That's a good point, but for all the things 'wrong' with the regulator design* - a quick look at the schematic shows it is
not sensitive to the status/quality of that 47uF axial cap.
That cap, one per rail, is used to pre-filter the feed to the zener reference: and even if the axial cap at , say ESR >+10ohms and cooked in old age - ripple at say 0.5Vrms at the reservoir caps only (nominally) rises to c. 100uV on the reg ( - based on simple approximation at 100Hz.)
So - fortunately - it's
much less critical than the axial 10uF parts for sure
*
@PigletsDad has posted detailed analysis on this, in last 3-5yrs. tl;dr: the whole regulator design is 'upside down' and
that gives up 20-30dB of potential performance - but to gain that margin requires homebrew & testing of something
entirely different. You cannot get there by simple parts mod on the original PCB.
So when the factory layout works well-enough to please most people, most of the time - and is as demonstrated above , easy enough to service even by people new to opening-up their power amp - I'd suggest ...don't worry about it. If you love your NAP250 / NAP135 as it is - it is very, very easy for
you to keep things that way.