nairb121

I'm far from an expert, and I don't know what all their sources were, but It doesn't seem that unlikely to me. The flight model adheres closely to the -1 performance data, but there's nothing in there tying it to true AOA. It seems very possible that some assumptions had to be made in development. Indeed, that's my report. But it's important that it be fixed correctly. I'm fairly certain already that there is an AOA indication error, due to the mismatch in reported stall values from what we have in DCS. I think the AOA limit is just a symptom. My theory/conjecture: An incorrect assumption was made on the relation of the True AOA in degrees to the Indicated AOA in units. The flight model was based primarily on available performance data and descriptions in the -1 manual, which describe the aircraft handling purely using Indicated AOA. Of particular note is the Stall AOA. The -1 manual states that stall occurs at 27-28 units (indicated) for our F-5E-3 - or 24 units in "older F-5E". The Taylor/Skow report linked in the AOA bug report, meanwhile, states that stall occurs at "approximately 23 deg angle of attack". The report does not precisely indicate which subvariant it concerns; however, based on the departure descriptions and the age of the data presented, an older variant seems much more likely. Testing in DCS shows that the AOA gauge indicates 27.5 units at a True AOA of only 20 degrees - this is wrong if Taylor/Skow is for F-5E-3 (unlikely), and very wrong if Taylor/Skow is for F-5E (which stalls earlier). Two more data points: takeoff (with nose hike) and landing roll. The video quality leaves much to be desired, but these appear to indicate about 11 units and 6 units respectively. Landing roll should be at 0° AOA; takeoff nose hike adds 3°. I have not yet confirmed these in DCS. https://youtu.be/WPtpS2wf-0Y Plotting these points gives a very rough idea of how "Indicated" and "True" AOA relate: How this connects to max AOA: visually extrapolating (and correcting a little for the noticeable, but unsurprising, nonlinearity) places 30 units indicated close to 30° true AOA. This aligns closely in my view with language in the -1 which indicates various departure onsets in this AOA region, and a small ability to pitch beyond stall. Simply increasing pitch authority to fit the expected 30° AOA would result in the ability to pull 50% beyond stall - this is not consistent with the language in the manual or the expected behavior of this type of aircraft. It's my belief that the entire pitch axis behavior is scaled improperly, and not just the peak AOA, but also the lift and drag polars need to be reevaluated. Another possibly related problem: the wing-snap issue. If the pitch rate is right, but the stall AOA is wrong - then you will reach max lift more quickly, the g-onset is too fast, and the wings are rapidly overloaded. One more ongoing concern for me is the departure resistance - while our IHQ F-5E-3 should be much improved over the older F-5E, few if any of the departures detailed in the -1 for the E-3 are present in DCS. I'm not sure if this connects to the AOA issue at all, but I thought I'd mention it while we're on the subject. Please let me know your thoughts - maybe I'm overthinking this, or have missed something critical, but that summarizes my thinking on this matter.

You're correct, knowing the "true AOA" is not relevant to typical piloting situations - all operational numbers are based on the gauge units. But the true AOA is relevant from a flight modeling standpoint, and as you know, there is fairly substantial evidence that the DCS F-5 has something amiss with its AOA modeling. I suspect that the root of this issue is an error in the correlation between indicated and true AOA - but I have no data apart from a single point, the stall AOA as mentioned above.

Both, really - I've been unable to find much to correlate the true AOA and indicated units, only scattered references to stall AOA - stated in one report to be 23°, and in the -1 to be 27-28 units - Notably, this does not seem to match DCS, which places 27.5 units AOA at about 20°. This doc seems promising though, I hope we can see it in full before too long.

@Curly I wanted to ask in a new topic to leave your aileron bug report relatively clean - in the Technical Description documents you have, is there any discussion of the relationship between true AOA in degrees and the displayed AOA units? I suspect there is a mismatch in our DCS F-5E, but have been unable to find much useful. See post here for previous discussion:

Wingtip tanks are a feature of the earlier A and B models - the E does not carry them.

Agreed, very good research and suggestion. Have you measured stick vs. aileron deflection in DCS to determine whether any such curve or gradient is implemented? Also, do you have a source for that document? Not that I am doubting its authenticity or applicability - I just want it for myself!

The difference is purely visual.

Currently rocket pod ripple-fire settings can only be changed while the aircraft is completely shut down. However, since the intervalometer is set by ground crew on the pod itself, it should be possible to change it anytime ground crew and loadout modification are available (i.e. stationary at a friendly airfield), even with engines running.

Is it possible that there's a calibration issue, and DCS thinks your throttle setting is slightly higher than 0%? That might prevent it from accepting the cutoff bind. Maybe set a small deadzone (with throttle axis bind set to slider) and see if that helps?

Can you replicate this and post the video so we can see it from the affected plane's POV? There could be clues in the engine indicators. Nozzle should be controlled only by throttle position and exhaust gas temperature - unless the EGT is unstable at that speed for some reason (which I wouldn't rule out), the nozzle in theory shouldn't be moving. Observing the nozzle position indicator would confirm though. At the AOA shown in the video (about 5 units), auto flaps should always be up. There's also no indication of a flap shift (barber pole) by the flaps indicator, and visually they also appear to be up on the other planes in the formation.

I agree that the work we've seen recently has been refreshing - I'm still a bit worried though, the recent fixes so far are mostly low-hanging fruit (the cannon example being a simple lua change), while some of the bigger issues (e.g. AOA and wingtip asymmetry) haven't been addressed.

FWIW, I've never had much luck with the -34 bombing tables either, even being near perfectly on the numbers. I ended up just coming up with my own numbers instead. Granted, this might be total user error on my part. I can't imagine why the tables would be any different, physics is physics... unless the drag values, or reticle depression are wrong.

I believe Wags was stating their intent, while noting that certain aspects were missing. Stills below from "Tigris Helveticus" (Youtube link) on the original Swiss F-5 acquisition. They definitely had INS, and what appear to be dual radios (though not the modern digital interface). Edit: to be clear though, our F-5 is from the second batch in 1981. (Serial 81-0844/J-3085 from the cockpit)

Track attached - just a bit of pure guns-only PvP airquake. There is a definite difference - not as dramatic as I was thinking it'd be, but a good shot is definitely likely to land more hits. F-5 guns update test.trk

I've not tried it myself yet, but I have been told that it's much improved. Hopefully I'll get to see for myself tonight (and find out whether it'll make me hit more, or less!)