You're right Ed, many things to try and if you have a decent parts trove, you can do lots of rx experimenting with little cash outlay.It surprised me what a small amount of loop inductance can generate. With my experimentation, I now think the best approach is to decide the lower cutoff freq of the loop, then adjust the secondary turns/inductance needed for the standard 4-5 times the impedance of your coax at that freq. I would add primary turns just enough to have the desired frequency spread, lets say 2200m to 10m, raise the rx noise level to somewhat above what is the resting noise floor of the rx without an antenna connected. Then, a preamp wouldn't be needed. Of course, with too small a loop that may never happen, so a preamp would be appropriate. I could be mistaken, but that's my "position" at this point of time. I'm not an expert. :-)
Those are a couple good tips you made concerning checking for "in house" noise.
I tried a loop on the ground for HF, but didn't really give it a thorough test. I may try it again at some point of time. Understood about your longwire not working well at 22m. Yes, things are a compromise when a large bandwidth is desired. I'm guessing distance of the station to be received might have some bearing on what polarization is best at that given time/distance. Lots of variables we experimenters have to "play" with.