The question is now:

would, if the FM was set to cut out at 0.1g, the acceleration measured ingame at the same place as in RL corresond to 0.1g?

I do not know but I could well imagine that what is shown as 0.5g ingame as acceleration is something different from what was measured during the RL flight trials and might just not correspond to 0.5g in RL.

So perhaps the 0.5g could well correspond to the RL 0.1g case (tbc by devs). At least at this moment we cannot exclude as we do not know how the ingame acceleration is obtained.

Are we also sure that the ingame acceleration given is the acceleration in the symmetrical plane of the aircraft? or just a lateral acceleration hence including side forces?.

I agree that historically pilots did not just push the stick forward, they would bank and dive away in a diving turn (just like in the movies) or even split S.

However the complaints seem to be that the cutoff occurs even in normal flight.

Given turbulence or a vigorous transition from climb to level flight, it should occur.

A second related document is in the UK National Archives "AVIA 18/1281 Tests of RAE devices for the reduction of "Negative G" engine cutting on merlin engined fighter aircraft". This document details flight test data on 3 devices (Including the Schilling orifice ... though its called the RAE Restrictor .... PC in action back in the 40's).

It compares each of the devices to an unmodified aircraft. In the tables presented the G used to induce cutout are in the order of -0.5G up to -1.5G. Though emphasis of the document is on the time taken to recover from cutout rather than preventing it, despite the document title.

Given the document is not looking at specifically preventing cutout itself but rather minimising the time of the cutout it needs to be put into perspective when using it to decide on initial cutout values. However it is of interest (imo) that reasonable values of Negative G were used (i.e. significantly less than 0G) in all the tests.... i.e. not just smooth nose position changes.

In our discussion here we are only interested in unmodified systems. The jpg below is from the document referring to an unmodified or "Normal Fuel System" aeroplane

http://i40.photobucket.com/albums/e2...eggdevices.jpg

That document clearly states:

Pretty clear cut but I am surprised at the amount of time it took for the Merlin to recover, 6-10 seconds and averaging ~8 seconds.

No wonder the Luftwaffe makes note of the effectiveness of bunting.

Agree 100%

In the sim the engine will frequently cut out when making fine adjustments to trim in level flight -- which seems a tad excessive.

Its worth pointing out that the Shilling modifications (which involved more than just fitting the famous flow constrictor) main effect was to substantially delay the onset of the second stage flooding cut-off. The shilling orifice was a stopgap.

The "Shilling Orifice" did not actually fix the problem, just delayed its onset a few seconds. Sustained inverted flight was still impossible in a Shilling equipped Spitfire, that required a pressure carburetor.

What you need is a G meter. That does seem kind of excessive. I got the game and can check it out.

The chart I made is based on acceleration measurements taken from the game and are effectively a G-meter (accelerometer):
cur_Plane.getParameter(part.ParameterTypes.Z_Overl oad, 0); is acceleration in the fore/aft plane
cur_Plane.getParameter(part.ParameterTypes.Z_Overl oad, 1); is acceleration in the lateral plane
cur_Plane.getParameter(part.ParameterTypes.Z_Overl oad, 2); is acceleration in the vertical plane, renamed G-force on the chart.

At the moment of the 0.5G cut there were very small fore/aft and lateral g-forces of -0.04g and +0.02g respectively (slight speed reduction and slight sideslip). Its close enough to indicate that the cutout occurs well before 0.1G.

Readings were taken every 300 mSecs. Unfortunately we cannot add instruments to the cockpit so we can only draw data from the game parameters.

Ok that already helps a lot. But I still do not know which point the acceleration provided by cur_Plane.getParameter(part.ParameterTypes.Z_Overl oad, 2) corresponds to (I assume z-axis here is body axis not lift-axis which should have to be confirmed by devs too). Does it really correspond to the acceleration measured by the g-meter in the flight tests that stated a 0.1g threshold for the second stage cut-out? It is important to make sure that we do not compare apples with pears.