Comet 4c - Plugin Development #2
22/06/13 12:00
After two months on photographic commitments, I was able to devote a whole week to the Comet.
The main effort now is programming. As I wrote in March, I blew the dust off my C programming text books, and spent the first few weeks of the year getting my head around the X-Plane SDK.
Engine Thrust / Revolutions
The thrust delivered at specific engine revolutions varies considerably between X-Plane 9 and X-Plane 10. The graph (above) compares the two, with XP 9.70 in blue and XP 10.21 in orange.
Why is it important to get this right? Well; it isn’t for take-off, which is at full throttle, but during all other parts of the flight, the revs are usually somewhere between 75-92% of maximum. Climbs, for example, should be at 7,400 RPM (92%). In XP9, those revs would result in 73% of available thrust, while in XP10 they would deliver 84%. Quite a difference — but which is right?
The second graph (below) shows the standard X-Plane responses as dotted lines. The real jet engine performance is shown in red (from the Rolls-Royce Avon 522-524 Maintenance Manual), and is typical of pure turbojets at that time — only delivering serious thrust as the revs wind up.
The green curve is the corrected thrust/RPM line for the Comet 4C in X-Plane. It is taken from actual test results using the latest plugin. Obviously, each X-Plane version requires a different correction formula, and the final result is the same in both XP 9.70 and 10.21. The benefit in using the plugin is that one can dial in the revs specified in the Flight Planning Manual and expect the model to perform accurately.
Two distinct steps in the red curve between 84-86% RPM are caused by air bleeds on the real engine. Initially, I built the steps into the program, but I found that when the cruise RPM was on one of the kinks, a single tap of the F1/F2 keys could cause the thrust to jump from one part of the kink to the next, resulting in some uncomfortable porpoising.
In a real Comet, the throttle levers were not geared linearly, but concentrated a larger portion of their travel to the steepest part of the curve. Although it would be possible to do this in X-Plane, it would require users to re-map hardware throttles and the F1/F2 keystroke shortcuts which would be a nuisance. So (currently) I have ignored the air bleed kinks, and smoothed the results.
Engine Thrust / Altitude
X-Plane simulates the thrust/altitude performance of bypass engines well, but not pure turbojets, which have a narrower, faster stream of propulsive gasses coming out of the exhaust. This gives them good performance at very high speeds, but they are very much less efficient (and much noisier) than modern bypass engines overall.
Again, X-Plane 9 and 10 give different results. The graph (below) shows the results of a full-throttle climb at a fixed IAS of 225 Kts. Using exactly the same engine parameters in Plane Maker, XP10 produces less thrust at low altitude (low Mach).
My first experiment was to raise the Plane Maker setting for Maximum Efficient Inlet Mach in XP10 until the thrust at sea level was the same in both XP9 and XP10. The result was a curve for XP10 that followed XP9 exactly to about 15,000 feet, and delivered slightly more thrust above that, which is definitely a step in the right direction for turbojets.
It’s also a tip for a quick fix of an aircraft designed for XP9 when you want to fly it in XP10.
Actually, the thrust I needed for the Comet’s Avon engines is even greater at high altitude. I confirmed this in two ways:
First, by running the engine specifications through NASA EngineSim 1.7, which gave me the gross thrust (without installed losses). This is shown on the graph below as a dotted red line.
Secondly, by conducting a series of tests to establish the minimum drag of the model, and therefore the minimum thrust, which is the shown by the green line on the graph (below). The real engines could well have produced thrust somewhere between the green and red curves, but because other aspects of performance also fell into line nicely with the green curve, I decided to use this as the target.
The delta for each (the increase in thrust that I had to provide via a plugin) is shown by the solid blue and orange curves on the far left of the graph. Finding a formula for these corrections was straight forward, and now the engines perform identically in XP 9.70 and XP 10.21 (64-bit) on Mac OSX.
The next step will be to see how the supersonic drag performance has changed in X-Plane 10 …
--
GMM-P (22/06/2013)
The main effort now is programming. As I wrote in March, I blew the dust off my C programming text books, and spent the first few weeks of the year getting my head around the X-Plane SDK.
Engine Thrust / Revolutions
The thrust delivered at specific engine revolutions varies considerably between X-Plane 9 and X-Plane 10. The graph (above) compares the two, with XP 9.70 in blue and XP 10.21 in orange.
Why is it important to get this right? Well; it isn’t for take-off, which is at full throttle, but during all other parts of the flight, the revs are usually somewhere between 75-92% of maximum. Climbs, for example, should be at 7,400 RPM (92%). In XP9, those revs would result in 73% of available thrust, while in XP10 they would deliver 84%. Quite a difference — but which is right?
The second graph (below) shows the standard X-Plane responses as dotted lines. The real jet engine performance is shown in red (from the Rolls-Royce Avon 522-524 Maintenance Manual), and is typical of pure turbojets at that time — only delivering serious thrust as the revs wind up.
The green curve is the corrected thrust/RPM line for the Comet 4C in X-Plane. It is taken from actual test results using the latest plugin. Obviously, each X-Plane version requires a different correction formula, and the final result is the same in both XP 9.70 and 10.21. The benefit in using the plugin is that one can dial in the revs specified in the Flight Planning Manual and expect the model to perform accurately.
Two distinct steps in the red curve between 84-86% RPM are caused by air bleeds on the real engine. Initially, I built the steps into the program, but I found that when the cruise RPM was on one of the kinks, a single tap of the F1/F2 keys could cause the thrust to jump from one part of the kink to the next, resulting in some uncomfortable porpoising.
In a real Comet, the throttle levers were not geared linearly, but concentrated a larger portion of their travel to the steepest part of the curve. Although it would be possible to do this in X-Plane, it would require users to re-map hardware throttles and the F1/F2 keystroke shortcuts which would be a nuisance. So (currently) I have ignored the air bleed kinks, and smoothed the results.
Engine Thrust / Altitude
X-Plane simulates the thrust/altitude performance of bypass engines well, but not pure turbojets, which have a narrower, faster stream of propulsive gasses coming out of the exhaust. This gives them good performance at very high speeds, but they are very much less efficient (and much noisier) than modern bypass engines overall.
Again, X-Plane 9 and 10 give different results. The graph (below) shows the results of a full-throttle climb at a fixed IAS of 225 Kts. Using exactly the same engine parameters in Plane Maker, XP10 produces less thrust at low altitude (low Mach).
My first experiment was to raise the Plane Maker setting for Maximum Efficient Inlet Mach in XP10 until the thrust at sea level was the same in both XP9 and XP10. The result was a curve for XP10 that followed XP9 exactly to about 15,000 feet, and delivered slightly more thrust above that, which is definitely a step in the right direction for turbojets.
It’s also a tip for a quick fix of an aircraft designed for XP9 when you want to fly it in XP10.
Actually, the thrust I needed for the Comet’s Avon engines is even greater at high altitude. I confirmed this in two ways:
First, by running the engine specifications through NASA EngineSim 1.7, which gave me the gross thrust (without installed losses). This is shown on the graph below as a dotted red line.
Secondly, by conducting a series of tests to establish the minimum drag of the model, and therefore the minimum thrust, which is the shown by the green line on the graph (below). The real engines could well have produced thrust somewhere between the green and red curves, but because other aspects of performance also fell into line nicely with the green curve, I decided to use this as the target.
The delta for each (the increase in thrust that I had to provide via a plugin) is shown by the solid blue and orange curves on the far left of the graph. Finding a formula for these corrections was straight forward, and now the engines perform identically in XP 9.70 and XP 10.21 (64-bit) on Mac OSX.
The next step will be to see how the supersonic drag performance has changed in X-Plane 10 …
--
GMM-P (22/06/2013)
blog comments powered by Disqus
