Fran recommended a test to me that I might get done this weekend. One of the beta testers of the previous attempt had wanted all 3 bars coming out the same side:
This turned out to be a problem. As long as the bars are all like this:
__ __ __
The other bars' magnetic fields won't be able to affect the sensor in this situation. That's because their field will pass straight through the top of the sensor, and the sensor only picks up the fields that pass through the sensor from left to right (or right to left). The old right hand rule!
So, the trick is to put the sensor in a location so that when you run current through a virtual B+ or B- bar (positioned like it would be in the controller), you shouldn't see any change in the voltage output of the current sensor. And modify the position of the sensor until the external affects are minimized.
So my current haha plan is to see if I can get away with not using a shield.