Fig.1: Approach function in progress; tests carried out over the English Channel, West of Weymouth.

Approach tests were another use of the flight test plug-in. In this case, an approach function plonked the aircraft at a specific point in space over Weymouth, and either continued in level flight, going through various permutations of weight, flaps and undercarriage, or descended with a gradient of -3.0 degrees (to simulate a standard glide slope). X-Plane's built-in autopilot was not helpful for this, having only a pitch or vertical speed lock. Glide slope lock was no use except on an actual glide slope, but it was easier for me to write a function that took complete and direct control of the elevators, with a PID to control the gradient. Tests were carried out over sea, so no hillock or other unexpected land mass ended the test prematurely!

Having established lift (including individual flap positions), and the drag in clean condition, the primary variable during approach tests was flap drag. It's also the only test where a comparison can be made between undercarriage up and down, so undercarriage drag was tuned, too.

The de Havilland manual quoted the approximate engine RPM to maintain a specific airspeed during each phase of approach, with a tolerance for different weights. Real pilots would just have flown the plane, but the figures are extremely useful for me, and also for new X-Plane users getting a feel for what level of thrust to use.

I chose a landing weight of 105,000 lb., being the "normal" landing weight for a Comet 4C with a full complement of passengers and luggage, and still carrying the appropriate reserve and emergency fuel allowance. It also happens to be the median weight between the lightest landing weight quoted in de Havilland documentation of 90,000 lb. and the maximum landing weight of 120,000 lb.

The real manual gave a tolerance of ± 150 RPM. In X-Plane it was ± 180 RPM.

This is how the model performs in X-Plane:

Weight (lb.)	Airspeed (kt.)	Flaps	U/C	Target RPM	Actual RPM	AoA (deg.)	Gradient (deg.)	Var. (%)	Ave. (%)
120,000	150	20	0	6,500	6,653	2.8	0.0	2.4	1.6
120,000	150	20	0	6,250	6,336	2.8	-3.0	1.4
120,000	140	20	1	6,750	6,843	3.9	0.0	1.4
120,000	140	20	1	6,500	6,589	3.9	-3.0	1.4
120,000	135	40	1	6,750	6,855	1.8	0.0	1.6
120,000	135	40	1	6,500	6,617	1.8	-3.0	1.8
120,000	124	60	1	6,750	6,858	2.6	0.0	1.6
120,000	124	60	1	6,500	6,587	2.6	-3.0	1.3
110,000	150	20	0	6,500	6,585	2.3	0.0	1.3	0.6
110,000	150	20	0	6,250	6,223	2.3	-3.0	-0.4
110,000	140	20	1	6,750	6,797	3.2	0.0	0.7
110,000	140	20	1	6,500	6,503	3.2	-3.0	0.0
110,000	130	40	1	6,750	6,801	1.7	0.0	0.8
110,000	130	40	1	6,500	6,514	1.7	-3.0	0.2
110,000	120	60	1	6,750	6,802	2.4	0.0	0.8
110,000	120	60	1	6,500	6,587	2.4	-3.0	1.3
105,000	150	20	0	6,500	6,553	2.0	0.0	0.8	-0.1
105,000	150	20	0	6,250	6,152	2.0	-3.0	-1.6
105,000	140	20	1	6,750	6,776	2.9	0.0	0.4
105,000	140	20	1	6,500	6,439	2.9	-3.0	-0.9
105,000	128	40	1	6,750	6,773	1.6	0.0	0.3
105,000	128	40	1	6,500	6,520	1.6	-3.0	0.3
105,000	118	60	1	6,750	6,773	2.3	0.0	0.3
105,000	118	60	1	6,500	6,476	2.3	-3.0	-0.4
100,000	150	20	0	6,500	6,517	1.7	0.0	0.3	-0.4
100,000	150	20	0	6,250	6,169	1.7	-3.0	-1.3
100,000	140	20	1	6,750	6,750	2.6	0.0	0.0
100,000	140	20	1	6,500	6,452	2.6	-3.0	-0.7
100,000	126	40	1	6,750	6,736	1.5	0.0	-0.2
100,000	126	40	1	6,500	6,434	1.5	-3.0	-1.0
100,000	116	60	1	6,750	6,735	2.2	0.0	-0.2
100,000	116	60	1	6,500	6,502	2.2	-3.0	0.0
90,000	150	20	0	6,500	6,442	1.1	0.0	-0.9	-2.0
90,000	150	20	0	6,250	6,040	1.1	-3.0	-3.4
90,000	140	20	1	6,750	6,683	1.9	0.0	-1.0
90,000	140	20	1	6,500	6,339	1.9	-3.0	-2.5
90,000	121	40	1	6,750	6,643	1.3	0.0	-1.6
90,000	121	40	1	6,500	6,342	1.3	-3.0	-2.4
90,000	111	60	1	6,750	6,638	2.0	0.0	-1.7
90,000	111	60	1	6,500	6,326	2.0	-3.0	-2.7

As you can see, it's been a productive week!

Next step: moving on to other bugs on the beta-testers' list.

--
GMM-P
(12-08-2016)

Tags: Performance, Plugin, Programming, Testing, Flight Model