So with installation and an audio panel, I’m getting on for a £ 20,000 upgrade on a £ 30,000 plane.
I’d never make that sort of decision in my business – as you say it makes no sense! ………..But that is G.A. for you!
There is nothing inherently wrong with that. My above comment was a generic one on GA and investment patterns. Before I got the TB20 I was looking at a PA28-181 (40k) and spending another 40k on it, and ending up with an IFR tourer with a FL140 capability for 80k. I am glad I went for the TB20 instead of course (FL200 capability, 2x the range, etc) but that happened only when I was able to get more money. If you have a plane which is in a basically good condition (as a 40k PA28-181 would have been in 2001) and which does the desired mission profile, then spending money on it is fine. It also avoids buying somebody’s can of worms – almost every used plane you buy is a can of worms and generates a list of issues for the first year or so.
3kg of silica gel is a lot 
I use 0.5kg and have found (by RH measurement) that it lasts about a week. The effectiveness is limited by the continuous ingress of water vapour (unpressurised i.e. non-sealed airframe) to replace the vapour which has been sucked into the silica gel bags.
I find baking them for 1 night at 120C does it, but not just a few hours it seems.
My three aircraft live in a hangar they is better heated than my house and I have an avionic inservicability rate close to zero, this reflects my experience with aircraft that are hangered.
With greater avionic problems, more corrosion, faiding paint and the chance of a "pond life " smell in the aircraft I would suggest that hangarage is cost neutral vs the extra maintanance costs of leaving the aircraft outside.
I can’t prove this but based on recent experience I reckon a lot of these bad display digit segments come back if the plane is left outside in the sun, with no cover on, so the inside gets really warm (but obviously don’t do this in +30C OAT). That seems to dry things out pretty well.
