I initially used a 1:1 balun to feed the loop and it matches quite well.
Now I can switch between the full size dipole for 80 and the full size loop for 80m.
Noise, and signals are lower on the loop. The real test is measuring signal to noise ratio and I've been switching back and forth while receiving WSPR to get an idea on this.
So far I can't see that SNR is better on one or the other. HF conditions vary all the time so perhaps I need to set up two receivers to really compare.
The loop is quite low to the ground, perhaps 2m high for much of it which might explain why it's not superior to the dipole which is quite a bit higher up.
Thanks to Frank, K4FMH, for bringing his excellent presentation on optimising horizontal HF loops to my attention:
Clearly my height above ground is too low for 80m use. The shape of my loop is far from square, more like 5 sides attached to trees.
Here's an SWR plot using nanovna-saver with the 1:1 balun I used at first:
After watching Frank's excellent talk, I changed to a 4:1 balun and things are improved on the higher bands: