I have tried an addon beefy 200J 10R NTC to decrease the startup inrush, to prevent my fuse from blowing again. And the behaviour was really not what I would have predicted. I actually have 2 notionally identical 2nd hand compressors - as they're cheap I got one of them as a spare - and to vacuum down during the build. Turns out one has a relay +80uF cap to start it up, the other is simpler with a PTC in series with the startup winding. Anyway, they behave very differently. After a lot of startups and measurements...
PTC compressor startup 16A rms for 230ms, I^2t = 59A^2s
PTC compressor +10R NTC startup 12A rms for 429ms, I^2t = 62A^2s
Relay+80uF cap startup 17A rms for 79ms, I^2t = 23A^2s
Relay+80uF cap +10R NTC startup 11A rms for 19ms, I^2t = 19A^2s
Conclusions:
The version I am using, is the wrong one. I installed the cheapie inferior startup one.
The relay + 80uF cap version is MUCH easier to startup, with 3x lower I2T. This extends the expected life of the fuse type I am using to >1million startups
Adding an NTC to either drops the peak current to below 13Arms, but oddly hardly affects I2T.
There are forms to fill in in the UK if the ASHP/GSHP has a startup of >13A, so swapping over will both save the fuse, and also means I'm vaguely legit:-) I think I can just swapout the elec bit - it's just a plug in object + solder on the cap, no need to change the actual compressor, no brazing needed.
There is a subtlety - I will add a 60kR 5W 500V resistor in parallel with the 80uF startup cap. Otherwise by default it stays charged, and there is nasty DC available down the power cables after power is removed. 60K will dissipate <1W, and should decay to 60V in <10s.
Attached are some pics, the scale is the same every time. One shows the PTC style compressor startup plus a schemtic of the compressor, the other the relay style, both have with and without an addon series 10R NTC.