Hey Chris;
There's really 3 considerations here (arrived at while walking the dog...in the rain...uphill both directions...)
1) The amp. I don't see why it can't be used down here; I'm pretty sure the transformer/inductors will have enough oomph around 500 kHz. The cores are FT50-77, and should be fine down there. It is designed for a balanced center-tapped winding, though.
2) The winding. The amp has a pair of built-in NTE618s in back to back configuration, which give an effective tuning capacitance of around 10 to 250 pF. This was designed primarily for the 2000 to 8000 kHz range. This dictates the required winding inductance which at C = 200 pF, F = 490 kHz needs to be around 530 uH. That is a lot of inductance for an air core loop. To maintain directionality, the total length of the winding conductor should be a small fraction (10-15%) of the operating wavelength. The aspect ratio (the length of a side divided by the depth) needs to be less than 4 or 5, so they say.
3) The system loaded Q. This is required to establish a useful Effective Height (the relationship between the incoming signal strength and the terminal voltage of a parallel tuned tank). The tuning diodes have a specified minimum Q of 100 at 1 MHz; I tend to see better than that on the average. They should be okay, not great, at 500 kHz. The winding will need to have a very good Q - maybe around 500? - to keep the system Q up. It will be big, and maybe not the easiest thing to swing on a rotor.
The loaded Q also needs to be high to help protect against IM products from strong MW stations. The loop needs to act as its own bandpass filter.
I am wondering (and I've thought this in the past) that the best implementation would be ferrite rod loop, at least for a quick and cheap attempt. To make a center-tapped (balanced) 500 uH winding on a rod, using Litz since it would be effective down at these frequencies, shouldn't be that difficult. This rod and the amp should be fairly light, and easy to pigglyback on any rotor you already have up and running. I have been meaning to give this a try, but I just haven't gotten to it yet.
Re directionality: loops will obviously give, at best, a cardioid response pattern.