Adaptive nulling LF antennas?

Anyone have any good ideas/write ups about an adaptive nulling antenna array (2 or more antennas) for the reception of AM broadcast band radio programming. I am trying to receive a distant AM station (1470 KHz) with a strong interferor that is close by, on the same frequency, but luckilly approximately 90 degrees different azimuth from the one I am trying to receive.

You can probably use a loop antenna and put the unwanted one in the null.

Yeah, I am going to try that with a big honkin' loop, and try to rotate it just right. But it seems like if I had say 4 ferrite antennas spaced ~ 50 feet apart, I could let a computer automatically keep a null on the nearby antenna, by adjusting the phase and amlitude of each antenna going into a summing junction. I just am not too sure how to do the algorithms.

It is hard to detect a null, and you only have to be a fraction of a degree off to lose some of the nulling. I was looking for a way to find the null and adjust the coefficients with some algorithm. In that case, I probably would need a digital algorithm to adjust the 4 amplitude and phases.

One idea I was pondering: I know that the Rubidium atomic clock people find a null by looking at the derivative of the transmitted light power as the frequency is dithered.

An alternative might be to use the big single loop antenna, but have a feedback control on the mechanical rotation so that you can find and maintain the null. Here I would only have to detect one null and control one parameter--the rotation angle, probably with some sort of analog control loop.