That's indeed interesting, I am sincerely surprised they actually pulled some code out of their database and provided it.
It does mean a 150pF capacitor is actually a worst case scenario to test on a BM235 (but this averaging might be deactivated in capacitance mode). I guessed there would be a switch at about 20LSBs given the 220pF was quick while the 150pF was not, but this is the first time I hear a manufacturer provide such an detailed characterisation of such a behaviour. I applaud them for that!
It does mean the BM235 takes 3,2s to settle after a small change (16 updates @ 0,2s/ud) and will trigger "fast mode" at 80counts/s. The 869s has the same update rate, so will apparently settle in 1,6s in normal mode (don't know if the same figure counts for hi-res mode) and will trigger at 40counts/s. Please note the 869s should be faster in responding to changes because it has higher resolution, so its 40counts/s will be equivalent to 4counts/s for a BM235 in most modes.
Thanks for checking that with the manufacturer, this is very interesting.
PS Averaging does not make the result more precise unless the meter has internal noise/jitter needing averaging, but it does not seem to be any more precise with averaging. It does help agains digit flickering, which is quite desirable and I don't mind the presence of the functionality. I like the values chosen for the 869s better though, and I'd have chosen the values for the 235
lower rather than higher to compensate for the lower resolution. Right now the 235 is powerless agains a multi-turn trimmer.
PS2 [nerd mode] I reread the "code", but it is not completely accurate. It obviously resets the average counter too after renewing the display value, else the value would jump to the correct one, but jump back when the average is back within 16 counts, this does not appear to happen.[/nerd mode]