Fix ADC gain formula

[briangow]

As pointed out by @tcpan in #515 there was a mistake in one of the formula's for calculating ADC gain. The calculation needs to compute the digital range first before mapping the physical range.

[tompollard]

@briangow could you add a test for this?

[briangow]

@tompollard , I've added tests of the gain calculation for when the baseline exceeds the boundary. For each boundary (MAX_I32 and MIN_I32) the test:

• creates a signal which will trigger the boundary overflow
• checks that we enter the logic for adjusting the gain and baseline
• verifies that a round-trip test of writing and reading produces the expected result within some tolerance
• makes sure the ADC gain formula produces the expected result