TCAS/ACAS Origins and Guidance Material
From late 1981 to early 1982 an early version of TCAS II was installed in two Piedmont B727s, and in 1984 N857N was cleared for use by all Piedmont's B727 crews. I have a copy of their operations manual before me as I write, and later in that decade I both flew their TCAS-equipped B727 flight simulator and observed from within the flight deck how the crews used this equipment in line operations.
Just a word about TCAS and ACAS. The acronym 'TCAS' stands for 'Traffic Alert and Collision Avoidance System' and was used in a generic sense in the early years when its equipment status was MOPS 6, 6.02, 6.04 and 6.04A. Subsequently TCAS II with MOPS 7 and 7.01 installed has become a solution to the operational standards inherent in ACAS II. TCAS I would not be capable of posting RAs, TCAS II can post RAs that advise avoidance manoeuvres in the vertical plane, and TCAS III would post avoidance manoeuvres in the horizontal plane as well as in the vertical plane. (As a guest of the FAA I flew a B727 in an encounter trial near Atlantic City with a breadboard version of TCAS III installed - it worked just fine, but subsequently the addition of horizontal advisories was never considered necessary.)
My involvement with TCAS/ACAS came about because as a Flight Operations Inspector in the UK CAA's Flight Operations Department I was tasked to participate in the likely introduction of TCAS II as a safety modification to UK commercial air transport/public transport aeroplanes. In the course of our early meetings that included representatives of NATS, airlines and the Ministry of Defence (MOD) Procurement Executive, the latter explained that fast jet pilots would probably not want to have a Resolution Advisory (RA) facility but rather a means by which they could be told/shown where to look to see any potentially conflicting aircraft and then make their own decisions as to what manoeuvre would be appropriate. Within a few years they produced a trials version of such a device that I seem to recall was installed in a Tornado but I don't know if a general modification programme followed.
To go back to Retired BA/BY's comment that at the time of the Uberlingen accident in July 2002 there was no guidance as to how pilots should respond when an ATC instruction disagreed with a TCAS RA: this was not the case in Europe where explicit guidance existed. First, the UK CAA published Civil Air Publication 579 'Airborne collision avoidance systems (ACAS) guidance material' in March 1991 (before any UK air operator had installed this equipment or begun training their crews). This CAP included the text in paragraph 6.2.4 (b), "If pilots receive simultaneously an instruction to manoeuvre from ATC and an RA, and both conflict, the advice given by the RA should be followed'. As the main author of this CAP and also the UK member on the Joint Aviation Authorities' Flight Operations Committee (later, 'Sectorial Team') I inserted a similar text into the first version of Temporary Guidance Leaflet No 11 published in October 1998 thus, "If pilots simultaneously receive instructions to manoeuvre from ATC and an RA which are in conflict, the pilot should follow the RA". Finally, as the UK nominee to the ICAO Operations Panel (yes, I was quite busy for a while!) I was able to ensure that similar text was written into the guidance contained in PANS-OPS Volume I Chapter 3 'Operation of Airborne Collision Guidance System (ACAS) Equipment' published in November 2008, "In the event of an RA, pilots shall .... follow the RA even if there is a conflict between the RA and an air traffic control (ATC) instruction to manoeuvre". The rationale is that the ACAS refresh rate was some ten times faster than information displayed to air traffic controllers, and of course the information transfer was much faster within a flight deck than from ground to air.
The point here is that it is essential that all pilots who have ACAS equipment in the aircraft they are flying (General Aviation as well as Commercial Air Transport) must have first received training in accordance with the ICAO guidance - this should ensure that pilots throughout the world comply with the same rules. Both ICAO Annex 6 Part I (CAT) and Annex 6 Part II (GA) contain this Standard. The only exception I can envisage is military fast jet aeroplanes when not in transit.
One other point: it is really important that all aircraft that are required to be equipped with SSR use it with altitude reporting enabled, for this will allow them to be detected by ACAS-equipped aircraft and the risk of conflict mitigated accordingly. Without altitude-reporting enabled in one aircraft, pilots in the ACAS-equipped aircraft might have a Traffic Advisory (TA) posted even though the former is some 7,000ft below (or possibly above). This is a real distraction as the pilots with an ACAS display don't know whether the intruder is co-altitude, and hence a hazard, or well distant and thus not a concern.
Finally, ACAS doesn't care how a potential conflict has come about, whether through ATC error, pilot non-compliance with ATC instruction, altitude bust on climb or descent, emergency descent, incorrect setting of altimeter sub-scale or whatever. If the ACAS algorithms detect a breach by an intruder of their time-to-CPA (closest point of approach) they will post accordingly - and this is regardless of ATC separation standards.
I hope that this submission will help resolve a few issues!