A lot depends on timing.
...
Where in the 10 years does the air cooled V-12 fall?
1st gen in service by early 1930s, 2nd gen by late 1930s, 3rd gen by end of ww2. WIth 16 and 24 cylinders starting from mid-30s.
quite true. But the 12Y was also a late 20s engine and had never been intended to operate at the pressures and RPM that later engines were. Using early 1930s air cooled cylinders in the basic engine design and then claiming you can use the higher power per cylinder that later models were operated at using 100 octane fuel isn't quite right.
The 1st tests 12Y engines were from 1932.
We can use the cylinders from second half of 1930s on engines that benefit from 100 oct fuel.
Ans what were the changes?
Strengthened crankshaft and it's bearings, increase of finning area of cylinder & head (0.83 sq m on Hornet, for BMW 132A at 1.2 sq m, and for the latest iterations to 1.4 sq m). Fins were thinner, so there was posssible to increse the number of them -> greater area. Use of new lubricant oil, better fuel. CR increased to 6:1. Main weight increase was with versions featuring reduction gear vs. the ones without it.
There was no change in external dimensions of heads.
translation problem? replace might mean instead of, depends on context. But I understand your meaning to be that in the late 20s/early 30s the engine designers go with the V-12s (or X-16s/X24s of H engines) instead of the two row radial. However there is a trap/dead end there.
Yes, here the manufacturers do not make 2-row radials, but air-cooled in-line engines instead.
As the radials were developed for more power they got a lot more finning. Already mentioned the Pegasus, the R-1820 went from from about 600 sq in of fin area on the head alone (not counting the cylinder barrel) of 600 sq in in 1931 to 1000 sq in in 1934. in 1935 they went to just under 2000 sq in and by 1939 they were at about 2300 sq in , including the barrel they were over 2800 sq in. the depth of the cylinder head fins went from about .75in in 1931 to 1.5in in 1935 to 2.25 in 1939.
Thank you.
If you design a V-12 crankcase to hold 1931 cylinders you won't have enough space to fit 1939 cylinders without cropping the fins. If you allow for extra space between the cylinders to allow for future growth you have a longer, heavier crankshaft and crankcase than needed for the power being made in the early 30s and you have made the torsional vibration problem worse. Without the late 1930s cylinders you won't get the power you are looking for no matter what fuel you put in the tank because you can't cool the cylinders without enough fin area.
RR designed, in 15 years, 6 different engine lines (Buzzard, Kestrel, Merlin, Exe, Vulture, Pennine, Eagle 46), plus 4 substantial redesigns of previous engines ('R', Goshawk, Griffon, Peregrine). Plus change in head and block of Merlin. P&W designed 5 different types of multi-row radials, plus 2 off-springgs of the 'main' types. Junkers designed 4 different in-line engines + diesels + jet in 10 years. Bristol designed 4 different 2-row radials in 10 years (that is without the stillborn Hydra and paper-only Orion). Mikulin was not sticking with cylinders from M-17 when going to M-34either ; AM-38 or 42 were far cry from M-34. Even the AM-38 was not just AM-35A with different S/C gearing.
Or, in other words: there is really no need to stick with early 1930s tech in 1943-44-45, even if it makes sense to use it in mid/late 1930s.
In principle yes, in practice no.
In practice there was the R-4360 - granted, not inline, but still rows of tight-packed cylinders each needing the direct air cooling, and RR Pennine of 2700+ HP. That is/was the top-end.
Hornet vs. BMW 132A (note the change of fin density & thickness); cutaway Hornet vs. 132 and overview of the 132 versions: