DCS: F-14 Development Update - AIM-54 Phoenix Improvements & Overhaul - Guided Discussion

[Machalot]

18 minutes ago, SuperEtendard said:

Burn rate and thrust varies depending on the temperature of the propellant, and this will be affected depending on the altitude of launch (external ambient temperature) and speed of the launching plane (heat caused by air drag).

This is true about the burn time, but I'm skeptical that the propellant mean bulk temperature (PMBT) would change fast enough to really respond to fight conditions. The temperature response time is typically measured in hours.  I think for most DCS sorties the PMBT will pretty much be what it was before takeoff, which will be whatever it was where the missile was stored. I could be wrong for smaller missiles. 

[roobarbjapan]

37 minutes ago, SuperEtendard said:

Burn rate and thrust varies depending on the temperature of the propellant, and this will be affected depending on the altitude of launch (external ambient temperature) and speed of the launching plane (heat caused by air drag). Air densitiy also seems to cause slight differences.

Here is an example for the rocket motor in the Soviet R-3S missile:

Thrust in Kgf in the y axis, seconds of burn time in the x axis.



While -54ºC to +60ºC is a big temperature range, you can see that say +15ºC at sea level in a standard day and -40ºC at 30K feet (standard day as well) would mean a rather different thrust profile for the same missile, you would also have to add the drag heating on top of it, which would also be different based on altitude.

Currently in DCS we have fixed thrust and burn time values, so this imposes a limitation on rocket motor modelling.

Thank you for the explanation.

[Wolfman289]

 I just played several quick start missions where you engage multiple targets with the AIM54, like the Mariana's mission with the 4 JF's and Final Hour 2. I can't get the AIM54 won't hit anything. I've tried it several times and different speeds, altitudes, and ranges. I've had 2 hits out of 32 or so missiles fired. I'm typically firing in TWS at around 35,000 ft to 40,000, 30-40 miles away and around Mach 1.1 or 1.2. Some of these missiles are going as high as 70,000 ft during lofting and then start to drop down on the target, but cruise right by it. Some don't even take a correct trajectory to dive at enough of an angle to hit the target.  I am centering the T symbol and I'm firing with the jet level. Meanwhile, I'm already having SD-10's fired at me that are tracking just fine.

What is going on? I can't figure this out, but it's pretty frustrating. Also, I've tried every variant of AIM54 that's available on these two missions. I've had one hit with the A and one with the C. I've even tried at 25 miles away. Still nothing.

Thanks all. I really appreciate any help or pointers, especially if I'm just doing something wrong.

Edited September 4, 2022 by Wolfman289

[SuperEtendard]

21 minutes ago, Machalot said:

This is true about the burn time, but I'm skeptical that the propellant mean bulk temperature (PMBT) would change fast enough to really respond to fight conditions. The temperature response time is typically measured in hours.  I think for most DCS sorties the PMBT will pretty much be what it was before takeoff, which will be whatever it was where the missile was stored. I could be wrong for smaller missiles. 

Yeah it's true it will depend on the propellant specific heat and their total mass (I guess also the chamber diameter and lenght ratio) This PMBT change you mention of hours for which example would it be, something like an ICBM / satellite launch rocket? Maybe there are studies that show how this scales down to the sizes of air to air missiles.

 

[Machalot]

57 minutes ago, SuperEtendard said:

Here is an example for the rocket motor in the Soviet R-3S missile:

Thrust in Kgf in the y axis, seconds of burn time in the x axis.

Another consideration is that burn time is defined differently by different sources.  It can be defined as the point when the thrust curve drops below say 90% of its peak, or drops below say 10% of average, or when some percentage (e.g. 90%) of nominal impulse has been delivered, or chamber pressure drops below a threshold like 2000 psi, etc.  So it's not even necessarily when the thrust goes to "zero" (if such a point can even be reliably measured and is reproducible).  With that in mind, some possible burntime values from the -54C curve in your graph could be 3.7, 3.5, or 3.2 sec.  For a differenr missile with a 25 sec burntime that could be a span of several seconds. 

[captain_dalan]

3 hours ago, Golo said:

... either that shot should never be allowed by WCS...

Unless you have intimate knowledge of the F-14 WCS that you willing-allowed to share with ED and/or HB, i don't think this statement holds. As of right now, the AMRAAM can be fired in the F-15, F-16 and F-18 well outside its battery life or ballistic range. Why would the F-14 be any different? 

1 hour ago, Wolfman289 said:

 I just played several quick start missions where you engage multiple targets with the AIM54, like the Mariana's mission with the 4 JF's and Final Hour 2. I can't get the AIM54 won't hit anything. I've tried it several times and different speeds, altitudes, and ranges. I've had 2 hits out of 32 or so missiles fired. I'm typically firing in TWS at around 35,000 ft to 40,000, 30-40 miles away and around Mach 1.1 or 1.2. Some of these missiles are going as high as 70,000 ft during lofting and then start to drop down on the target, but cruise right by it. Some don't even take a correct trajectory to dive at enough of an angle to hit the target.  I am centering the T symbol and I'm firing with the jet level. Meanwhile, I'm already having SD-10's fired at me that are tracking just fine.

What is going on? I can't figure this out, but it's pretty frustrating. Also, I've tried every variant of AIM54 that's available on these two missions. I've had one hit with the A and one with the C. I've even tried at 25 miles away. Still nothing.

Thanks all. I really appreciate any help or pointers, especially if I'm just doing something wrong.

 

Some tacviews would be helpful. I just spent the better half of the night doing BVR instant actions (exported into my own mission folder so i can tweak and experiment with them) and observing the AI's behavior. It looks like the latest patch incorporated some new AI behavior, right out of the GS server meta mindset. The AI's are now highly reluctant to engage beyond minimum engagement range AND they go cold as soon as the missile fired reaches the scripted range of 10 miles, probably 75+% of the time if not more. This means they will likely outrun any initial shot, unless that shot is performed under very specific circumstances. This also means that they are even less likely to support their own missiles then they used to be. BVR engagements are right now essentially turkey shoots. Fire at long range, the AI will forget about its own shots and turn cold, then run them down as you are both faster and carry more fuel. Heck, i just fought a MiG-31 10 times in a row. The poor thing didn't even get to fire twice in those 10 engagements. AI fights haven't been this boring in a long, long time...

EDIT: this of course extends to your wingman as well. Expect a completely impotent fire support from him. Sometimes he gets knocked out without firing a single shot as well. 

Edited September 5, 2022 by captain_dalan

[Wolfman289]

17 minutes ago, captain_dalan said:

Unless you have intimate knowledge of the F-14 WCS that you willing-allowed to share with ED and/or HB, i don't think this statement holds. As of right now, the AMRAAM can be fired in the F-15, F-16 and F-18 well outside its battery life or ballistic range. Why would the F-14 be any different? 

Some tacviews would be helpful. I just spent the better half of the night doing BVR instant actions (exported into my own mission folder so i can tweak and experiment with them) and observing the AI's behavior. It looks like the latest patch incorporated some new AI behavior, right out of the GS server meta mindset. The AI's are now highly reluctant to engage beyond minimum engagement range AND they go cold as soon as the missile fired reaches the scripted range of 10 miles, probably 75+% of the time if not more. This means they will likely outrun any initial shot, unless that shot is performed under very specific circumstances. This also means that they are even less likely to support their own missiles then they used to be. BVR engagements are right now essentially turkey shoots. Fire at long range, the AI will forget about its own shots and turn cold, then run them down as you are both faster and carry more fuel. Heck, i just fought a MiG-31 10 times in a row. The poor thing didn't even get to fire twice in those 10 engagements. AI fights haven't been this boring in a long, long time...

EDIT: this of course extends to your wingman as well. Expect a completely impotent fire support from him. Sometimes he gets knocked out without firing a single shot as well. 

 

Thanks for the reply. What's interesting is that when I try the Final Hour mission, the MiG's really don't turn cold after I fire. They angle themselves, but still are pointed towards my general area. On The Mariana's mission with the JF's, they continue towards me head on, nose hot and fire SD-10's at me and my wingman. I still can't get any AIM54's to hit on either mission, even with the JF's that stay nose hot.

I'd attach some tac views, but I don't have any, sorry!

[HeavyGun1450]

Yes. currently the AI do not care what setting the target size switch is set to in the F14, they immediately go defensive when they detect a missile 10nm from them, not when the missile goes active because its a ED problem. When they go defensive they proceed to immediately notch the missile, resulting in the missile flying right past them.

https://forum.dcs.world/topic/301471-bandit-ai-defending-aim-54-at-10nm-regardless-of-the-set-active-range/#comment-4974983

Edited September 5, 2022 by HeavyGun1450

[captain_dalan]

53 minutes ago, Wolfman289 said:

Thanks for the reply. What's interesting is that when I try the Final Hour mission, the MiG's really don't turn cold after I fire. They angle themselves, but still are pointed towards my general area. On The Mariana's mission with the JF's, they continue towards me head on, nose hot and fire SD-10's at me and my wingman. I still can't get any AIM54's to hit on either mission, even with the JF's that stay nose hot.

I'd attach some tac views, but I don't have any, sorry!

I'll try the missions myself tomorrow if i can. I remember the SD-10 being a real pain in the behind, much more so then the AMRAAMs back in the day. Unfortunately i don't own the JF-17, so i could never test its weapons and ordnance performance. The AMRAAMs are currently in a rather bad shape, but the SD-10's may not be. If so, they could be a much more dangerous missile then the AIM-54, regardless of variant.

41 minutes ago, HeavyGun1450 said:

Yes. currently the AI do not care what setting the target size switch is set to in the F14, they immediately go defensive when they detect a missile 10nm from them, not when the missile goes active because its a ED problem. When they go defensive they proceed to immediately notch the missile, resulting in the missile flying right past them.

https://forum.dcs.world/topic/301471-bandit-ai-defending-aim-54-at-10nm-regardless-of-the-set-active-range/#comment-4974983

 

It's more then that. Before this patch, an AI set to veteran (not to ALL mission and campaign creators, please don't set your AI's to ace, it cripples them), was just as likely to do a Split-S at 10 miles, as it was likely to offset for a few seconds and then notch. That is, the chances between few more seconds of missile support and a complete trash were about 50-50. This doesn't mean anything for FOX-1's, as they got trashed anyways, being dependent on the mothership support all the way in. But it meant precious few seconds of support for the FOX-3 shots. With the latest patch the almost ALWAYS goes for the Split-S, ditching all support. It even ignores notching. Just a plain 180, about turn, and burning the hell out of Dodge. Almost every fight ends up in a chase scene. What ever they did, they turned the PvE and SP into a bore fest. 

[WarthogOsl]

I feel like since the update, I'm losing way more tracks and having AIM-54A's never go pit bull then previously.  I'm often going against a pair of bandits, and one missile will track just fine, while the other will just continue straight on, and I never get an active indication on the TID.

I've attached a track file of one of these engagements against a pair of F-5's that just happened, and wonder if someone could explain what is going on.  The lead gets hit by my first missile.  The trailer starts to make a defensive maneuver when the second Phoenix is over 16 miles away.  This raises a few more questions:  If I never got an active indication, what is the bandit responding to?  And why is he reacting to the missile beyond the 10 miles out that AI usually reacts at?  Then the turn he makes seems hardly enough to notch the missile (though I guess that doesn't matter if the missile never really turned his radar on).  Anyway, I'm not sure what's up.

noactive.zip.acmi

[Lykurgus]

Does the missile as modeled in dcs gradually lose weight as the propellant burns during the flight? The highest acceleration should be the seconds right before it burns out. From all the test shots I have been doing it seems like even extremely faborable scenrios like starting at 40k at mach 1.4 the c model with mk47mod1 and mk60 only reach mach 3.6 or 3.7. 

I love the patch because the old missiles were definitely over performing. However it seems like they may be just a tad too slow even up high. 

[howie87]

This is a good point. From my testing last night it also seems like missiles fired in TWS fly a loft profile designed to arrive at the position the bandit is at when the shot is fired (rather than an intercept course). With the height of the loft, the control surfaces don't seem to have much authority for mid course corrections and the missile goes long and loops behind the bandit (front aspect shots fired at Rmax). Wil post a track tonight.

Edited September 5, 2022 by howie87

[Lurker]

I'll just leave this here, it's an interesting tool: https://www.engineeringtoolbox.com/air-altitude-pressure-d_462.html

How does this compare to the air pressures we get in DCS World. I feel like there has always been a pretty large discrepancy between "REAL" Air-pressure at different altitudes\temperatures compared to what is simulated in DCS World, which is why we get somewhat strange discrepancies in missile performance across the board. 

Edited September 5, 2022 by Lurker

[RoyalWaffles]

I just flew a single mission, angels eeeeh about 40000 against a non maneuvering IL-78 dead hot aspect at 40000 feet with a aim 54c mk 47, this glorious missile topped out at a blistering mach 3.12 in its loft and DIED at 64.29NM from its launch point completely falling limp short of the target by like 20 miles (i launched at 100) 
 

Limp aim 54.acmi

 

Sorry wrong tacview, here you go folks

Edited September 5, 2022 by NotAFlanker
Sorry sorry wrong tacview

[KlarSnow]

39 minutes ago, NotAFlanker said:

I just flew a single mission, angels eeeeh about 40000 against a non maneuvering IL-78 dead hot aspect at 40000 feet with a aim 54c mk 47, this glorious missile topped out at a blistering mach 3.12 in its loft and DIED at 64.29NM from its launch point completely falling limp short of the target by like 20 miles (i launched at 100) 
 

Sorry wrong tacview, here you go folks

 

Your missiles battery died 200 seconds in. You launched it at .8 mach, against targets that are going .6 mach.

This is what the actual 110 mile shot profile looked like, from the Outsiders View of the Awg-9 weapons system.

F-14 at Mach 1.5, Target at Mach 1.5

[KlarSnow]

And here is what that looks like in DCS. I fired a 54C-Mk60 and Mk47, and then repeated with a 54A-Mk60 and Mk47 so you can see that all missile variants meet the performance. Fired in PD-STT. Only difference from the real test is the AI wont fly at 50,000 feet. Highest I could get the Backfire was 43600 feet. All missiles impacted with ~50 seconds of battery life to spare.

110NmPhoenix shot.zip.acmi 110NmPhoenix-A shot.acmi

Edited September 5, 2022 by KlarSnow

[KlarSnow]

Just one thing I wanted to note on this, in the outsiders view for the 110 mile shot, it states the phoenix topped out at 103,500 feet and travelled a distance of 72.5 nautical miles to impact the target.

Note how closely the current AIM-54A Mk47 matches that. Tops out at 103,943 feet (103,500 for the real world) and travelled a ground distance of 74 Nautical miles to impact the target. That's.... quite close.

[RoyalWaffles]

On 9/3/2022 at 6:52 PM, DSplayer said:

For those who want to visualize the difference between the nozzle_exit_areas I've calculated (0.04525 m^2 vs 1e-6 aka 0.000001 m^2), here are some graphs. The performance difference down low is expected to be pretty marginal but once up high, its absolutely stunning to see the difference. Excuse the colors on the graphs since this was pretty rough and quick and I just reused the previous graph's colors.

 

Click here for link to charts so you can hover over the values to compare with the motor performance from prior to the September 2nd patch

 

500m

  Reveal hidden contents

 

6km

  Reveal hidden contents

 

 

12km

  Reveal hidden contents

 

 

Conclusion: The AIM-54, both in the Mk47 and Mk60 configurations, have an additional peak Mach speed of ~0.6 Mach at 12km with the nozzle_exit_area of 0.04525 m^2.

EDIT: Here are some bar graphs for Peak and Average speed for the missiles at 12km.

Peak

  Reveal hidden contents

 

Average

  Reveal hidden contents

 

Here's the differences between the Weapons.lua (stock and new nozzle) that I used in order to perform these tests: https://www.diffchecker.com/v3DOf6fU

 

THIS is interesting, If its as simple as changing one or two values in the lua why hasn't heatblur adjusted it already? Or has heatblur opted to try to compensate for this just by increasing the rocket motors thrust output by a certain percentage?

[BIGNEWY]

Hi all, 

just a reminder, please keep to the forum rules when posting. 

thread cleaned

[KlarSnow]

So just to check since it was so close. The 110Nm Shot in the real world test terminated at 157 seconds of flight, assuming the target maintained a constant mach 1.5. If so, the AIM-54A in that shot travelled at an average speed of Mach 2.9 (1660 knots) to consumate the intercept at 72.5 Nautical miles of travelled distance.

In my Shot against a slightly lower altitude target, the missile traveled a ground distance of 74 Nautical miles in 152 seconds. An average speed of Mach 3.05 (1750 Knots). So even now while this is definitely within any reasonable margin of error, the AIM-54A Mk-47 is just barely overperforming the real world.

There is of course enough room for error in either my performance in flying and employing, or in details that are unknown from the real world test shot for this to be used as any more than a curiousity and a datapoint. Overall that it is THAT CLOSE is quite cool.

 

Just to put some percentages to it, thats a time error of 3% shorter, a distance travelled error of 2% further, and a speed error of 4% faster.

I'm going to dig up the old pre-patch motor values If I can find them and redo the shots and see how far off they were.

 

Ok, Redid the shot above with the Pre-Patch Phoenixes just for comparison. The Old Mk-47A completed the shot in 138 seconds, travelled 77 Nautical Miles in that time, and flew at an average mach of ~3.35

percentage errors from the real shot are 13% shorter on time, 6% further on distance, and 15.5% faster on speed.

The Old Mk-60 completed the shot in 121 Seconds, Travelled 80 Nautical Miles, and flew at an average mach of ~3.97

percentage errors from the real shot are 23% shorter on time, 10% further on distance, and 37%  faster on speed.

 

110NmPhoenix-A PREPATCH shot.zip.acmi

110NmPhoenix-A CURRENT PATCH shot.acmi

Edited September 5, 2022 by KlarSnow

[howie87]

I think that speaks volumes for the work that Heatblur have been doing. It sounds like the kinematic performance is largely correct. I think most of the remaining issues are guidance and loft related. Specifically, the TWS loft profile not accounting for target lead 

[The_Tau]

20 minutes ago, howie87 said:

I think that speaks volumes for the work that Heatblur have been doing. It sounds like the kinematic performance is largely correct. I think most of the remaining issues are guidance and loft related. Specifically, the TWS loft profile not accounting for target lead 

Well and magic AI spotting missile at 10nm no matter what

[IronMike]

4 hours ago, NotAFlanker said:

THIS is interesting, If its as simple as changing one or two values in the lua why hasn't heatblur adjusted it already? Or has heatblur opted to try to compensate for this just by increasing the rocket motors thrust output by a certain percentage?

Unfortunately it is not as simple as just setting the value to what DSplayer has suggested above. Without going into detail, the gest is: the output from the value is skewed - this is unfortunately a legacy issue from how the phoenix was modeled originally. This value has been intentionally kept as is thus far.

If you would feed the correct value now, it would make the missile drastically over-perform. And we are currently investigating with ED how to best approach this offset. So we cannot just simply change a value in the lua, to get the correct outcome, if that makes sense.

Please also note: the nozzle exhaust area is not the silver bullet some believe it would be. If you feed the correct value now, as mentioned, it over-performs, so this is likely where some high hopes originate from. EDIT: to redact predictions here again, thinking twice it is a bit premature to make definitive claims, we'll have to see, but I would advise caution against having too high hopes for the NEA to have a massive impact overall.

Edited September 5, 2022 by IronMike

[DSplayer]

17 minutes ago, IronMike said:

Unfortunately it is not as simple as just setting the value to what DSplayer has suggested above. Without going into detail, the gest is: the output from the value is skewed - this is unfortunately a legacy issue from how the phoenix was modeled originally. This value has been intentionally kept as is thus far.

If you would feed the correct value now, it would make the missile drastically over-perform. And we are currently investigating with ED how to best approach this offset. So we cannot just simply change a value in the lua, to get the correct outcome, if that makes sense.

Please also note: the nozzle exhaust area is not the silver bullet some believe it would be. If you feed the correct value now, as mentioned, it over-performs, so this is likely where some high hopes originate from. The end result should show a slight increase in performance up high, and practically no or almost no visible change in performance down low.

 

Ok it’s just odd how other missiles that use the older FM scheme/API like the R-27s and some of the AIM-9s are able to maintain an accurate nozzle_exit_area while maintaining a thrust and total impulse values that, in the AIM-9L’s case at least, matches up with now declassified SAC/SMC documents that describe the motor’s performance at sea level but I can assume that limitation comes from the older missile API and older Phoenix modeling like you said. 

[IronMike]

18 minutes ago, DSplayer said:

Ok it’s just odd how other missiles that use the older FM scheme/API like the R-27s and some of the AIM-9s are able to maintain an accurate nozzle_exit_area while maintaining a thrust and total impulse values that, in the AIM-9L’s case at least, matches up with now declassified SAC/SMC documents that describe the motor’s performance at sea level but I can assume that limitation comes from the older missile API and older Phoenix modeling like you said. 

Primarily due to the initial phoenix modeling.