Electronics Guide

Why Cooperative Identification Was Necessary

The problem announced itself as soon as radar entered service. British Chain Home stations could detect formations over the Channel at ranges that gave real warning, but the plot on the display was a blob of energy with no label. Returning bombers, friendly fighters climbing to intercept, and inbound raids all rendered the same way. The Royal Air Force improvised a procedure called "pip-squeak," in which a fighter's radio transmitter was keyed automatically for fourteen seconds in every minute so that ground direction-finding stations could fix its position and correlate that fix with a radar plot. Pip-squeak worked, but it consumed the radio channel, required ground stations dedicated to direction finding, and identified only aircraft that had been assigned to the scheme.

The deeper difficulty is that radar cross section is a poor identifier. It varies by orders of magnitude with aspect angle, so the same airframe can present a large return in one second and a small one in the next, and it is influenced by frequency, polarization, external stores, and the resonance of structures whose dimensions fall near the wavelength. Techniques do exist that extract identity-bearing features from a skin return, including jet engine modulation, in which rotating compressor blades impose a spectral line structure on the echo, and high-range-resolution profiling, which resolves the target into a sequence of scattering centers. Modern fighter radars carry them, but they classify rather than identify: they suggest an aircraft type, not the specific aircraft or its allegiance.

A transponder converts identification from an inference about scattered energy into a communication protocol with a defined message set, and the link budget improves dramatically in the process. A skin echo suffers the fourth-power range loss of the two-way radar equation, since the transmitted power spreads on the way out and the scattered power spreads on the way back. A beacon reply suffers only the square-law loss of a one-way link in each direction, which is why a modest interrogator with a few hundred watts obtains replies at ranges where the same antenna could not detect the aircraft's skin at all. Ground beacon interrogators routinely reach beyond two hundred nautical miles, limited by the radio horizon rather than by power.

Surveillance built on this principle is called dependent, because its output depends on equipment aboard the aircraft functioning correctly and honestly. Primary radar is independent: it needs nothing from the target. The two properties are complementary rather than ranked, and every serious surveillance architecture retains both. A dependent system tells the controller who is there; an independent system tells the controller that something is there whether or not it consents to be seen.

Wartime Origins: Mark I to Mark III

The first operational transponder identification system was IFF Mark I, fitted experimentally by the Royal Air Force in 1939. It used a regenerative receiver tuned across the Chain Home band in the region of twenty to thirty megahertz, sweeping continuously so that it would amplify and re-radiate whichever radar happened to illuminate it. The retransmitted signal added to the aircraft's own echo, causing the blip on the radar display to grow and shrink in a characteristic rhythm that an operator could learn to recognize. The design was elegant in that it required no separate interrogator, but it was fragile. The gain had to be set precisely: too little and the effect was invisible, too much and the set oscillated and swamped the radar.

Mark II, introduced in early 1940, replaced the manual gain adjustment with automatic gain control and added motorized tuning that stepped through several radar bands in turn. It was more reliable and much more widely fitted, but it inherited the central weakness of the concept: the transponder answered the surveillance radar itself, so every new radar band required a new tuner section, and the sweeping receiver spent most of its time listening on frequencies where nothing was transmitting.

The decisive change came with Mark III, which entered service during the war and became the standard system for the Western Allies. Mark III abandoned the idea of replying to the search radar and moved interrogation to a dedicated band of its own, well separated from the radars it served. A radar site that wanted identification transmitted a separate interrogation on the IFF band through its own antenna, and the transponder replied on that band alone. This is the architecture that has survived unchanged in principle to the present day, and its advantages are worth naming explicitly. A dedicated channel means the transponder receiver can be fixed-tuned and sensitive. It means the reply waveform can be designed for information content rather than forced to imitate a radar echo. It means one transponder serves every radar in the fleet, because compatibility is now a matter of agreeing on the interrogation format rather than on the radar's operating frequency. It also means the interrogation itself can carry a code, so that different questions may be asked and different answers expected.

Standardization on a common interrogation format across allied forces was as important as any circuit in the set. Identification is only as good as the agreement behind it: a transponder that answers questions nobody asks is inert, and an interrogator that asks questions nobody understands receives silence indistinguishable from hostility. That lesson — that the value of the system lies in the shared protocol rather than in the hardware — has governed every subsequent revision. The wartime development of radar itself, of which IFF was one strand, is treated in more detail elsewhere in this guide.

The Postwar Settlement: One Channel, Two Communities

After the war the system was rebuilt around a pair of frequencies that remain in use today: interrogations at 1030 megahertz and replies at 1090 megahertz. The system that established them was designated Mark X, and the letter is a reminder that the designation originally denoted an experimental series rather than a tenth iteration. Mark X introduced the Selective Identification Feature, which added coded pulse trains to what had been a simple presence indication, so that an aircraft could return a numeric identity rather than merely announcing that it was equipped.

Placing interrogation and reply twenty megahertz apart in the same band solves a real problem cheaply. A single-frequency system would require the transponder receiver to be blanked during its own transmission and the interrogator receiver during its own pulse, with all the timing complication that implies. A twenty-megahertz split allows a fixed diplexer and a modest filter to isolate the paths, while keeping both inside one antenna's bandwidth so that a single quarter-wave blade on the aircraft belly serves for transmit and receive alike.

Crucially, the civil aviation authorities adopted the same channel and much of the same signal structure. Civil secondary surveillance radar and military IFF are not parallel systems that happen to resemble each other; they are branches of one system, sharing frequencies, sharing pulse formats for the modes they hold in common, and sharing the airborne equipment in aircraft that need both. A military transport flying in civil airspace answers the civil interrogations of an air traffic control beacon interrogator using the same transponder that answers a fighter's military interrogation moments later. That sharing is the source of the system's greatest strength, universal interoperability, and of its most persistent weakness, a single congested channel that every user must contend for. The international technical provisions are published by the International Civil Aviation Organization in Annex 10 to the Convention on International Civil Aviation, whose fourth volume covers surveillance and collision avoidance systems, and it is that document, together with the corresponding RTCA and EUROCAE minimum operational performance standards, that fixes the numbers quoted below.

Interrogation: The Pulse Pair That Selects a Mode

The classical interrogation is disarmingly simple. The interrogator transmits two pulses, each 0.8 microseconds wide, through the directional antenna. These are named P1 and P3. The information is carried entirely in the interval between them: the transponder measures the spacing and uses it to decide which question has been asked and therefore which reply to compose. A third pulse, P2, plays a different role and is treated in the next section.

The mode assignments follow a simple ladder of spacings. A P1 to P3 interval of 3 microseconds selects Mode 1, a military mode. Five microseconds selects Mode 2, also military. Eight microseconds selects Mode 3/A, the mode shared between military Mode 3 and civil Mode A, which returns an identity code. Seventeen microseconds selects Mode B, twenty-one microseconds selects Mode C, which returns pressure altitude, and twenty-five microseconds selects Mode D. Modes B and D were allocated for regional or future use and are not employed in ordinary operation.

The economy of this scheme deserves comment. Mode selection requires nothing in the transponder but a time-interval discriminator and a small decoding network. No demodulator, no clock recovery, and no synchronization are needed, because the pulses are unmodulated bursts of carrier and the only measured quantity is their separation. In vacuum-tube and early transistor implementations that mattered enormously, and it is why a working transponder could be made small enough and reliable enough to fit in a fighter in the 1950s. The cost appeared only later: a protocol with no addressing has no way to speak to one aircraft rather than all of them.

An interrogator normally alternates modes on successive interrogations, most often interleaving Mode A and Mode C so that each antenna scan produces both an identity and an altitude for every aircraft in the beam. The interrogation rate is constrained on both ends: high enough that the beam sweeping past an aircraft elicits enough replies to establish presence and azimuth, low enough that the shared channel is not saturated and that transponders, which are rate-limited by design, are not driven into their protective duty-cycle limits.

Side-Lobe Suppression and the Need for an Omnidirectional Control Pulse

Every practical antenna radiates outside its main beam. A secondary radar antenna with a beamwidth of two or three degrees will still put appreciable energy tens of degrees off boresight through its side lobes, and because the beacon link is so much stronger than a skin echo, a transponder illuminated by a side lobe can easily receive enough power to trigger a reply. The consequence on the display is a ring of false targets at the correct range but at wildly wrong azimuths, historically called ring-around. The strength of the beacon link, which is the system's great advantage in range, is precisely what makes this failure mode severe.

The remedy is interrogation side-lobe suppression, and its mechanism is worth studying because it is a clean example of solving an antenna-pattern problem in the protocol rather than in the antenna. The interrogator transmits a third pulse, P2, exactly 2 microseconds after P1, but radiates it from a separate omnidirectional or broad-pattern control antenna whose gain is set to exceed the side lobes of the directional antenna while remaining well below its main-beam gain. The transponder compares the received amplitude of P2 with that of P1. If P1 is stronger, the transponder concludes that it is being illuminated by the main beam and processes the interrogation normally. If P2 is equal to or stronger than P1, the transponder concludes that it heard P1 through a side lobe, discards the interrogation, and suppresses itself for a period specified as 35 microseconds, with a tolerance of plus or minus 10 microseconds.

Several design details follow. The P2 position at 2 microseconds is deliberately shorter than the shortest mode spacing of 3 microseconds, so the suppression pulse can never be mistaken for a P3 and cannot itself select a mode. The suppression interval covers the remainder of a side-lobe interrogation, including the longest mode spacing, yet ends well before the next interrogation from another station is likely to arrive. And because the comparison is one of relative amplitude within a single interrogation, it is insensitive to range: an aircraft at ten miles and one at a hundred apply the same test with the same result, which no scheme based on an absolute power threshold could do.

A complementary mechanism, receiver side-lobe suppression, operates at the ground station. Here the interrogator's receiver compares the amplitude of an incoming reply on the directional channel with the amplitude on the control channel and discards replies that appear stronger on the control channel, since those must have entered through a side lobe of the receiving pattern. Interrogation side-lobe suppression prevents the unwanted reply from being generated at all, which conserves channel capacity; receiver side-lobe suppression catches what remains. Practical systems use both.

The Reply Train: Identity in Octal, Altitude in Gillham Code

A reply is a pulse train on 1090 megahertz whose structure is fixed regardless of mode. Two framing pulses, F1 and F2, bracket the message and are spaced 20.3 microseconds apart. Between and including them the interval is divided into 15 time slots, each 1.45 microseconds wide, and a data bit is asserted by placing a 0.45-microsecond pulse in the corresponding slot. The slot sequence, in transmission order, is F1, C1, A1, C2, A2, C4, A4, X, B1, D1, B2, D2, B4, D4, F2, with a further slot for the Special Position Identification pulse positioned 4.35 microseconds after F2.

The interleaving of the letter groups is intentional. Rather than sending all four A bits together, then all four B bits, the format alternates among the groups so that a burst of interference or a partial overlap with another aircraft's reply damages one bit of several digits rather than obliterating one digit completely. The slot labeled X is not used in normal civil operation and is reserved. The framing pulses serve as the reply's own synchronization: a ground receiver searching for replies looks for two pulses 20.3 microseconds apart and only then attempts to read the slots between them.

Mode A: A Four-Digit Octal Identity

In Mode A the twelve information pulses are read as four octal digits, A, B, C, and D, each formed from its 1, 2, and 4 weighted bits. Four octal digits give 4,096 distinct codes, the familiar squawk that a controller assigns by radio and a pilot dials into the transponder's four thumbwheels or keypad. Certain codes are reserved worldwide and are recognized by every ground system: 7500 indicates unlawful interference, 7600 indicates radio communication failure, and 7700 indicates a general emergency. Conspicuity codes for aircraft operating under visual flight rules are assigned regionally, 1200 in the United States and 7000 in much of Europe among them.

The Special Position Identification pulse is the mechanism behind the controller's instruction to "squawk ident." Pressing the ident button on the transponder appends the SPI pulse to every reply for a period of some seconds, causing the aircraft's symbol on the display to blossom or blink so that the controller can pick it out of a crowded picture. It is the one piece of the protocol whose purpose is purely human.

Four thousand and ninety-six codes sounded generous in 1955 and became a scarce resource. Because codes must be assigned so that no two aircraft within view of the same set of interrogators share one, and because blocks are reserved by state, by function, and by facility, code shortage became a genuine constraint on traffic growth in busy regions long before the channel itself saturated. This scarcity is one of the principal reasons the system eventually needed a unique-address successor.

Mode C: Pressure Altitude in a Gray Code

In Mode C the same twelve slots carry pressure altitude referenced to the standard datum of 1013.25 hectopascals, in increments of 100 feet, encoded in the Gillham code. Gillham is a modified reflected binary code, a member of the Gray code family, and the reason for choosing it over ordinary binary is mechanical rather than mathematical. Early encoders were optical or brush-contact discs geared to an aneroid capsule, and in a Gray code exactly one bit changes between adjacent values. A misaligned brush at a transition therefore produces an error of one increment, not the catastrophic multi-bit error that a binary encoder can produce when several tracks change at once. The D1 slot is not used in Mode C, so eleven pulses carry the altitude.

Reporting to 100 feet was adequate when vertical separation minima were a thousand feet, but it is coarse for any use that compares two aircraft's altitudes closely, and it is unsuitable for the reduced vertical separation minimum of a thousand feet above flight level 290 in the way that modern automation would prefer. Mode S therefore introduced a 25-foot altitude report, carried in a different field, and modern systems prefer it wherever the transponder supports it. Gillham encoding survives in the interface between older altitude encoders and transponders, and the conversion from Gillham to a plain integer remains a small but real piece of avionics firmware.

FRUIT and Garbling: The Price of a Shared Channel

Two pathologies dominate the classical system, and both follow inevitably from the decision to let every interrogator ask every transponder the same unaddressed question on one pair of frequencies.

The first is FRUIT, an acronym expanded as False Replies Unsynchronized In Time, or equivalently False Replies Unsynchronized with Interrogator Transmissions. An aircraft replies to every valid interrogation it receives, from whatever station. A ground interrogator therefore hears not only the replies it elicited but also replies that other stations elicited from the same and from other aircraft, arriving at times bearing no relation to its own transmissions. In dense airspace with many overlapping interrogators, FRUIT can outnumber the wanted replies substantially.

The defense is called defruiting, and it exploits the one property the wanted replies possess and the unwanted ones do not: correlation with the interrogator's own timing. A defruiter stores replies from successive interrogations and passes only those that recur at a consistent range on successive interrogations of the same target. A genuine reply appears at nearly the same delay scan after scan, because the aircraft moves only slightly between interrogations. FRUIT, arriving at random delays, fails the test. The method is simple and effective, but it costs something real: it requires several interrogations of each aircraft before a target can be declared, which raises the interrogation rate and therefore contributes to the very congestion it is defending against.

The second pathology is garbling, and it is more insidious because the corrupted reply looks like a valid reply. Two aircraft at similar azimuth whose ranges differ by less than the length of a reply train will have their replies overlap at the receiver. Since the reply occupies just over twenty microseconds, and a microsecond of round-trip delay corresponds to about one hundred and fifty meters of range, aircraft within roughly 1.7 nautical miles in slant range on the same bearing produce overlapping pulse trains. The receiver may then read a valid-looking but entirely wrong code composed of pulses from both aircraft, or may fail to find framing pulses at all. Synchronous garbling is the term for the case where the overlap persists from interrogation to interrogation because the two aircraft hold formation, in which case no amount of correlation processing separates them.

Degarbling algorithms attempt to resolve overlapping trains by finding two consistent sets of framing pulses and assigning intermediate pulses to one target or the other, and the interleaved bit ordering described earlier helps. But the problem is fundamentally one of insufficient information: when two independent transmitters emit unaddressed messages into the same time interval, no receiver can always separate them. Formation flight, closely spaced parallel approaches, and holding stacks are precisely the situations where controllers most need reliable identity, and precisely the situations classical beacon surveillance handles worst.

Underlying both problems is a capacity argument. The number of replies on 1090 megahertz grows with the number of aircraft multiplied by the number of interrogators that can see them, because every interrogator asks every aircraft. That product grows faster than traffic alone, and in the busiest regions it grew fast enough to threaten the channel. Any fix had to break the multiplication, which meant giving interrogators a way to ask one aircraft a question without provoking a reply from all the others.

Mode S: Selective Addressing

Mode S, where the letter stands for select, is the answer to all of the preceding problems and, not coincidentally, the foundation of everything built on the beacon channel since. Its central idea is a unique address. Every Mode S transponder is assigned a 24-bit aircraft address, allocated in blocks to states by the International Civil Aviation Organization and thence to individual aircraft, giving a space of 16,777,216 values. The address is a property of the aircraft, not of the flight, and it does not change when the flight number or the Mode A code changes. With an address available, an interrogator can direct a question to one aircraft and expect exactly one reply.

Mode S was designed to coexist with the installed base rather than replace it, and the backward compatibility mechanisms are ingenious. A Mode S interrogation begins with P1 and P2 spaced 2 microseconds apart, which is exactly the pattern a classical transponder interprets as a side-lobe suppression pair; the older equipment therefore suppresses itself and remains silent for the duration of the Mode S transmission that follows. That transmission is a long pulse, P6, carrying data by differential phase-shift keying at four megabits per second, preceded by a synchronizing phase reversal that establishes the chip timing. Two lengths exist, a short interrogation carrying 56 data bits and a long one carrying 112.

In the other direction, an all-call interrogation intended to elicit replies from aircraft whose addresses are not yet known uses the classical P1 and P3 pair followed by a fourth pulse, P4. The width of P4 selects the audience. A short P4 of 0.8 microseconds is the Mode A/C-only all-call: classical transponders reply normally and Mode S transponders recognize the short P4 and stay silent. A long P4 of 1.6 microseconds is the Mode A/C/S all-call, to which Mode S transponders reply with their address. A separate control pulse, P5, transmitted from the omnidirectional antenna and overlapping the synchronizing phase reversal of a Mode S all-call, provides side-lobe suppression for that interrogation by destroying the transponder's ability to recover the sync phase reversal when the aircraft lies outside the main beam.

The reply waveform is entirely different from the classical train. It opens with a preamble of four pulses occupying eight microseconds, whose spacing is chosen so that it cannot be produced by any legal Mode A or Mode C reply, and follows with a data block encoded by pulse-position modulation at one megabit per second: a one is a half-microsecond pulse in the first half of the bit interval, a zero a half-microsecond pulse in the second half. Every bit therefore contains exactly one pulse, which gives the receiver a transition in every bit period for timing recovery and makes the format robust against the amplitude variation that plagues a channel shared with much stronger nearby transmitters. Replies come in the same two lengths as interrogations, 56 and 112 bits. The first five bits of any message name its downlink format, DF, or uplink format, UF, and thereby its interpretation.

All-Call and Roll-Call

A Mode S interrogator runs a two-phase cycle on each scan. In the acquisition phase it transmits all-call interrogations to discover aircraft entering coverage; a Mode S transponder that is not already locked out replies with a DF11 all-call reply containing its 24-bit address. Having acquired the address, the interrogator moves that aircraft to the roll-call phase, in which it addresses the aircraft individually with selective interrogations scheduled into the dwell time available as the beam sweeps past. Because each selective interrogation produces exactly one reply from exactly one aircraft, garbling between Mode S targets is eliminated, and because the interrogator knows in advance which replies to expect and when, defruiting becomes nearly trivial.

Lockout completes the mechanism. Having acquired an aircraft, the interrogator commands its transponder to stop answering all-call interrogations from that interrogator, so the aircraft does not clutter the acquisition phase on every subsequent scan. Lockout must be per-interrogator rather than global, or an aircraft acquired by one station would become invisible to its neighbors. Mode S therefore assigns each interrogator an identifier: a four-bit interrogator identifier, II, giving fifteen usable values, and later a six-bit surveillance identifier, SI, extending the space to sixty-three where the density of ground stations demanded it. Lockout state is maintained separately for each identifier, and it times out if the interrogator stops refreshing it, so an aircraft flying out of one station's coverage becomes acquirable again automatically.

Error Control by Address Overlay

Mode S protects every message with a 24-bit cyclic redundancy check, and it does something clever with the result: rather than transmitting the parity separately from the address, it exclusive-ORs the two together into a single 24-bit field at the end of the message. In a reply, the parity is overlaid on the address of the replying aircraft. In an interrogation, the parity is overlaid on the address of the intended recipient.

This overlay buys two properties at once for the price of one field. First, a transponder decoding an interrogation computes the parity of the received message, exclusive-ORs it with the received parity-address field, and compares the result with its own address: if they match, the message is both error-free and addressed to it, and the two tests have cost one operation. A transponder that is not the intended recipient sees a mismatch and simply discards the message, requiring no separate address comparison. Second, a ground receiver decoding a reply recovers the sender's address from the same computation, so the address need not occupy any additional bits in a short 56-bit message where every bit is scarce. The generator polynomial is chosen so that the code detects all burst errors up to 24 bits, and in favorable circumstances a receiver can even correct short bursts, which matters on a channel where the dominant impairment is a colliding transmission of finite duration rather than white noise.

Mode S Data Link: Comm-A Through Comm-D

Once a protocol has addressing and error detection, it is a communication link, and the designers of Mode S exploited that. Beyond the basic surveillance formats, which return altitude or identity in reply to a short interrogation, Mode S defines message fields that carry general-purpose data in both directions.

Comm-A is a long uplink interrogation carrying a 56-bit message field from the ground to the aircraft, combined with a surveillance request so that the same transmission also elicits an altitude or identity reply. Comm-B is the corresponding downlink: a long reply carrying a 56-bit message field from the aircraft, again combined with surveillance data. Comm-B transfers may be initiated by the ground, which requests a specific data register, or by the aircraft, which announces that it has data waiting and is then polled for it.

For payloads too large for a single 56-bit field, Mode S defines the Extended Length Message service. Comm-C is the uplink ELM, in which the ground transmits a sequence of message segments and the aircraft acknowledges the set as a whole rather than segment by segment; Comm-D is the downlink ELM, in which the aircraft transmits a sequence of segments in response to a single ground request. Batching the acknowledgment is what makes the service efficient: a link with a per-message round trip would waste most of a scan's dwell time waiting.

The data carried in Comm-B are organized into numbered registers, referred to as BDS registers and written as two hexadecimal digits separated by a comma. Two service levels are defined for air traffic surveillance use. Elementary Surveillance, ELS, requires the aircraft to supply its own identification — the flight identity as filed, carried in register BDS 2,0 — along with transponder capability reporting, altitude in 25-foot increments, flight status, and the ability to operate with surveillance identifier codes. The gain is immediate: with the flight identity downlinked directly, a controller no longer needs a Mode A code to correlate a radar track with a flight plan, which relieves the code-shortage problem that had constrained busy airspace for decades.

Enhanced Surveillance, EHS, adds a set of downlinked aircraft parameters drawn from the flight management and air data systems. The principal registers are BDS 4,0, the selected vertical intention, which reports the altitude selected in the mode control panel; BDS 5,0, the track and turn report, which carries roll angle, true track angle, ground speed, track angle rate, and true airspeed; and BDS 6,0, the heading and speed report, which carries magnetic heading, indicated airspeed, Mach number, and vertical rate. The operational value is that a controller can see what the aircraft has been told to do, not merely what it is doing. A discrepancy between the selected altitude in BDS 4,0 and the cleared altitude is a level bust in the making, and detecting it before it happens is worth a great deal. In European airspace, carriage of ELS and EHS capability is mandated for appropriate categories of aircraft under the European Union's surveillance performance and interoperability rules, which is why the great majority of transport aircraft operating there report these parameters.

Mode S also underpins airborne collision avoidance. TCAS II interrogates other aircraft on 1030 megahertz and listens on 1090, addressing Mode S intruders selectively and falling back to Mode C interrogation otherwise, and it coordinates its resolution advisories with other equipped aircraft over the Mode S air-to-air link so that two aircraft in conflict never both choose to climb. Collision avoidance is therefore not a separate radio system but an application layered on the identification channel.

Monopulse Secondary Surveillance Radar

Classical beacon interrogators estimated azimuth by the sliding-window method, sometimes called the centroid or beam-splitting method. As the antenna rotates, an aircraft begins replying when the leading edge of the beam reaches it and stops when the trailing edge passes. The processor records the azimuth at which replies started and the azimuth at which they stopped, and reports the midpoint. The accuracy of that estimate depends entirely on how many replies fall inside the window and on how consistently they are received, so the method requires the interrogator to transmit twenty or more interrogations per beam dwell simply to define the edges. Every one of those interrogations adds to the loading of the shared channel, and any reply lost to garbling or FRUIT biases the estimate.

Monopulse secondary surveillance radar replaces this with an angle measurement made on a single reply. The antenna is fed to produce two simultaneous patterns: a sum pattern with a maximum on boresight, and a difference pattern with a sharp null on boresight and opposite phase on either side. The ratio of difference to sum, evaluated with the sign taken from their relative phase, is a monotonic function of the off-boresight angle across the main beam. A calibration curve converts the measured ratio into an off-boresight angle, and adding that to the antenna's instantaneous pointing angle yields the target's azimuth from one reply.

The consequences are substantial. Azimuth accuracy improves by roughly an order of magnitude over sliding-window extraction, sharpening the track and reducing the jitter that forces controllers to apply larger separation buffers. The number of interrogations per scan falls from twenty or more to a handful, since angle no longer requires a population of replies, and that relieves channel congestion for every other user of the band. Monopulse also tolerates garbling far better, because the angle is computed from the received amplitudes and phases of one reply rather than from the pattern of which interrogations produced replies.

Monopulse also happens to be a prerequisite for Mode S operation. A Mode S interrogator addresses each aircraft only a few times per scan, so it cannot afford an angle-estimation method that needs twenty replies. The two technologies arrived together for that reason, and virtually every modern secondary radar is a monopulse Mode S installation, usually designated MSSR, frequently mounted on the same pedestal as a primary radar so that skin returns and beacon replies share an axis and a rotation.

ADS-B: Broadcast Without Interrogation

Automatic Dependent Surveillance — Broadcast removes the question. Rather than waiting to be interrogated, an aircraft periodically broadcasts its own state: position and altitude derived from a global navigation satellite system receiver, velocity, identity, and a set of quality indicators describing how much the receiving system should trust the position. It is automatic because no crew action initiates it, dependent because it relies on airborne equipment, and broadcast because it is addressed to no one in particular.

The dominant physical layer is the 1090 megahertz extended squitter, and its name explains its lineage. A squitter is an unsolicited Mode S transmission; Mode S transponders already emitted a short acquisition squitter periodically so that airborne collision avoidance systems could find them without interrogating. The extended squitter simply uses the long 112-bit Mode S downlink format, DF17, to carry a 56-bit payload of ADS-B data on the same schedule. Because it is a Mode S reply in every structural respect, the existing 1090 megahertz receiver at a ground station or aboard another aircraft decodes it with the machinery already present. Position and velocity messages are broadcast at a rate of roughly one to two per second, identification less often, and the transmissions are deliberately randomized in time so that two aircraft do not settle into a persistent collision.

Position is encoded by compact position reporting, a scheme that exists to fit a globally unambiguous latitude and longitude into a message field far too small for the full values. CPR transmits position relative to a grid of zones, alternating between two slightly different grids designated even and odd, so that a receiver holding one message of each kind resolves the ambiguity and recovers the absolute position; a receiver that already knows roughly where the aircraft is can decode a single message locally. The cost is subtle. A decoder that pairs messages from different times, or applies local decoding against a stale reference, produces a position wrong by a whole zone rather than slightly wrong, so correct implementations apply consistency checks that reject such results.

The United States adopted a second link for general aviation, the Universal Access Transceiver on 978 megahertz. UAT was designed from a clean sheet rather than layered on Mode S: it uses continuous-phase frequency-shift keying at approximately one megabit per second with Reed-Solomon forward error correction, and it structures each second into a segment for aircraft broadcasts and a segment for ground uplinks. That uplink capacity is what allows UAT to carry the Flight Information Service — Broadcast weather and aeronautical products, and Traffic Information Service — Broadcast, which relays the positions of aircraft that ground surveillance sees but that a UAT receiver would not otherwise hear. The trade is fragmentation: an aircraft on 978 megahertz and one on 1090 cannot see each other directly, and only the ground rebroadcast services close the loop between them.

The infrastructure consequences of removing interrogation are the deepest part of the change. An ADS-B ground station is a receiver, an antenna, and a data link. It has no transmitter, no rotating antenna, no high-power amplifier, and no mechanical maintenance burden of consequence, so it can be installed where a radar could never be justified — offshore platforms, mountain ridges, remote regions — at a small fraction of the capital and operating cost. Surveillance coverage becomes, for the first time, something that can be extended cheaply, and the update rate improves from one report every several seconds as an antenna rotates to roughly one per second.

The limitations are equally structural and should be stated plainly. ADS-B messages are neither encrypted nor authenticated. Any receiver can decode them, which is why public flight-tracking web sites exist, and, more seriously, any transmitter capable of producing a correctly formatted DF17 can inject a message that a receiver has no cryptographic means of rejecting. Security researchers demonstrated this openly beginning in 2012, and the vulnerability is inherent in the standard rather than in any implementation. Mitigations in operational use are indirect: correlating ADS-B tracks against independent surveillance such as primary radar or multilateration, checking kinematic plausibility, and verifying that a claimed position is consistent with the direction and timing from which the signal actually arrived at the ground station. Integrity is also dependent on the navigation source. The position is only as good as the GNSS solution behind it, which is why the message set includes the navigation integrity category, navigation accuracy category, and source integrity level fields, and why an aircraft whose receiver loses integrity monitoring must degrade or cease its reports rather than continue broadcasting a position of unknown quality. Finally, 1090ES adds traffic to the same congested channel that Mode S was designed to relieve, and channel loading in the densest airspace remains an active concern.

The deployment of ADS-B ground station networks and the wider air traffic management modernization programmes that depend on them are covered separately in this guide; the concern here is the signal and its relationship to the interrogation protocol from which it descends.

Multilateration: The Passive Complement

Multilateration turns the beacon channel's greatest nuisance into an asset. Every transponder reply and every squitter is a transmission from a known aircraft at an unknown position. If several receivers at surveyed locations record the time of arrival of the same transmission against a common clock, then each pair of receivers yields a time difference, and each time difference constrains the emitter to a hyperboloid. The intersection of enough hyperboloids fixes the position, in two dimensions with three receivers and in three dimensions with four, and additional receivers over-determine the solution and permit error estimation. This is the same time-difference-of-arrival principle that underlay the LORAN navigation system, applied in the reverse direction.

The system is passive in the sense that matters most: it adds nothing to the 1090 megahertz channel, because it listens to transmissions that would occur anyway. Some installations add a low-power interrogator to elicit replies from aircraft that would otherwise be silent, but the surveillance function itself is a listening one.

Three properties make multilateration valuable alongside the systems already described. First, it works with any transponder, including a plain Mode A/C unit with no Mode S and no ADS-B capability, so it extends modern surveillance quality to the oldest equipment in the fleet. Second, its position estimate is independent of anything the aircraft asserts: a transponder can broadcast a false ADS-B position, but it cannot alter the propagation time of its own transmission to four surveyed receivers. Third, its accuracy is set by receiver geometry and timing precision rather than by antenna beamwidth, so a well-sited network achieves accuracy on the airport surface that no rotating radar can match — which is why surface movement multilateration became the standard method of tracking aircraft and vehicles on taxiways and around terminal buildings.

The costs are geometry and timing. Position accuracy degrades sharply when the receivers are poorly distributed relative to the target, a condition analogous to poor dilution of precision in satellite navigation, so wide-area deployments require careful site planning. Timing must be held to nanoseconds across the network, which demands either GNSS-disciplined clocks at every site or a reference transmitter at a surveyed location whose known position lets the network calibrate itself continuously.

Military IFF: Mark XII and the Modes That Matter

The military system in current use is Mark XII, which operates on the same 1030 and 1090 megahertz channel as the civil system and supports the civil modes alongside its own. This dual capability is not a convenience; it is a requirement, because military aircraft operate in civil airspace routinely and because civil traffic must be identifiable to military air defense.

Mode 1 returns a short octal code used to indicate mission type or aircraft category, a coarse classification useful for sorting a raid picture. Mode 2 returns a four-digit octal code, giving 4,096 values, conventionally assigned to a specific unit or airframe and used for individual identification within a force. Mode 3/A is literally the same mode as civil Mode A, which is why the designation carries both names, and Mode C is the same pressure-altitude report used by air traffic control. Modes 1 and 2 are transmitted in the clear, exactly as the civil modes are, which means they can be interrogated by anyone with a compatible interrogator and can be exploited by an adversary's electronic support measures.

Mode 4 was the first cryptographic mode, added with Mark XII. Its structure is a challenge and response: the interrogator transmits an encrypted challenge derived from a code that changes daily, the transponder decrypts it with the key loaded into its cryptographic computer, and replies only if the result matches the expected value. The reply itself is a short three-pulse train, and the delay before it is transmitted is derived from the challenge rather than fixed, which denies an adversary the ability to use a known reply delay to fix the transponder's range. Mode 4 was a substantial advance over an unencrypted code, but it is a design of its era. The keys must be physically loaded and periodically changed, the reply carries no information beyond the assertion of validity, and the challenge-response exchange consumes channel time.

Mode 5 is the modern replacement, and the programme that fields it is designated Mark XIIA. Mode 5 retains the frequency pair but replaces the waveform with a spread-spectrum, cryptographically secured format whose interrogations and replies are keyed to a shared time reference, so that a recorded exchange cannot be replayed later and an interception yields nothing reusable. Two service levels are defined: Level 1 provides the authenticated identification that Mode 4 provided, with far better resistance to exploitation and jamming, while Level 2 adds reporting of the platform's own position and other data, which is to say a secure military counterpart of the information ADS-B broadcasts in the clear. A broadcast form allows a platform to announce itself without being interrogated, which reduces the emissions that interrogation would otherwise require. NATO adoption proceeded through the 2010s under the alliance's identification standardization agreements, with fielding of Mode 5 capable interrogators and transponders continuing across member forces.

The engineering constraint shaping all of this is retrofit. An identification system is worthless unless nearly every platform carries it, and forces field aircraft, ships, and ground-based air defense systems with service lives measured in decades. Mode 5 therefore had to occupy the same frequencies, work through the same antennas, and interoperate with Mode 4 and the civil modes throughout a transition lasting many years. That is why Mark XIIA is an upgrade programme rather than a new system, and why the civil modes survive in it unchanged.

Combat Identification and Its Residual Difficulty

It is tempting to read the progression from Mark I to Mode 5 as a problem steadily being solved. The operational record does not support that reading, and the reason is a logical asymmetry at the heart of the concept.

A cryptographically valid reply is strong evidence of friendship. Silence is not evidence of hostility. A friendly aircraft may fail to reply because its transponder has failed, because its antenna is shadowed by the airframe at that geometry, because it is set to the wrong code for the tactical area, because its cryptographic keys have expired or were loaded incorrectly, because it is operating under emissions control, or because the interrogation never reached it. The system is therefore asymmetric by construction: it can positively identify friends, and it cannot identify foes at all. Everything that fails to answer falls into a residual category that includes both hostile aircraft and friendly aircraft having a bad day.

The shootdown of two United States Army UH-60 Black Hawk helicopters by two United States Air Force F-15C fighters over northern Iraq on 14 April 1994 illustrates the failure directly. All twenty-six people aboard the helicopters were killed. The accident investigation found that the helicopters' response to a Mode 1 interrogation was negative, because they were squawking a code appropriate to a different area, and that Mode 4 gave a momentary positive response and thereafter responded negatively for reasons the board could not conclusively determine. The helicopters were squawking correct Mode 2 codes throughout. The board's conclusion — that the transponders on the F-15s, the Black Hawks, or both did not operate correctly for unknown reasons — is the most instructive sentence in the report, because it names the residual failure mode exactly: a system whose negative answer is ambiguous will eventually produce a negative answer at the worst possible moment.

The destruction of Iran Air Flight 655 by the USS Vincennes on 3 July 1988, in which all 290 people aboard the Airbus A300 died, shows a different pathology. The aircraft transmitted civil identification codes throughout its flight. An identification supervisor aboard the ship interrogated the contact without resetting the range on his equipment and received a brief Mode 2 response, characteristic of a military aircraft, most plausibly from an unrelated aircraft on the ground at Bandar Abbas. That erroneous reading, once attached to the track, propagated through the tactical picture and outweighed the correct civil identification that was continuously present. The lesson is not about the transponder at all. It is about correlation: an identification is a claim about a particular target at a particular range and bearing, and a system that lets an operator associate a reply with the wrong track has failed no matter how correctly the radio link performed.

Modern combat identification therefore treats cooperative identification as one input among several rather than as an oracle. Non-cooperative recognition from the radar return, situational data from surveillance and battle management networks, the track's history and its conformance with known flight corridors, and positional reporting from friendly forces over data links such as Link 16 all contribute. Doctrine reinforces the technology with procedural measures — corridors, altitude blocks, timing windows — whose purpose is to make friendly aircraft predictable so that identification does not rest solely on a radio exchange. None of this eliminates the asymmetry. It reduces the number of situations in which a commander must decide about an unidentified contact under time pressure, which is a real gain, and it is honest to describe it as no more than that.

Conclusion

Cooperative identification exists because a reflection carries no name. Everything that follows — the 1030 and 1090 megahertz pair, the pulse spacings that select a mode, the framing pulses that bracket an octal identity, the omnidirectional control pulse that rejects side-lobe interrogations — is engineering built on that one observation, and much of it is remarkably economical. The classical format required nothing more elaborate in the aircraft than a time-interval discriminator and a pulse generator, and on that foundation the world built the surveillance system that made dense controlled airspace possible.

Its weaknesses were the direct consequence of its simplicity. An unaddressed protocol on a shared channel produces FRUIT because every interrogator hears every reply, and garbling because two aircraft close in range cannot be separated in time. Both problems grow with the product of aircraft and interrogators, which is why they became critical before traffic alone would have suggested. Mode S broke that product by introducing a unique 24-bit address, and in doing so it converted a beacon into a data link. The overlay of parity on address is the small piece of design that best captures the system's character: one 24-bit field performs error detection and addressing simultaneously, because the engineers had no bits to spare.

The line continues outward. Monopulse angle measurement made a single reply sufficient to fix azimuth, which is what allowed Mode S to interrogate sparingly. ADS-B removed the interrogation entirely and with it most of the cost of a surveillance site, at the price of depending on satellite navigation and broadcasting an assertion no one can verify cryptographically. Multilateration listens to the same transmissions and derives position from propagation time alone, which is exactly the independent check a broadcast system needs. Each layer solves the previous layer's principal problem and introduces one of its own — the ordinary shape of progress in a system that must never stop working while it is improved.

What has not changed since 1939 is the asymmetry. The system answers the question "is this a friend?" and it answers it well when the equipment works, the codes are current, and the reply is correlated with the right track. It has never answered the question "is this a foe?", and no version of it will, because silence has too many causes. Engineers building identification systems should design for that fact rather than around it, and the fratricide record is the reason to state it plainly rather than to let a positive identification rate stand in for a capability the system does not have.

Related Topics