I have an Alpha Delta DX-EE antenna in my attic. The antenna is 40 ft long and my attic is about 5ft high by 55 ft long. There's a lot of other stuff up there--a chimney, Romex, Cat 6 cable, RG-6 cable, a TV antenna, junction boxes, etc. It tried to place the antenna so anything metal was at least a foot away. I ran about 60 ft of LMR400 up to the attic (I soldered N connectors on it).
I borrowed a friend's brand new Icom IC-7300. He hadn't tried it before. I connected the chassis ground to the ground prong on an electrical outlet and connected the antenna to the radio with a PL259 to N adapter. We didn't have much success receiving any stations on any bands. I'm in the Pacific Northwest and the band conditions were roughly same as pictured below:
The Icom IC-7300 has a feature where it can make a bar-chart SWR plot. It transmits at 30W when doing so. I tried 10M and a believe a few other bands. No matter which frequencies I picked, all the bars in the chart were always pegged at an SWR of 3:1.
So I recently tested my antenna (and feedline) with a VNA. I only measured reflection on one port, however. These are my results (please ignore the phase plot--I got the feedline length compensation factor wrong):
I've marked the points with the lowest VSWR. The numbers are even better than in the review article I linked to above.
Why are my VSWR results so different between the radio and the VNA? Could a bad adapter cause this? I used a different N adapter for the VNA as opposed to the radio. Could this difference have anything to do with the radio operating at 30W and the VNA at -14 dBm?
Edit - 23-Mar-2019
After reviewing K1LI's answer, here are zoomed in versions of the plot above that show his point. The house is a split-level with two stories at one end of the antenna and one story at the other. On the two-story end, the antenna is about 21 ft AGL.