...
You seem to think manufacturers never make mistakes.
FYI I just measured a collection of random diodes I manged to arrange to get a reading of 3.053V on the F87. I measured that voltage drop in parallel (quicker than making a whole new setup) with a Keithley 2015. It said 3.05242V. I don't even need to bother doing the maths, that's clearly well within the 2%+1 LSD spec, and well within the combined uncertainty of the meters.
I put it to you that the 3.000 figure in the manual is a typo, and should read 3.099V. Or perhaps they just felt 3.000V looked neater than 3.099V, such things have been known.
If they haven't bothered to correct the Hi Res firmware issue (afaik), why would they bother with a trivial typo?
EDIT: Incidentally, while the manual gives a figure of 0.6mA (typ) for diode test current, I measured mine as 0.78mA and change. That's 30% away from "typical". Make of that what you will.
Again, can you
GUARANTEE that the readings what you've just done, stays within the specified limits for
ALL specified conditions for
ALL the manufactured DMM's(calibrated, OC)?
Because that's what the specification changes means.
And BTW, what is the required accuracy for "typical"? Old Trafford or playing field?
Anyway, I'm getting tired of this discussion, never mind you've barking at the wrong tree; I strongly recommend to contact Fluke, and continue the discussion with them, since they are responsible to the issues you've raised.