[The Philpott]

Hmm. Are you working from a known accurate voltage source? I am interested to know whether the correct readings are occurring as you escalate the voltage up the scale, or when you diminish the voltage. The situation i am familiar with is a slightly lazy movement which underreads as you escalate, but shows (better) accuracy as you take the voltage down in stages. This, generally, i attribute to excess mechanical resistance between the jewels and their sprung pivots. I haven't, however, messed with anything later than a Mk III so will leave it to someone else to comment further on this.. and what you might be able to do about it.