EDIT: I've re-thought this log, and decided to take a more scientific approach to evaluating ENIG finish in this application. See later logs for details.

So, I'm thinking about the sensor design - I need a small PCB to carry the MLX90393, then another remote board to hold a microcontroller (spoiler alert, I'm thinking PIC here), programmable current source for electromagnet experiments, and USB-UART bridge. I naturally started wondering about the nickel used in the ENIG finish on many boards, and what magnetic properties it might impart. Electroless nickel immersion gold (ENIG) plating uses a thin layer of nickel over the copper traces, and a very thin layer of gold over the nickel. Nickel, like iron, is ferromagnetic (but less so): it's attracted to magnets, and can be made into permanent magnets.

To see if an ENIG finish interacted with magnetic fields, I suspended a PCB with ENIG finish from a thread, and tried to attract it with a strong NdFeB magnet, as shown in this video:

The effect is very weak - you can't pick up the board with a magnet or anything like that - but very real. My concern for the sensor head is that the nickel on the board may become magnetized during the scan, skewing the results, or act as a magnetic flux concentrator, distorting the measured field.

I noticed that the MLX90393 evaluation board uses a gold finish of some type, as does the HMC5883L breakout board from Adafruit, although I'm not sure if either finish is ENIG. Have they done the calculations and/or tests to show that this doesn't affect the measurements? Who knows. I don't think I'm going to take any chances, though - it's HASL for the sensor board.