yet nothing on screen. The final arbiter of "better" has to be the TV itself.
Because the transmitters are waaaaay over the horizon (under, actually...), one effect I hadn't anticipated is spacial diversity. Moving the antenna up or down just 10 feet can make a big difference -- particularly in the early morning and evening when temperature gradients are causing changeable patterns in the propagation. Which height is better depends on temporary conditions: of course, normally higher will be better -- but refraction can be tricky.
So the end of the story is "the wife wins" -- the dish goes, because it offers no advantage in my situation and it has been a butt-ugly novelty.
Sigh.
More updates after I get some more performance data.
Last edited by davodavo; 27-Jun-2020 at 3:40 PM.
|
