When calibrating the performance of the Comet 4C model, everything starts with a stall test. Stalls in clean configuration allow me to adjust the aerofoils for the right amount of lift. Once that's set, I can move on to other aspects of performance.

Stall tests by de Havilland were carried out at 10,000 feet, with throttles closed. The stall was defined as the moment the aircraft could not maintain 1.00g normal to the flight path. This avoided the aircraft actually dropping a wing, or otherwise over-stressing the airframe. Instantaneous readings were taken using a Royal Aircraft Establishment (Smith Barnes) accelerometer, which had an accuracy of 1%.

For the X-Plane stall-test plug-in, I used a PID controller to maintain a constant vertical path of zero (level flight). As the Comet model approached the stall, the rate of change of lift was too quick for the PID controller to keep up with. The controller was tuned to respond as rapidly as possible, without surging or noise, but the flight path still curved down prematurely. I had to find some way to allow for this.

The test was also set up to record the angle of attack, which I could compare with the maximum angle of lift for the aerofoils: Alpha(MAX). Consistently, this happened when the vertical force = 0.99 g, so I have used this as my offset in the table of results, below. An accuracy of 1%, repeatable time after time, is an awful lot better than I was achieving manually!


Fig.1: Lift and vertical acceleration against indicated airspeed during a stall.
The green bands are the figures for the real Comet 4C, as targets with a 1% tolerance.


Fig.2: Vertical flight path, angle of attack and elevator angle during a stall. Maximum lift is achieved when the yellow curve intersects the dotted green line at "Alpha(MAX)" (12.5 degrees) as shown. The elevators reach maximum travel at -23 degrees, shown by the blue line.

Aerofoils	Weight	Flap	U/C	Elevator	Alpha	Target	Actual	Accuracy
Mod	Lbs	Degrees	Up/Down	Degrees	Degrees	Speed (kt.)	Speed (kt.)
127	100,000	0	UP	17.2	12.3	103.0	103.5	0.5%
127	100,000	20	UP	22.9	11.2	98.0	98.9	0.9%
127	100,000	40	DOWN	20.8	9.8	93.0	95.0	2.2%
127	120,000	0	UP	16.5	12.3	113.0	113.0	0.0%
127	120,000	20	UP	21.8	11.2	107.0	107.8	0.7%
127	120,000	40	DOWN	18.9	9.7	102.0	103.6	1.6%
127	140,000	0	UP	16.0	12.4	123.0	121.5	-1.2%
127	140,000	20	UP	21.2	11.3	116.0	116.0	0.0%
127	140,000	40	DOWN	18.7	10.1	110.0	111.0	0.9%
127	160,000	0	UP	15.7	12.4	133.0	129.6	-2.6%
127	160,000	20	UP	20.9	11.4	125.0	123.6	-1.1%
127	160,000	40	DOWN	18.2	10.2	118.0	118.3	0.3%

Fig.3: Table of stall results for a range of weights (updated for X-Plane 10.45 on 30/06/2016)

Aerofoil settings are a compromise between low and high weights. These figures were achieved using Javafoil to calculate lift for a NACA 63A011 + 8.0%, which is realistic for a wing with modest leading edge droop.

Next step: Verify that there is sufficient lift using a take-off test plug-in.

--
GMM-P
(14-03-2016)

Tags: Performance, Aerofoils, Plugin, Programming, PID controller, Testing, Wings, Flight Model