Missile malfunction - bug or feature?

[ZaneWuzHere]

I always they fired slowly at first because it was launched off of the pylon, then the motor fired and the missile zoomed to it's unlucky target.

[Fritz]

In one of the earlier realeses 1.65 or close above I had a Python, Not sure what it was called now.

If the same problem existed even with the classic Python models it can't be some detail in the missile_launch_position as I suspected. Perhaps the problem exists with all big/long ships? Does anybody fly a Boa or Anaconda with missiles? I can't afford one to test it.

[Anonymissimus]

In one of the earlier realeses 1.65 or close above I had a Python, Not sure what it was called now.

Does anybody fly a Boa or Anaconda with missiles? I can't afford one to test it.

I have a Boa 2 Cruiser save. Yes, can reproduce. ECM hardened missile shot onto an asteroid at full speed, flying straight.

[Fritz]

I played around with an Adder (how can anybody aim with this over-sensible little thing? ). It also shows the hesitating effect of the missile but it doesn't cause trouble. Obviously the problem lies somewhere in the core routines that handle the missile launching: The missile object is created with a very low speed (zero?) instead of the ship's speed.

It can probably be solved easily by moving the missile launching position a bit to the front. But I don't know how to interpret these coordinates:

missile_launch_position = "0.0 -5.1 19.2374"; // Adder

missile_launch_position = "0.0 -2.2 63.5"; // Python

[Smivs]

It is the hesitation that is the problem, not the launch position. The launch position should be 'on' the ship otherwise it looks silly. My brief test last night left me with the impression that the missile seems to launch 'engine off', then a moment later it ignites and the missile takes off.
The other mystery is why it only seems to happen in standard view, but not when viewed externally.
Just for full disclosure the tests were done using trunk v1.3.0.6557.150601-3f1779f (yes I know that's quite old now).

The coordinates are the position on the ship relative to its centre, defined on the x,y and z axes. So the Adder's position is central horizontally, -5.1m below centre vertically and 19.2374m forward of centre.

[Fritz]

The other mystery is why it only seems to happen in standard view, but not when viewed externally.

It does happen! I just tested it in the rear left view, and I had 3 malfunctions with 5 missiles fired under identical conditions. For me the main mystery is that it is a random thing: What part in the missile software except the reaction to ECM could be random?

I think the launching position should be lower because it looks at if the missile is fired from the front view "camera". The position of the latter is positioned at the tip of the nose on the Python:

view_position_forward = "0.0 1.18304 62.0972";
missile_launch_position = "0.0 -2.2 63.5";

The missile is launched from a position (2,2 + 1,183) = 3,383 m lower, which isn't much in this scale (the missile diameter is bigger!), but this isn't the problem.

The problem is that the Python is only (63,5 - 62,1) = 1,4 m behind the launching position, and even a small delay will cause it to hit the missile - at least in theory, I don't know how hitting is calculated by the software.

I moved the launch position 10 m to the front: No real change. 20 m? No change. 50 m? No change.

100m? Better, but still only 4 of 5 fired successfully. It can be seen in the side view that each missile indeed behaves differently. Even some of the successful launches just avoided being hit by the ship! With 175 m I still had a mishap, but 200 m seem to work reliably. But of course it doesn't look good any more...

Edit: And even with 200 m I still managed to hit a missile that was turning to a position below the ship...

[Fritz]

And why doesn't it happen with the Adder or the Cobra Mk III? They do catch up with the missiles too, but the latter are simply flying lower than the bottom of the ships, so nothing happens.

So the timing problem exists with all ships, but the missiles are hit only by some models.

[Fritz]

Here you can see the missile actually flying behind the ship!



I wonder how this works in reality. Do wing-mounted missiles drop behind for a moment? This seems dangerous!

[Stormrider]

And why doesn't it happen with the Adder or the Cobra Mk III?.

The Cobra Mk III missile launch position is well below the actual ship. At least when I converted the dat to an obj and loaded it into wings I found a discrepancy. The real issue is that the missile doesn't even launch at ship velocity and the momentary pause before the missile engine fires up. I wonder about the AI, as I said the missile racks in the armoury oxp don't seem to behave this way.

[Diziet Sma]

I played around with an Adder (how can anybody aim with this over-sensible little thing?)

Oolite has a precision mode for fine control. If you're playing with the keyboard, holding the <Ctrl> key down while using the roll/pitch/yaw keys gives you precision mode. If using a joystick, you can set one of the buttons on the 'stick to toggle precision mode on/off (though correct editing of the joystick response curves will give you even better control than precision mode is capable of).

Comes in handy for docking, as well as in combat.

[Norby]

There is a 0.5s delay after launch before a core missile ignite its rocket, so regardless of got velocity from the ship at start the game think in the first half second as if the missile would like to stop so use all thrust to break to zero as soon as possible. This cause the missile left behind in the case of a moving Cobra3 where there is enough space below and hit the Python where is not.

Now I made a [wiki]Missile Booster[/wiki] OXP which prevent the initial collision by boosting the launch speed of missiles. Should work even at injector speeds also.

Please install in the in-game expansion manager within the Mechanics category.

[Smivs]

There is a 0.5s delay after launch before a core missile ignite its rocket...

Yes, I noticed that in the AI. The odd thing is that it's always been there (as a check over some old versions confirms) yet this problem seems to be a new one.
Thanks for the OXP fix, but this still looks like it might be a core issue, and if so needs to be sorted out properly really.

[Cody]

Curiously, from my Griff Cobra I can fire missiles with no problems (even at injector speed) - is it only the bigger ships?

[Smivs]

- is it only the bigger ships?

No. I just tried several ships (dependant on what saves I had loaded) with mixed results.
Cobra 2X (tiny!) Problem present
Cobra 3 (BullDog One - I got Jack to help ) Successful launch
Asp-X (medium sized) Unsuccessful launch
Boa Clipper (same as BCC) - unsuccessful launch

[Stormrider]

- is it only the bigger ships?

No. I just tried several ships (dependant on what saves I had loaded) with mixed results.
Cobra 2X (tiny!) Problem present
Cobra 3 (BullDog One - I got Jack to help ) Successful launch
Asp-X (medium sized) Unsuccessful launch
Boa Clipper (same as BCC) - unsuccessful launch

Did you guys miss this?

The Cobra Mk III missile launch position is well below the actual ship. At least when I converted the dat to an obj and loaded it into wings I found a discrepancy.

I noticed this missile problem about a year ago but nobody replied to my post so I figured that no one else was having an issue. Apparently nobody noticed because most players are flying the cobra.

I like to launch while on injectors to add extra velocity to the missile but this has made that tactic useless. When I fire a missile with my ship at a complete stop the missile appears in front of me then comes to a complete stop, pauses for about a half second then takes off. I had a look at the missile rack AI and there is a pause after launch with it too, yet they fire consistently and the exhaust plume is always present when they appear. In fact the only difference I can think of is that the missile racks don't use the js AI. I have noticed that the js AI for my oxp station defence ships is much slower to respond to an attack that the old plist AI.