OK, I’ve got some results of my testing.
TL;DR: My antenna’s better but let me tell you why 
Let me set the stage first…
- I live in a very strong RF field generated by multiple FM and TV transmitters.
- I live about 27nm due west of Denver International Airport (KDEN) but because it’s dead flat for hundreds of miles to the east of me, and because I’m about 1000ft higher, I can see the main terminal from my roof.
- Directly to the west of me are the Rocky Mountains. If you’re not familiar with the Rockies, they just “stick up” with no preamble. They’re a 5 minute walk from me.
- Both my antenna and the FA antenna are mounted on my north-facing deck, which if not for one tree, would also be able to see KDEN.
- Both antennas are wired as follows:
- 10m RG-174 N-male (antenna end) SMA-male (entrance panel end) They’re the length they’d be if mounted on my roof.
- SMA-female bulkhead connector through the entrance panel.
- 1m RG-316 SMA male-male.
- RadarBox “Green” 1090 FlightStick with built-in filter.
- No other filters (but that changed later)
- Two instances of dump1090-fa running, configured the same, except for the serial number
/usr/bin/dump1090-fa --device RBX10900 --gain 40 --max-range 360 --fix --modeac <networkstuff>
This is the 8-hour plot for the FightAware 1090MHz ADS-B Antenna: (all distances in nautical miles)
Over the past few days, I’d been playing with filter/no-filter and at about 0900 today, I took the filter OUT to get a baseline reading. With no filter, reception was DISMAL and if you look at the Signal Level chart, you’ll see why. The receiver is just overloaded. Even though the FlightSticks are supposed to have filters, no amount of gain adjustment could improve the stats. At about 1140, I added a RadarBox Blue ADS-B filter back in and you can see the difference double filtering made. I tried asking RadarBox support where their filter is in the receiver chain but haven’t heard back. I’d suspect it’s NOT at the front :). No amount of gain tweaking improved the stats better than what’s shown.
Oh, the spikes in tracks are from tis-b rebroadcasts from KDEN.
This is the 8-hour plot for my "semi-spider" (more later):
The scales are the same for the FA antenna charts. There is NO additional filter in this chain. Just the one in the FlightStick. Adding the additional filter didn’t change the results enough for me to notice. As with the FA antenna, I started with the gain at 40 and at about 1300 increased it to 43. The signal levels went up but I don’t think there’s a difference in usable messages. I’d have to leave it there for a while to tell.
Conclusions: My antenna performed “better” IN MY ENVIRONMENT than the FA antenna and demonstrates that while a commercial antenna will beat a custom one 95% of the time, there are just some situations where you have to go custom. In this case, I think the combination of the flatter far field and the sharp bandpass of my antenna offset the high field strength of those FM and TV transmitters that look down on me from Lookout Mountain.
For reference, here's the 8 hour plot from my "production" rooftop setup with an earlier version of my the spider: These plots are NOT in the same scale as those above.
And finally, here's the semi-spider:
I took the suggestion @geckoVN made earlier in this thread and used double length wire for the radials and just wrapped them around the corner screws. Less drilling, fewer screws to fiddle with, performs the same. Just FYI, a simple 4 legged spider does NOT do as well. Still needs the 8 legs.
