It seems to be "accepted fact" since the days of Lindberg that increasing the CR is a useful thing (the price being paid is the need for lead, etc) but I read somewhere that the SFC of an engine is proportional to the CR.
Is that true i.e. if you put say 10:1 pistons instead of 9:1 you would get 10% more power, and the fuel flow would be the same?
Where does the extra power come from? Obviously the MP won't change, and neither will the combustion chamber volume (significantly) so the mass of air getting sucked in at a given RPM will be the same, so for a stochiometric combustion the fuel flow must be the same, so any increase in the power output must be due to the higher CR. There cannot be a need to tweak the fuel servo for a higher fuel flow.
I am informed that a CR of 9.5:1 or lower is OK for 91UL, too.
The other question this touches on is whether the power output of an engine can reduce over time, if there is no increase in friction, no massive compression loss, and no change to the camshaft shape etc. I can't see how it can.
No, that "accepted fact" is incorrect. For an Otto engine, there is an optimum efficiency:
efficiency = 1 - (1 / compression powered by 0.4)
So higher compression does increase the efficiency of the Otto engine but the advantage gets smaller as you increase compression. At some point you have the detonation issue. In your 9:1 vs 10:1 example, the difference in efficiency is 3.5%. How that exactly translates to HP output depends on a lot but I guess it should be roughly linear.
If you increase compression, temperature rises as well and you have more thermal losses. Note that higher compression cylinders mean smaller cylinders (same amount of air/fuel mixture gets compressed more by the same pistons) and the ignition timing might have to change for optimum power output as the flame needs less time to traverse through the now smaller chamber.
Normally aspirated car engines use compression ratios of 10:1 to 12.5:1, racing engine up to 15:1, motorbikes 13:1. Mazda has a new remarkable engine with 14:1 cylinders. That will be very interesting to see how they do in the field.
Obviously the MP won't change, and neither will the combustion chamber volume
Well, the change in combustion chamber volume is exactly what increases the compression :)
Well, the change in combustion chamber volume is exactly what increases the compression
I was referring to the amount of air sucked in, which the the volume of the cylinder at BDC.
To a first degree approximation one would assume the pressure is equalised throughout the inlet system before the inlet valve shuts.
I've just plotted the =1-(1/A2^0.4) equation and get this
7.5 0.553341612 8 0.564724718 8.5 0.575153123 9 0.584756353 9.5 0.593640368 10 0.601892829 10.5 0.609586991 11 0.616784624 11.5 0.623538244
which suggests going up from say 8.5 to 9.5 would give another 3.2%.
Intrestingly the TB21 with its CR of 8.0 should be 1.8% less efficient than the TB20 (8.5). But the difference seems much bigger in reality.
The other question this touches on is whether the power output of an engine can reduce over time, if there is no increase in friction, no massive compression loss, and no change to the camshaft shape etc. I can't see how it can.
That's an often discussed topic and it does sound so plausible and helps the business case of overhaulers as well. I think it is wrong, our aircraft engines make their rated power pretty much always when cylinders fire. There have been cases where the most rotten engines with 20/80 compression were put on test stands and they delivered book performance.
You can easily do the test yourself. Apply brakes, prop full forward, maximum MP and check the RPM. It will be less than the max RPM the governor allows and should always be the same value, usually given in the POH.
Remember Emir's TB20 performance discussion?
Intrestingly the TB21 with its CR of 8.0 should be 1.8% less efficient than the TB20 (8.5). But the difference seems much bigger in reality.
There are other factors like the back-pressure in the exhaust system and the additional resistance in the induction system.
Where does the extra power come from? Obviously the MP won't change, and neither will the combustion chamber volume (significantly) so the mass of air getting sucked in at a given RPM will be the same, so for a stochiometric combustion the fuel flow must be the same, so any increase in the power output must be due to the higher CR.
The easiest way to understand where the extra power comes from is to understand that CR is actually expansion ratio. A 9:1 engine will therefore expand and cool the charge a bit more than an 8:1 engine. The extra power comes from the exhaust having less energy, similar to the effect on efficiency that results from using a turbocharger.
if you put say 10:1 pistons instead of 9:1 you would get 10% more power
I thought that that was the difference between the Lycoming IO-360-B (180bhp rated) and the IO-360-C (200bhp rated). But the CR is 8.5:1 vs. 8.7:1, sounds like it doesn't explain the difference in rated power.
I'm starting to wonder why the IO-360-B is approved for MOGAS while the IO-360-C isn't.
I also wondered how they achieve 300HP+ in the IO540 in say the Extra.
I vaguely recall it is 10.5:1 or something like that but that doesn't explain it.
It is sone with the prop RPM. The power is the cube of the RPM so even a small increase dramatically increases the power.
An engine develops its power from the fuel it burns. As Silvaire explains, higher compressions allow the combustion process to extract more energy from the fuel but not that much more.
Apart from increasing the efficiency of the combustion process which is only possible to certain extent, there are two main ways to burn more fuel:
Both are used in aircraft engines. My turbocharged O-540 produces 235hp because it is flatrated to 31"/2400rpm. The smaller TSIO-520 produces 310hp (with lower compression cylinders) because it is rated to 34"/2700rpm if I recall correctly (Cessna P210).
PS: The difference between 2400rpm and 2700rpm is one tiny adjustable screw at the governor... 
The power is the cube of the RPM ...
I'm afraid you are confusing something here! Power is the product of torque with RPM, so it depends in linear way.
