Which brings me to the AIM-54. Called the Phoenix but known as the Buffalo, despite a 'Mach 4 class' (Mach 3.8, AIM-54A Rocketdyne Mk.47 motor) and even 'Mach 5 class' (Mach 4.5, AIM-54C, Aerojet Mk.60 motor) what you have in this weapon is a 'trainwreck' (or a thundering herd of buffalo, driven off a cliff) mechanic in which a very slow start and midcourse also employs
terminal dive attacks that (though greatly complicating endgame intercept geometry for the missile) reenergize it's final approach.

As an example of how bad this 'getting there' deficit can be- In April 1973, a single Tomcat, flying a standard loiter at about Mach .67 and 25,000ft (i.e. a preplaced FORCAP orbit), with one AIM-54 Phoenix aboard (minimum weight and drag) was turned to intercept a BQM-34E which was _itself_ closing at Mach 1.5 and 50,000ft. Starting from initial detection at 132nm, the F-14 flew an additional 20nm (1 minute) to achieve firing parameters of Mach 1.5 and 44,000ft. The missile then flew for 2.62 minutes or 157 seconds. To achieve a downrange intercept at a mere 72.5nm. DURING THIS TIME, average /missile/ velocity, including the parent boost and a specially tailored (engineers spent all night tweaking the analogue autopilot gains, something which would never happen in the fleet) profile of 103,500ft altitude (as near zero-drag vacuum as you can get) was no more than 1,656 knots. Or roughly Mach 2.93.

What happens if, the target turns away? If the target is going slower? If the target performs a beaming or 'notch' maneuver? If the target postholes down into the clutter where the Hi-PRF looses it in the clutter? The missile misses that's what.

In another 'miracle mile' event, an F-14, firing at a BQM-34A with an initial setup of 10,000ft and Mach .72 vs 50ft and Mach .75 at 22nm separation showed the weapon intercepted at around 16nm in 54 seconds. DESPITE being a SARH-PDSTT all the way (no firing lag to account for time share TWS on multiple missiles) in nominally 'snap down' assisted conditions for acceleration, where there was no time spent climbing to the loft. DESPITE the fact that the missile was indeed /powered throughout the flight/. Average Missile Mach was only 1.88.

In the 'so impressive' 6x6 engagement, the aircraft _could not_ achieve 'maximum kinetic assist' because, with six Phoenix aboard, it didn't have the gusto to do more than about Mach 1.2 and could not achieve even this _in the time available_ to initiate TWS tracking on 'fighter sized' (3-5m augmented drones) targets. And the crew were pressed so hard already (in retaining adequate radar scan volume overlay) that, instead of a wall or conventional shelf, the UT-33 and BQM-34 were arranged in an 'extended card' type formation with azimuth spacings on the order of 5nm in frontage (three sets of two for 15nm) and with upwards of 35nm in trail (83 vs. 110nm) yet only 5,000ft of altitude separation. What this allowed the F-14 to do was generate a maximum X minimum scan of 120` X 2 bars in a giant pie-slice sectoring of sky that kept everybody visible with missiles in the air and with no AWS-27 (E-2) uplink.

Even so, at the outset of engagement, a completely bogus 'orchestration' of formation behaviors had to occur so that none of the targets drifted out of the scan volume or steamed right through it so that the initial drones flew at Mach .6 and the trailing aircraft at Mach .8 while the last was a 'sprinter' coming in at about Mach 1.2 to play catchup. The F-14 initiated
firing at a mere 31nm and continued to do so over a steady-flight period of 38 seconds, opening up on the closest targets /last/ (exposing itself to their weapons systems) to ensure that 'all missiles impacted within a few seconds of each other' for a virtual simultaneous seeming engagement. As a part of this exercise in idiocy, it 'maxxed the dot' (ASE starboard) so that
it could bias it's TWS volume into the target lane while setting the geometry physically to engage the final AQM-37 (sprint) target coming up the far right side of the engagement.

/Conveniently/, not only did the drones all arrive at co-pole distance with the missiles due to their careful range distribution, but they actually /curved inwards/ to follow the Tomcat (like a drunk crossing lanes into oncoming traffic) so as to better stay in-volume.

'And So', over a total period of 3.92 minutes (235 seconds, 33 miles at Mach .9, 55 miles at Mach 1.5, a /veritable eternity/ in fighter vs. fighter ops) the Tomcat killed all but the furthest-out (lefthand biased) 2 targets, thereby proving that multi-on-multi _did not_ work. Because even with all this grooming of the engagement variables, the AWG could not keep everybody
under track long enough to get a missile out to each of them, dumping one target completely before the AIM-54 could hand off. While the other drone had its FQ augmentation now so far out of field that the AIM-54 itself could not maintain the target track at the severe crossing angle.

Keep in mind that NONE of these were 'valid kills' because despite the nominally /enormous/ LAR or 'Launch Acceptability Region' of the Phoenix itself, the combination of scan lag and limited PRF ability to handle various low closure/high crossing angle targets through the Hi/Lo interleave ensured that TWS was unavailable until a point (roughly 50nm) at which the structured missile flyout sequencing necessary to get all six targets challenged the assumption of killing before being killed. And so, regardless of supposed simultaneity, the entire raid behavior was suspect, not only for being designed to bring the targets into the Phoenix envelope ONLY as the weapons came to bear. But rather for what it did NOT require the Tomcat crew to do so as to avoid threat bypass or direct engagement of the F-14 itself.

This is something which no halfway competent (threat) fighter pilot would 'step into' as he:

1. Doubled the altitude separation so as to force the F-14 RIO to compress his azimuth scan field to deepen the bar search.
2. Transited the combat area at a MINIMUM 550 knots or Mach .95 to compress the flyout vs. SARH timeshare problem even more.
3. Maximized his formation frontage densities to make sure the Tomcat had to open fire at closer to maximum (TWS interleave) of 50-60nm to have a hope of killing all targets in a very tight separation of missile guidance updates. For which sudden, drastic, formation changes would leave little or no ability to adjust final missile update steering into handoff conditions.

All of which leads to the generalized sarcasm of "Ignoring the Phoenix shots..." among those AF pilots in particular who sparred with the USN Tomcat community off Rota Spain during the late 70's and early 80's when the AIM-54/AWG-9 was at the height of its 'mystical' powers (achievement is the inverse of expectation, the Pentagon Paradox).

Indeed, the saddest element of this story is that no Tomcat has ever flown a Fleet Defense Bravo loadout of six missiles in active (cruise) service. They cannot safely recover or (single engine ROC) launch with that much weight. And further they cannot themselves maintain adequate smash to aggressively maneuver at altitude to set the geometry vs. fighter targets without going
supersonic which both eats fuel and instantly compresses the fight. Indeed, most of the squadrons did even not deploy (before 1988 anyway) with the outboard horn rails because they were draggy as hell and a pain to mount/dismount in trade for the more common Sparrow or even Sidewinder (6X2 or 4X4) alternate loadouts.

Lastly, the USN only produced about 5,000 AIM-54s and of those, only about half were the AIM-54C+ 'ECCM/Sealed' (either as new or by conversion) which had the seeker, warhead fuzing and autopilot upgrades to be any good against more than the dumbest threats. The /total/ number is only sufficient to allow every Tomcat a Fleet Defense Alpha loadout of 4 missiles. One time. And the actual magazine count during cruise was never more than a fraction of even this (back when there were actually two squadrons of 12 Tomcats on every deck).