Don't know for sure if this has any bearing, but the Nikon FE and FG use the historic 60/40 weighting, while the F90X/N90s were set up with 80/20 bias.
Hi, thanks -
I really don't think it would make a difference with the front-lit, uniform scenes I measured (green hedges throughout the frame, blue featureless sky, brick wall, all filling the frame, all uniformly sun-lit from behind).
Also, a second control (a Canon Eos 3 with its own 'flavour' of centre weighted metering) agrees with the F90X and with my Sunny 16 estimates.
I really think my FE and FG are faulty - 3 pages in, and I'm not convinced the 2 stops errors I'm seeing have much to do with subtle centre-weighted pattern weighting differences, or with user error, or with measured 'difficult' light tricking the meters.
I'm more and more suspecting that one should exercise caution when on the market for one of these old cameras. Test the meter if you're planning on relying on it.
I haven't seen any clear proof in support of or against meter cell aging issues in cameras of this vintage.