Jump to content

PSA: F-14 Performance/FM Development Status + Guided Discussion


IronMike

Recommended Posts

I'm guessing transonic drag is bonkers? And it's been so for ages now. The A model won't even go supersonic with some loads above certain altitude. 

As for the reaction to dropping ordnance.....have you guys ever fired a shoulder mounted Sparrow while in a 4x2x2 or 6x2? The thing immediately drops to the heavier side. 
 

Modules: FC3, Mirage 2000C, Harrier AV-8B NA, F-5, AJS-37 Viggen, F-14B, F-14A, Combined Arms, F/A-18C, F-16C, MiG-19P, F-86, MiG-15, FW-190A, Spitfire Mk IX, UH-1 Huey, Su-25, P-51PD, Caucasus map, Nevada map, Persian Gulf map, Marianas map, Syria Map, Super Carrier, Sinai map, Mosquito, P-51, AH-64 Apache

Link to comment
Share on other sites

On 5/15/2022 at 3:38 PM, Noctrach said:

See, I also felt the F-14 seemed slow to accelerate in the transonic region, so I did a 1:1 test with DCS compared to the AAP-11 scenario below.

Test parameters and method: Syria map, 15C, QNH 29.92. Stabilized level flight at FL35, IMN 0.70, engage full afterburner while maintaining level flight (verified altitude and level flight with info panel in transonic region to avoid instrument inaccuracies). Flare release to indicate IMN milestones for TacView analysis. (TacView shows TMN)

Taking some error margin due to inaccuracies in manual testing and interpreting the graph lines, our F-14B isn't too far off in total timing accelerating from 0.9 to 1.20. However, this is mainly due to its overperformance in going from 0.7 to 1.0, after which it hits a wall of drag that takes over 50% more time to cross than per manual. It also accelerates significantly faster  past IMN 1.60->1.65->1.69, taking a massive 3 minutes less total time to reach.

Both ways it does seem like something that might be worth looking at.

Maximum Afterburner - 66,000 lbs gross weight - 35,000 feet - ICO standard day
AIM-9*2, AIM-7*2, AIM-54*2, 2*370 Gallon External Tanks, Gun+AMMO

IMN DCS Time Mnl Time Delta DCS Dist Mnl Dist Delta
0.7 0 0 0 0 0 0
1.0 0:46 1:14 -0:28 6.6 10 -4.4
1.2 2:03 1:53 +0:10 20.3 17.5 +2.8
1.4 3:07 2:42 +0:25 34.1 28.7 +5.4
1.6 4:06 4:32 -0:26 48 56.3 -8.0
1.65 4:24 5:51 -1:27 53.2 74.9 -21.7
1.69 4:37 >7:30 -2:55 56.8 107 -50.2

 

 

Very interesting, are there data at sea level or other altitude? If the answer is yes, may i ask you if could you perform a comparison with DCS performance? Thank you a lot.

Link to comment
Share on other sites

See, I also felt the F-14 seemed slow to accelerate in the transonic region, so I did a 1:1 test with DCS compared to the AAP-11 scenario below.
Test parameters and method: Syria map, 15C, QNH 29.92. Stabilized level flight at FL35, IMN 0.70, engage full afterburner while maintaining level flight (verified altitude and level flight with info panel in transonic region to avoid instrument inaccuracies). Flare release to indicate IMN milestones for TacView analysis. (TacView shows TMN)
Taking some error margin due to inaccuracies in manual testing and interpreting the graph lines, our F-14B isn't too far off in total timing accelerating from 0.9 to 1.20. However, this is mainly due to its overperformance in going from 0.7 to 1.0, after which it hits a wall of drag that takes over 50% more time to cross than per manual. It also accelerates significantly faster  past IMN 1.60->1.65->1.69, taking a massive 3 minutes less total time to reach.
Both ways it does seem like something that might be worth looking at.

Maximum Afterburner - 66,000 lbs gross weight - 35,000 feet - ICO standard day
AIM-9*2, AIM-7*2, AIM-54*2, 2*370 Gallon External Tanks, Gun+AMMO

IMN DCS Time Mnl Time Delta DCS Dist Mnl Dist Delta
0.7 0 0 0 0 0 0
1.0 0:46 1:14 -0:28 6.6 10 -4.4
1.2 2:03 1:53 +0:10 20.3 17.5 +2.8
1.4 3:07 2:42 +0:25 34.1 28.7 +5.4
1.6 4:06 4:32 -0:26 48 56.3 -8.0
1.65 4:24 5:51 -1:27 53.2 74.9 -21.7
1.69 4:37 >7:30 -2:55 56.8 107 -50.2
 
Very nice table, thank you very much!

It's not so strange that relationship DCS/manual sort of breaks down near the asymptote (i.e. max IMN). The same would likely be true for real aircraft/manual.

But there does seem to be rather large discrepancies across a large speed range. I think you have a small problem in your testing in that you should have started at Mach 0.6725 rather than 0.7 if gross weight was 66,000 lbs.

I will try to reproduce your results, and then I'll see if I can get some data for 25000' and 4 x aim-7 + 4 x aim-9, because I think this is worse accuracy-wise. Based purely on my gut feeling, but let's measure it and see.
Very interesting, are there data at sea level or other altitude? If the answer is yes, may i ask you if could you perform a comparison with DCS performance? Thank you a lot.
There is data for 5, 15, 25 and 35 thousand feet for 3 different loadouts. I.e. 12 different charts each covering a range of gross weights, so any sort of exhaustive testing is going to be very tedious.

Link to comment
Share on other sites

2 hours ago, Katj said:

I will try to reproduce your results, and then I'll see if I can get some data for 25000' and 4 x aim-7 + 4 x aim-9, because I think this is worse accuracy-wise. Based purely on my gut feeling, but let's measure it and see.There is data for 5, 15, 25 and 35 thousand feet for 3 different loadouts. I.e. 12 different charts each covering a range of gross weights, so any sort of exhaustive testing is going to be very tedious.
 

 

Bust out the ol' VX-4 livery and see if it at least takes the edge off, imo

  • Like 1
Link to comment
Share on other sites

Looking at the results of that test, i think that while the overall acceleration is all over the place, not all parts of the envelope are created equal. That is, the time you'll spend being above mach 1.4 or the situations that warrant it, is much less pronounced then the time you'll be between 0.9 and 1.2. At least for most BVR. BTW, has anyone done this for the A model? 

Modules: FC3, Mirage 2000C, Harrier AV-8B NA, F-5, AJS-37 Viggen, F-14B, F-14A, Combined Arms, F/A-18C, F-16C, MiG-19P, F-86, MiG-15, FW-190A, Spitfire Mk IX, UH-1 Huey, Su-25, P-51PD, Caucasus map, Nevada map, Persian Gulf map, Marianas map, Syria Map, Super Carrier, Sinai map, Mosquito, P-51, AH-64 Apache

Link to comment
Share on other sites

I tried to reproduce the results and while I didn't get exactly the same numbers it was basically the same. Heatblur F-14 overperforming to Mach 1.0, and then having abysmal performance to Mach 1.2 and also 1.4, then overperforming.

I also tested 25k with 4+4 loadout 60,000 lbs and my gut feeling was incorrect. It was the same overperformance to Mach 1, abysmal performance to 1.2, but then overperformance all the way to the top.

Going from Mach 1.0 to 1.2 takes almost double the time it should according to the charts.

Also looking at the specific exess power charts, acceleration should start increasing at about Mach 1.05, but in DCS you're hitting that brick wall all the way up to Mach 1.2.

I haven't had time to compile tables, and Honestly I don't know if I will.

Link to comment
Share on other sites

I tested acceleration in the 0/4/4 55.6k 5,000ft configuration for the B and there was nowhere that it over performed.  .7-1.1 it definitely underperformed.  Used the EM plot Ps values.

And as far as I am aware there is a transonic drag bug for weapons, which is why you see the dismal performance around M1.


Edited by Spurts
Link to comment
Share on other sites

54 minutes ago, Spurts said:

I tested acceleration in the 0/4/4 55.6k 5,000ft configuration for the B and there was nowhere that it over performed.  .7-1.1 it definitely underperformed.  Used the EM plot Ps values.

And as far as I am aware there is a transonic drag bug for weapons, which is why you see the dismal performance around M1.

 

Please could you share the datas of your test? Thanks

Link to comment
Share on other sites

I tested the

"MAXIMUM AFTERBURNER CLIMB TO 45,000 FEET AND INDICATED MACH 1.4
AIRCRAFT INITIAL GROSS WEIGHT: 62, 161 POUNDS"

And it took me about 4:10 (should be 3:26), or about 120 % of manual time. It consumed 6,700 lbs of fuel, should be 5,3.

The 64k loadout (2+2+2) took about 4:40 (should be 3:45) or 124 % of manual time. Fuel consumed was 7400 lbs (5700).

The 68k loadout (2+2+2+tanks) was pretty spot on. Took me about 5:30 (about 5:45 per manual), 96 % of manual time and 7500 lbs (8000).

... but I'm no pilot so ymmv and all that.

Link to comment
Share on other sites

Sounds like the missile transonic wave drag rearing it's ugly head again.  And apparently the tanks have little/no transonic wave drag?  only way I can account for the difference in time being smaller when adding tanks, unless it's just a deviation in test as we are not professional test pilots here.

Link to comment
Share on other sites

Sounds like the missile transonic wave drag rearing it's ugly head again.  And apparently the tanks have little/no transonic wave drag?  only way I can account for the difference in time being smaller when adding tanks, unless it's just a deviation in test as we are not professional test pilots here.
I don't have time to analyze it further right now but the same loadout performed badly in the transonic region per the level acceleration discussion above. So my guess is that it overperformed at either the subsonic or supersonic phase of the flight.

The 0+4+4 loadout looked pretty close at the subsonic and supersonic climb phases, but the transonic phase was way slow.
Link to comment
Share on other sites

21 hours ago, Spurts said:

Sounds like the missile transonic wave drag rearing it's ugly head again.  And apparently the tanks have little/no transonic wave drag?  only way I can account for the difference in time being smaller when adding tanks, unless it's just a deviation in test as we are not professional test pilots here.

Wasn't there an engine bug a while back where empty tanks had either negative effective weight or negative effective drag (or both)?  If that was/is the case, getting performance right in any case where tanks are present is going to cause everything else to be wrong.

 

It's got to be frustrating to be in HB's shoes.  Lots of knobs to turn and they all interact.  Same for everyone else trying to make a high-fidelity module.

Link to comment
Share on other sites




It's got to be frustrating to be in HB's shoes.  Lots of knobs to turn and they all interact.  Same for everyone else trying to make a high-fidelity module.

Seems to me like exactly the sort of thing that would be hell to do manually.

I guess one could view it as an optimization problem. You want to minimize the error when compared to the published performance graphs. This isn't my field, but it would surprise me greatly if it is a convex problem. It is thus likely that you end up in local minima when you turn the knobs to try to match the graphs. That risk is probably even greater if you start out focusing at matching only a few graphs (e.g. STR). In that case you might very well end up in a situation where the best thing to do is to just start over from the beginning.

It's probably very doable to implement a numerical solver, though.
  • Like 1
Link to comment
Share on other sites

You know all of this could be solved in seconds if ED wasn’t so protective of every little aspect of the games backstage functioning. You could just SEE what the game thought it was doing. 

I don’t think it should be open source to users but there should definitely be that level of transparency between ED and 3rd party devs.
  • Like 1
Link to comment
Share on other sites

On 5/28/2022 at 2:35 AM, r4y30n said:


I don’t think it should be open source to users but there should definitely be that level of transparency between ED and 3rd party devs.

You mean Open OUTPUT. Just enumerating all the sim variables is not "Open Source"

Link to comment
Share on other sites

On 3/23/2022 at 6:42 PM, fat creason said:

Have you put any personal effort into finding out? Would you even bother to test or notice a FM change without me telling you it happened? Personal initiative and curiosity are good things.

 

Also: Comparing performance among two aircraft in DCS is a fool's errand unless you know for sure one aircraft's performance is 100% correct. This is extremely unlikely.

 

Please, can you tell us something about this? Thank you

 

ScreenHunter_1085 May. 31 12.32.jpg


Edited by maxsin72
Link to comment
Share on other sites

2 hours ago, maxsin72 said:

Please, can you tell us something about this? Thank you

I wouldn't expect anything to come of that video.  "Oh hey maybe our MF drag is a tad off" is a maybe and "Our testing scripts allow us to hold the precise AoA more accurately and shows more accurate results" is more likely, and I would welcome either equally.  

Link to comment
Share on other sites

It would be interesting to understand a little bit about the methodology or toolchain though. Does the DCS SDK allow the external flight model to be tested in a scripting harness, or is the flight model itself tested with fixed inputs against the baseline in some other tool, then exported for use in DCS.

If it was the latter then you can introduce errors simply because the inputs can be off, just as an example OAT might introduce an error compared to the external tool.

  • Like 1
Link to comment
Share on other sites

7 hours ago, Spurts said:

I wouldn't expect anything to come of that video.  "Oh hey maybe our MF drag is a tad off" is a maybe and "Our testing scripts allow us to hold the precise AoA more accurately and shows more accurate results" is more likely, and I would welcome either equally.  

I thought the same but any good answer is wellcome. I also ask to myself if, when the script give the right result, could be a good idea if a real person (a skilled one who knows what to do) test the FM to see if he is able to obtain the same result as the script because, in the real world, there were not a script to pull the plane to his limit.

Link to comment
Share on other sites

7 hours ago, RustBelt said:

But, in the real world, no two planes ever performed exactly to the book either. Although being clear on the margin of acceptable deviation would also make it easier to discuss any results realistically.

I think it's enough to simulate a standard airplane who reproduces well NATOPS standard, don't you think? 


Edited by maxsin72
Link to comment
Share on other sites

1 hour ago, maxsin72 said:

I think it's enough to simulate a standard airplane who reproduces well NATOPS standard, don't you think? 

 

But then you have to also take into account NATOPS charts are extrapolated test data and therefore not entirely accurate to real life performance either... from the video you screengrabbed it seems to me all of this is very much within the 5% margin of simulation error. This is fast approaching what the dutch call "geneuzel in de marge" i.e. putting way too much stress on unimportant details.

Does it really matter whether it's -1200 or -1000 Ps if the end result is still going to be that you're wallowing at stall speed...

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...