Is it really a requirement to watch every single mv? surely not, 5mv steps or even 10 mv steps should be enough.
First, the range we are looking at is microvolt (uV) not millivolt (mV) - but your point is still valid.
The requirement for accuracy is dependent on the circumstance. For most day-to-day situations I encounter at under 12V, 50mV accuracy is quite good enough - but there are some occasions where I found this would be totally inadequate and I needed those smaller units of resolution.
It is to
those situations where this discussion is speaking - and I, for one, am not really happy about the idea that someone has made a decision to include an arbitrary "dead spot" -
especially one that is not advertised.
Call me weird, but I would much rather have the "best guess" of the meter's ADC with a large error bar than not have any idea what's going on. For example, if it gave me a reading of 2uV with an error range of +/- 2uV, then a change from zero would tell me
something is happening. With a 5uV dead spot - you just wouldn't know.
This is the thing about test and measuring equipment - they get used in all sorts of different situations where you cannot always make design decisions in that equipment that affect
fundamental expectations without at least making those very clear from the outset.
It is always useful for someone to know the limitations of their equipment - and it is
essential when they start pushing into those regions. If I had the choice in a situation like this, I would strongly lean towards having a noisy measurement than a 'cleaned up' one. At least with the noisy one, I would see the fluctuations - which would remind me of this limitation or prompt me to check if this was normal. With a "cleaned up" reading - I just wouldn't have any indication ... and I might spend many hours chasing a "problem" in my circuit that simply wasn't there.
I have little doubt that any "dead spot" techniques used in avionics would be specific to the task - and that the implications of doing so would have been well examined before implementing them in respect to anything that leaves the ground.