If modeling flight dynamics in FSX was simply a matter of entering the specs of the real aircraft, I would have been finished years ago. But it’s not, at least not when you are dealing with turbojets, delta wings, and supersonic flight.
The FDE in 4.1b2 was the best so far, but I was not really happy with the afterburner and how the aircraft behaved at high Mach. I have received comments on this in the forum and in emails, and I have to agree that performance on afterburner is not very impressive. Rate of climb is also way off the mark – especially at high speed, indicating that the wave drag table is not properly set.
So, I have been working hard on this all week and am happy to say that the FDE now seems to be “on the numbers” in almost all aspects. Climb rate, sink rate, acceleration, everything seems to more or less match the performance data I have.
The new FDE will be included in the third and final beta which will be released in a couple of days, but if you are impatient you can download the upgraded FDE separately and replace the existing one in 4.1b2. Just put the downloaded .air file in the Draken 4.1b2 aircraft folder, then edit aircraft.cfg to point to the new file:
sim = j35bm_fsx_55
Please note that this FDE will NOT work in any other version – it is specifically paired with the aircraft.cfg file in 4.1b2!
The new stick and throttle are in place on my desk, and I am finally back on track – although with some turbulence (read on).
While fine-tuning the FDE this week I experienced a new problem: when flying in altitude or attitude hold mode, the aircraft would suddenly become unstable and then bank hard to either the left or right. To regain control I had to turn off the autopilot. My first thought was that this was related to a conflict between the wing leveler in the FSX autopilot and the custom autopilot functions in Draken, but I soon found out that the phenomenon also appeared in fully manual flight, i.e. with both the FSX autopilot as well as the custom functions completely disabled.
After some digging on the Internet it dawned on me that it was probably a matter of the new FDE in Draken overreacting to the rather rough “turbulence” simulation in FSX. Stock aircraft seem not to be affected, but high-performance addons (such as Draken) can be much more sensitive. The problem is not autopilot-related, but with wing leveler enabled the effect will be more prominent as the plane may be in a hard roll before you have time to react and override the autopilot.
The solution was to simply disable turbulence completely in FSX and use the weather function in FSUIPC instead, which is a lot better. I sincerely recommend this.
Today and tomorrow working on model optimization, trying to reduce polys and texture load as much as I am able. Also adding a new “options” panel for enabling/disabling some visual elements, sounds and fx. Then on to writing documentation for the latest additions and changes, and finally compiling the installer package for 4.1b2.
Stay tuned /Tom
The new FDE is now “almost finished” – which really means that we are now in the kingdom of Pareto and Murphy. The final stage of any project is inevitably where most of the work is being done, and where everything inevitably goes wrong. That said, I am actually optimistic about releasing 4.1b2 within the next few weeks.
The big difference between the previous (current) version of the Draken FDE and this attempt is that I am now starting with a blank sheet and some good books on aerodynamics, instead of just tweaking a stock FDE until it squeaks. The 4.0 FDE performs OK within, say, 75% of the envelope. The new version will hopefully be able to “touch the edges”.
I started by remodeling the engine, which now produces net dry thrust within 5% of the real thing all the way from the runway to the “coffin corner”. It also has the correct thrust-specific fuel consumption and a more linear throttle response. All based on the sparse documentation from Rolls-Royce, Volvo and Saab that I managed to find.
The next step was to rework the lift and drag parameters to achieve the correct airspeed, angle-of-attack, pitching moment, etc. during climb and cruise. A great help in this was that I finally found some actual airfoil data for Draken. At the moment I am working on stability and control derivatives, which turned out to be the trickiest part. I don’t have the real values, so it is basically a lot of maths, guesswork, and test flights.
Putting the final touches on the derivatives today, then adding the afterburner and starting on the supersonic parts.
What ever happened to the release of 4.1 proper? Or the second beta?
The first beta turned out to be riddled with bugs and performance problems – some discovered by me in the weeks following the upload, some duly reported by users. I planned to put out a second beta that took care of most of these bugs, but the issues with flight performance were very troubling and I decided to postpone the second beta while working on the FDE. There is, after all, no point in uploading another version which still does not fly properly.
This is basically the problem: A lot of the engine parameters were totally wrong in 3.x and 4.0 due to my misunderstanding a crucial part of the SDK documentation. To compensate for the erroneous engine params I had been tweaking the aerodynamics to a point where the curves were completely off the chart. I have spent the last month reworking this, starting from scratch and actually following instructions this time. The result is a completely new FDE which gives a more or less correct engine thrust at all altitudes. The next step is to rework the lift/drag parameters to also get the correct airspeed and angle of attack, which is what I am working on right now.
Apart from this I have added a lot more 3D details in the VC and made some changes to the sometimes quirky radios and nav systems. When I eventually upload the next version I will try to include a detailed change list. No promises as to when this will be though, I really want to get the FDE right first.
Stay tuned, /Tom
One of the more annoying quirks of MSFS is how it (mis)calculates indicated airspeed from true airspeed. Unless you do some tweaking in either the .air file or in Aircraft.cfg, your IAS will be incorrect. The recommended way to deal with this in FSX is by tweaking the airspeed_indicator.n entries in Aircraft.cfg, which will set a scalar value and a base offset value for IAS calculation.
To my horror, I just discovered that in Draken (FS9 and FSX versions) I had set the airspeed_indicator.0 values way wrong, resulting in indicated airspeed in Draken being about 100 kmh too low. This of course negatively affects things like flight plan calculation and formation flying. Furthermore, since I have calibrated the alpha incidence angle (aoa) curves in the FDE to IAS rather than TAS, the whole envelope is slightly off. This is the main reason why I have failed to get realistic performance at high altitude.
Changing airspeed_indicator.0 and cruise_lift_scalar to the following values were successful in my case:
airspeed_indicator.0 = 0.93, 8.0
cruise_lift_scalar = 1.9
If you feel like experimenting, use these values as a starting point. Please note that other parameters (lift and thrust scalars) may also need some tweaking to get realistic performance.
Thanks to Johan at Novelair for reporting the airspeed error.