I have to deal with this same problem, but even at a smaller scale than what you're describing. I use two different power supplies. If you must use just one AC/DC conveter, I would recommend using two isolated DC/DC converters which are easily available. Filters will only help so much, but the best I have been able to do is by using two separate isolated supplies. If you do need a reference ground from the motor supply (as in my case), I would also recommend using a very short thin trace that connects the ground from one to the other. The other issue is proximity. If you are sure proximity is not causing the deficit then I would try the above. I have even tried replacing the ball bearings in the motor with ceramic ball bearings and making sure that no part of the motor makes electrical contact with any other portion of the chasis. Basically, isolating the motor from the acquisition sensor, altogether. I have found that just isolating the body of the motor works quite well and that over time (as the bearings set in, the grease dries up) that noise increases on my signal. The most bang for the buck is definitely two isolated supplies. Also, isolate the sensor from the chasis. I think that isolating the sensor may prove easier. These methods work for about 30db of isolation, after that averaging needs to be employed.