I’ve been wondering that, since I’ve been using it to null out noise sources, yet I had long forgotten what the pattern of such antennas is.
I couldn’t just use the antenna itself plus noise to find out because I live in town and there’s potential noise sources all around. I couldn’t use the antenna plus signals because the ionosphere compromises the source-directionality of signals that pass through it, and it’s an HF-only antenna. (I suppose I could have spent most of the day delving into the physics of it, and coming up with the answer, but that would have taken most of the day and I just wanted a quick answer.)
Checking on-line wasn’t very helpful. I found articles claiming that both the null and the lobes were “in the plane of the loop”! Part of the issue, I think, is that there are different geometries of loop antennas. What one would tend to think of as the “plane” of a large, flat, air-core loop is perpendicular to what one would tend to think for a compact, multi-turn ferrite loop stick antenna.
A simple experiment with one of my old tube radios (which have a large, flat air-core loop antenna for the medium-wave broadcast band) and KVI’s signal on 570 KHz povided the answer: for “flat” loop antennas with a large diameter and low number of turns, the null is along the plane the loop exists in. It was pretty definitive: when I got the radio aligned so I had difficulty hearing KVI, its back (on which the loop is mounted) was aligned directly on a line running from KVI’s Vashon Island transmitter to me.
So there you have it.
Note that the while the peaks of a loop antenna are very broad, the nulls are by contrast very sharp. I’ve noticed this when nulling out interference; a slight bump on the HF loop antenna (changing its position by mere inches) often makes a significant difference.