Fighter aircraft sea level climb and turn rates evaluations using "Tacview" versus their credible manual's data.

[MaverickSu-35S]

Hello,

 

           I am not here for criticism or to give a hard time to the wonderful people (flight model developers/engineers, etc.) who have put all this together in an effort to share a unique simulation of WW2 and WW1 aircraft flying behavior and experience, but to share them an insight of what I've found during some tests of mine, on what the fighter aircraft (Axis and Allies) performances actually are compared to the given specifications at their description. Some validate spot on, while some are quite far off...! Nothing is perfect and everything comes at a price and I know that pretty well when you have lack of available data and when it's very difficult to get something reliable from CFD analysis alone! Nobody is seeking perfection, but an error of less than 5..3% should be reasonable with a correct CFD only analysis! Right now, we have plenty of aircraft for which their performances error vs. reality (at least what the description tells, which I believe it's from their real flight manual) sometimes peaks over 30%, which is quite too much for a realistic performance.

 

I have done these tests only for WW2 fighter aircraft, not for bombers nor for any WW1 aircraft (at least not yet) and the results speak for themselves. I've used Tacview to analyze the tests recordings and try to see what respects the descriptions and what doesn't respect the descriptions and how well the results validate from the manual or how far off they can be.

 

As far as I can see from the tests, the "maximum performance turn at sea level" told by the description is the necessary time for the aircraft to complete a 360 degree turn while already flying at a constant turn rate, NOT when starting a turn from a high speed and a reference altitude and heading, then count the time it took to finish the turn while holding critical AoA. For the tests I have meticulously read and respected the prescribed conditions (weight and engine regimes) in order to not misinterpret the results. The "standard weight" is usually (except for different given fuel % conditions) the weight of the plane having default and unmodified 100% ammo with 100% fuel. For example, the YAK-7B has a given -no ammo- and -10%- fuel weight of 2630kg. The maximum fuel weight is 305kg. Thus 2630 + 0.9 * 305 = 2904 (no ammo, 100% fuel). The standard weight is 3002kg. It makes sense that the rest of 97.5kg is the 100% ammo weight. Thus, in most of the times, the "standard weight" is simply the unmodded plane's ammo with 100% load and 100% fuel. Please correct me if I'm wrong at interpreting the given "standard weight". Before starting the turn tests, I've carefully set the engine conditions as described at the specs and I've only took the average degrees per second (TRT in Tacview) rate from the final 5-10 seconds of the constant turns where I have settled the aircraft to fly with as constant variables as possible (IAS, AoA and vertical velocity), then divided 360 to that average constant turn rate and obtained the time in seconds for a constant 360 turn.

 

Here is a .rar file which contains list of Tacview files recordings (in the parentheses we have the climb rate and the 360 degrees constant turn times given by the manual and after the dash we have what actually takes place within the simulation).

 

I'm here for decent discussions and looking forward to receive answers in a similar way.

 

Kind regards,

Mav.

IL-2 BoX FM problems.rar

Edited July 6, 2021 by MaverickSu-35S

[JtD]

Good approach. I also take the standard weight to be the, well, standard weight.

 

However, trying to use TacView for turn rate testing myself, I found that turn rate in a constant turn is not constant in TacView. I didn't get to the bottom of it, but I think it is realated to the transformation of geodetic coordinates. TacView appears to use only these, and gets fairly inaccurate in the process.

 

The best way to measure sustained turning imho still is to do a dozen or so turns with (near) constant altitude and speed, and then check on a watch how long it takes.

[LColony_Kong]

I dont think the turn rate inconsistency in tacview is related to how tacview relates coordinates. It has to do with tiny pilot error. Ive measured the same turn time test on different airplanes with both taview and by recording the flight and timing it independently, and got the same results. After all, this is the reason a person needs a dozen or so turns to get a decent measure in either case. There are simply tons of tiny pilot errors with slightly oscillating speeds, bank angles, and angles of attack. It even varies between different aircraft, and in the same aircraft as weight and CG change depending on fuel or gun load etc. Some planes or configurations are simply easier to fly and result in far less variation in measured rate over time. You can also see large changes in measured rate in some planes if you trim the plane to counter torque, since the the torquing opposite to direction of turn makes for particularly fine control requirements on the pilot.

[JtD]

The errors I was looking into weren't random, they were systematic and related to the heading of the aircraft. I doubt that my piloting errors or other random errors constently caused turn rate to go up through the northeast and southwest quarters, and down through the other two (or whichever way round it was). Might be map related or even position on map related.

 

Fact is that our maps are too large to accurately work with exactly orthogonal coordinates, which I think they are using. If that's true, and we assume that in the center of the map north is actual north, you may get two degrees of error at the edges. Meaning that while the game says you're turning 90° at a constant rate, you're actually turning 92° in geodetic coordinates through say north east, and then 88° through say south east. This results in a 5% difference in turn rate, which I think is what TacView is doing.

 

Anyway, all assumptions, but would explain my issue to me, so I left it at that.

[MaverickSu-35S]

On 7/8/2021 at 7:39 AM, JtD said:

Good approach. I also take the standard weight to be the, well, standard weight.

 

However, trying to use TacView for turn rate testing myself, I found that turn rate in a constant turn is not constant in TacView. I didn't get to the bottom of it, but I think it is realated to the transformation of geodetic coordinates. TacView appears to use only these, and gets fairly inaccurate in the process.

 

The best way to measure sustained turning imho still is to do a dozen or so turns with (near) constant altitude and speed, and then check on a watch how long it takes.

Thanks for the reply!

 

Guess what, I've already done that initially, testing a plane at constant AoA/IAS at sea level (about 10-20m AGL) using just chrono to see how many seconds it takes for a precise 360 deg turn and with an error of about 0.1..0.3 seconds between the tests (which I find accurate enough anyway for just a chrono testing), the results were almost identical with Tacview as well. That's the reason why I later used just Tacview as I saw a very good cosistency of the calculated degrees per second turn rate and the chrono measured tests. So it's good and reliable, no matter what transformations it uses to get your TRT (turn rate dps).

Of course that the TRT is most of time wobbling a bit, but no more than around 0.1..0.2 dps if you maintain a perfectly constant turn during the recording. I have remembered a portion where during the recording I stood with the stick and rudder fixed in a position and on the same place in the Tacview analysis, indeed I had a perfectly constant TRT. So it matches and it's reliable. The TRT didn't change with even 0.1 dps, which confirmed that it was quite precisely calculating the turn rates from what I actually did in the sim.

 

Now, even if let's say..., Tacview would've messed the turn rates by some 1 degree per second (let's say...), the big differences between what turn rates (or resulted 360 degrees constant turn rate time) should a plane have according to it's specifications and what actually it performs during tests by carefully respecting the told conditions, still proves that there's some work to be done!

 

From my point of view, I'm not looking to sound smart or to "teach" the FM devs how to do their job, the corrections require readjustments of the 2 most important factors for flight performances, and only readjust the lift slope (CL vs AoA) if the stall speeds for given weights and flaps/slats conditions are also not correct (if the weights are already correctly implemented in the FM): CD vs AoA and propeller (or jet for ME-262) thrust with IAS/Mach. Too much turn rate and climb rate while the top speed is good? Increase only the drag vs AoA slopes above 4..5AoA, but keeping the low AoA ranges (up to 3..4 AoA) constant to not affect top speed at low AoAs, thus both the climb and turn rates decrease as the lift to drag ratio decreases for the used AoA without touching the prop. Too much climb rate, but worse turn rate, as it happens with LA-5FN? If the lift slope and CL max are already correct (stall speeds, etc.), work from both prop thrust and drag tables by increasing the drag power for the whole CD vs AoA function altogether and adjust the prop thrust vs speed to not affect the top speed if it's already where it should. I'm only hoping to give some hints of the stuff that I work with for X-Plane in improving flight models. Again, don't take me like I wanna look/sound smart, I'm only hoping to help IL-2 FM devs reduce the errors with reduced effort as well!

 

Regards!

On 7/8/2021 at 12:18 PM, JtD said:

The errors I was looking into weren't random, they were systematic and related to the heading of the aircraft. I doubt that my piloting errors or other random errors constently caused turn rate to go up through the northeast and southwest quarters, and down through the other two (or whichever way round it was). Might be map related or even position on map related.

 

Fact is that our maps are too large to accurately work with exactly orthogonal coordinates, which I think they are using. If that's true, and we assume that in the center of the map north is actual north, you may get two degrees of error at the edges. Meaning that while the game says you're turning 90° at a constant rate, you're actually turning 92° in geodetic coordinates through say north east, and then 88° through say south east. This results in a 5% difference in turn rate, which I think is what TacView is doing.

 

Anyway, all assumptions, but would explain my issue to me, so I left it at that.

 

Ok, let's assume that, but still, the error is much higher than that. So it's this problem with Tacview (if indeed confirmed, cause from my tests it was validated with in-game chrono tests) regarding the shown TRT (turn rate), what about the climb rates? Both what I could see on the plane's vertical speed as well as in Tacview were a match, so it makes a lot of sense that if a plane is overrated in terms of climb rate, it's a general logic that the turn rate also becomes overrated (greater than real thrust vs drag = greater turn rate). Cheers!

 

---------------------

 

-The following is a separate message, but it's forcibly merged and I can't do anything about it:-

Because those Tacview files that contain the planes performances specs and results are within the rar file uploaded here, this is the full list:

 

- BF-109E7 (14mps - 20.5s) - climb rate matches, but with 18s (overrated) for turn time;

- BF-109F2 (16.4mps - 23.6s) - does 18mps (overrated) with just under 21s (overrated) turn time;

- BF-109F4 (19.5mps - 20.3s) - climb rate matches, but with 18s (overrated) for turn time;

- BF-109G2 (21mps - 22.2s) - does 20mps (underrated) with 20s (overrated) turn time;

- BF-109G4 (20.1mps - 21.2s) - climb rate matches, but with 20s (overreated) for turn time;

- BF-109G6 (20.6mps - 21.7s) - the climb rates and turn times are a perfect match;

- BF-109G14 (19.2mps - 23s) - the climb rates and turn times are a perfect match;

- BF-109K4 (20.5mps - 24s) - does 18.5mps (underated) with 23.5s (match) turn time;

- FW-190A3 (16mps - 23s) - does 17.5mps (overrated) with 20s (overrated) turn time;

- FW-190A5 (15.4mps - 23.5s) - climb rate matches, but with 20.3s (overrated) for turn time;

- FW-190A6 (15mps - 23.5s) - climb rate matches, but with 21.1s (overrated) for turn time;

- FW-190A8 (13.8mps - 24.2s) - the climb rates and turn times are a perfect match;

- FW-190D9 (19mps - 20s) - does 18mps (underrated) with 21.2s (underrated) turn time;

- I-16 (16.7mps - 19s) - climb rate matches, but with 20.2s (underrated) turn time;

- LA-5 (18mps  - 23.4s) - does 15.5mps (underrated) with 28.35s (underrated) turn time;

- LA-5FN (20mps - 21s) - does 23.4mps (overrated) with 23s (underrated) turn time;

- LaGG ser.29 (14.9mps - 22.2s) - the climb rates and turn times are a perfect match;

- MC-202 (17.3mps - 22.6s) - does 14.9mps (underrated) with 22.2s (match) turn time;

- Me-262A (19.3 - 33.5s) - does 15.2mps (underrated) with 28.3s (overrated) turn time;

- MIG-3 (15.9mps - 22.4s) - the climb rates and turn times are a perfect match;

- P-38 (20.4mps - 20s) - the climb rate matches, but with 21.1s (underrated) turn time;

- P-39 (16.7mps - 21.5s) - the climb rate matches, but with 20.2s (overrated) turn time;

- P-40 (12.5mps - 24.3s) - does 11.5mps (underrated) with 21s (overrated) turn time;

- P-47D22 (18.5mps - 27s) - does 16.7mps (underrated) with 25.5s (overrated) turn time;

- P-47D28 (18.1mps - 27.5s) - does 15.3mps (underrated) with 26.5s (overrated) turn time;

- P-51D15 (18.1mps - 20s) - the climb rates and turn times are a perfect match;

- Spitfire MKIXe (21.5mps - 17.7s) - does 16.7mps (underrated) with 17.5s (match) turn time;

- Spitfire MKVB (12.9mps - 25s) - does 14.2mps (overrated) with 19.2s (overrated) turn time;

- Spitfire MKXIV(23.8mps - 18.1s) - does 21.8mps (underrated) with 18.4s (match) turn time;

- Tempest (21mps - 20s) - does 17.8mps (underrated) with 19.5s (match) turn time;

- Typhoon (+11 boost, 4 blades) - does 19.8mps (underrated) with 18.5s (certainly overrated) turn time;

- Typhoon (19.3mps - 23.4s) - does 16.7 (underrated) with 20s (overrated) turn time;

- YAK-1 (16.9mps - 19.2s) - does 16mps (underrated) with 19s (match) turn time;

- YAK-7 (16.9mps - 19.5s) - does 15mps (underrated) with 20.3s (underrated) turn time;

- YAK-9 (18.5mps - 17.5s) - does 14.7mps (underrated) with 20s (underrated) turn time;

- YAK-9T (16.5mps - 19s) - does 15.2mps (underrated) with 20.5s (underrated) turn time.

 

P.S.:
What is wrong with this forum system? Why can't I have 2 separate massages if I want to? The system merges them forcibly! Thank you!

Edited July 9, 2021 by MaverickSu-35S

[MaverickSu-35S]

On 7/8/2021 at 8:50 AM, LColony_Red_Comet said:

I dont think the turn rate inconsistency in tacview is related to how tacview relates coordinates. It has to do with tiny pilot error. Ive measured the same turn time test on different airplanes with both taview and by recording the flight and timing it independently, and got the same results. After all, this is the reason a person needs a dozen or so turns to get a decent measure in either case. There are simply tons of tiny pilot errors with slightly oscillating speeds, bank angles, and angles of attack. It even varies between different aircraft, and in the same aircraft as weight and CG change depending on fuel or gun load etc. Some planes or configurations are simply easier to fly and result in far less variation in measured rate over time. You can also see large changes in measured rate in some planes if you trim the plane to counter torque, since the the torquing opposite to direction of turn makes for particularly fine control requirements on the pilot.

Ok, bro..., let's grab a plane that you decide to test, and we'll both test it and see the results in order to clear your impression that I haven't flown the plane correctly either for the turn rate tests or climb rate tests. We'll both do in-sim direct display recordings as well as Tacview recordings, compare the errors that both you and I get between the in-sim and in Tacview in terms of performances and then see how far your results were to mine. I guarantee you that I've followed the best turn rate/time speeds closely to what the manual tells, with constant Beta (sideslip angle) required to keep the turn/slip ball in the middle and only done tiny bank corrections in order to maintain a horizontal turn as well as possible and I've only took as a good reference the turn rates where I had the most constant flight, so I wasn't fouling myself for the end results! For the climb rates, yes, I've mostly tested climbing rates from SL at a constant 250km/h IAS with the specs told engine regime. Some planes like the ME-262 and P-47 which are heavier and have a quite higher than optimal AoA at those low speeds I've used higher speeds to keep an optimal climb AoA (usually around 5-7 degrees).

 

Regards!

Edited July 10, 2021 by MaverickSu-35S

[gimpy117]

so when you say "overrated" you mean that it's overperforming?

 

so that means (if this data is correct) the all early 109's are over performing in terms of turn and that the majority of early Russian fighters are under performing? 

that's a big yikes ?

 

as to the P-40...the climb rate seems to check out for this aircraft being a wallowing pig. I'd be curious to see what the P-40 could do without the ridiculous engine timers... 

[JtD]

It means that in game performance is better or worse than the developers spec sheet.

[Angry_Kitten]

18 hours ago, gimpy117 said:

so when you say "overrated" you mean that it's overperforming?

 

so that means (if this data is correct) the all early 109's are over performing in terms of turn and that the majority of early Russian fighters are under performing? 

that's a big yikes ?

 

as to the P-40...the climb rate seems to check out for this aircraft being a wallowing pig. I'd be curious to see what the P-40 could do without the ridiculous engine timers... 

when you go online and google p-40e-1 engine testing, you get lots of data pop up. From the plane manufacturer, and THOSE tests use the 50 inch manifold pressure the GAME uses.  indicated airspeeds are slightly off in game, but when one adds "correction number of 100" to get "true airspeed" its very very close

[gimpy117]

10 hours ago, pocketshaver said:

when you go online and google p-40e-1 engine testing, you get lots of data pop up. From the plane manufacturer, and THOSE tests use the 50 inch manifold pressure the GAME uses.  indicated airspeeds are slightly off in game, but when one adds "correction number of 100" to get "true airspeed" its very very close

I will add though, Alison cleared 60" manifold and also admitted the aircraft was being run at 66-70" HG. so...the P-40 being the worst pig in the game I think is very severely hampered by pre war limits. It really couldn't hurt to call off the dogs on the Alison engine, since especially the P-40 is one of the bottom fighters in game 

10 hours ago, JtD said:

It means that in game performance is better or worse than the developers spec sheet.

what exactly does that entail? I'm guessing the spec sheets are supposed to be historic 

[Angry_Kitten]

2 hours ago, gimpy117 said:

I will add though, Alison cleared 60" manifold and also admitted the aircraft was being run at 66-70" HG. so...the P-40 being the worst pig in the game I think is very severely hampered by pre war limits. It really couldn't hurt to call off the dogs on the Alison engine, since especially the P-40 is one of the bottom fighters in game 

what exactly does that entail? I'm guessing the spec sheets are supposed to be historic 

 

But did they ever do any official testing of running the engine/ plane at 50 to 70" inches?

 

The programers HAVE admitted that if someone provided them actual data from the engine folks or the plane manufacturing company we COULD have a p-40e running an engine at 50 to 70 inches on the manifold.  

 

I would LOVE that because shes a great bird, even while gimped, she rolls well, dives well. She just cant RUN.  some planes on SP, if you dont do serious damage to an engine during that initial head on pass just forget it or run to emergency and hope  you dont blow a piston out before you can close.

 

[MaverickSu-35S]

On 7/12/2021 at 12:18 AM, gimpy117 said:

so when you say "overrated" you mean that it's overperforming?

 

so that means (if this data is correct) the all early 109's are over performing in terms of turn and that the majority of early Russian fighters are under performing? 

that's a big yikes ?

 

as to the P-40...the climb rate seems to check out for this aircraft being a wallowing pig. I'd be curious to see what the P-40 could do without the ridiculous engine timers... 

Yes, and I'm not biasing anything, I'm all here to help improve the realism, not being in the mood for role playing here. Overrated, as it says, means overperforming and underrated, underperforming, all according to the clear numbers given at each plane's specifications. If those are the planes specifications, then those planes should logically respect them, right? I believe that my tests shouldn't even take place and that the told performances should be taking place with little error, +/- 1m/s climb rate and +/- 0.5s for turning times. At least that's what I consider a negligible error for accurate performances. Again, I don't want to sound like a pain in the ass for the devs, all I want to say is that the corrections are doable and we should all agree that it might require some time until each one of them is corrected. All in the name of IL-2 to stay on top and to not disappoint those who try it and to leave them with a smile on their face as they recognize how authentic the plane are replicated.

 

The P-40, yes, it's a pig when it comes to climbing, and let's say that my chosen speed for it of 250km/h climb rate constant IAS might not be the correct one for the best climb rate of this plane, so, ok, let's say that being with 1m/s underrated wouldn't matter much anyway cause it's a very poor plane for climbing whatsoever, but if it is indeed a bit underrated, that should be easily corrected as I've already suggested. Speaking of witch, I have only used a 250km/h constant IAS for climb rate for almost all aircraft, the exception being the P47s for which I used 290.300km/h IAS (as it has a big wing loading compared to other prop planes, so it would have a rather too high AoA for best climbing rate) and 400+km/h for the ME-262A which would also have a too great AoA below 300km/h and the fact that being a jet engine, the turbojet engines gain a thrust increase as a function of Mach/IAS for a constant altitude.

 

Kind regards!

On 7/12/2021 at 6:06 PM, JtD said:

It means that in game performance is better or worse than the developers spec sheet.

The planes performances don't respect the dev spec sheet, exactly. I trust them and I know that they'll be able to sort it out and correct the aero and engine power or prop thrust data. Cheers!

Edited July 14, 2021 by MaverickSu-35S