Most fieldbuses solve the problem of moving a lot of data between intelligent devices. AS-Interface solves a different one: getting a few bits to and from hundreds of dumb devices, cheaply, without a terminal strip the size of a door.
A proximity switch has one bit of information. A solenoid valve needs one bit. Wiring each one back to an I/O card means two conductors per device, marshalling terminals, a numbered drawing, and an hour of someone’s time. AS-i replaces all of that with a flat yellow cable that a device clamps onto wherever it happens to be.
This guide covers how AS-i actually works — the signalling, the addressing, the fixed cycle time — and where it genuinely beats a fieldbus. Verified against IEC 62026-2, the international standard for AS-Interface.
Table of Contents
The core idea: data and power on two wires
AS-i is a master-slave system with a single master and up to 31 slaves — or 62 using extended addressing. The master sends data and parameters to a specific slave; the slave passes data to its output ports or processes the request, then returns its input data or the result.
What makes it distinctive is the cable. The same unshielded two-conductor flat cable carries both the communication signal and the power for the connected devices. No separate supply run, no shielded twisted pair, no termination resistors.
The standard fixes the cable colour: the AS-i standard cable shall be coloured yellow, with the conductors clearly identified — brown for +, blue for −. The yellow is specified against RAL 1012. That is why every AS-i installation looks the same, and why the cable is instantly recognizable in a cabinet photo.
Slaves connect through insulation-displacement contacts: the module clamps over the cable, piercing prongs make contact with both conductors, and the device is on the network. No stripping, no terminals, no polarity mistakes — the cable’s asymmetric profile only fits one way.
Addressing: 31, or 62 the hard way
Each slave has an address, and the address space tells you a lot about AS-i’s history.
Standard addressing gives addresses 1 to 31. Address 0 is reserved for unaddressed devices — a new slave out of the box sits at 0 and waits to be assigned.
Extended addressing doubles the count to 62 by splitting each address into an A and a B sub-address, so addresses run 1A/1B through 31A/31B. The standard describes this as doubling the number of slaves from 31 to 62 by introducing sub-addresses.
The catch is in a quiet note in the standard: in extended addressing mode, two cycles are needed to transfer data for all slaves in that mode. You get twice the devices at half the update rate for those devices. That is the whole trade, and it is worth knowing before someone specifies 62 slaves on a machine with a timing requirement.
Each slave carries an I/O configuration and an identification code, which together tell the master what kind of device it is. The master compares detected devices against its projected list — the standard defines a list of projected slaves, a list of detected slaves, a list of active slaves, and a list of slaves signalling a peripheral fault. Those four lists are the whole diagnostic model, and reading them is how you troubleshoot an AS-i segment.
The 5 ms cycle
AS-i’s headline number is its cycle time, and unusually for a fieldbus, it is essentially fixed rather than something you tune.
The master polls each slave in turn. Every transaction is a fixed sequence: master request, master pause, slave response, slave pause, send pause. The standard specifies the timings tightly — the slave pause is 1.5 to 2 bit times, and the slave must accept a new master request after 6 µs.
With all 31 slaves active, a full cycle completes in about 5 ms. The standard states the constraint directly: where there are 30 or fewer transactions per cycle the send pause may be extended, but in that case the AS-i cycle time shall not be longer than 5 ms, including management and inclusion phases.
Two things follow.
The cycle time is deterministic and known in advance. You do not calculate it from bus load or configure it — with a full network it is 5 ms, and with fewer slaves it is proportionally shorter but bounded by the same rule. For machine control at the sensor level, that predictability is worth more than raw speed.
Extended addressing costs you. Slaves in extended mode are served every other cycle, so their effective update is roughly 10 ms.
Each slave exchanges 4 bits of input and 4 bits of output per cycle in the base profile, plus a 4-bit parameter value that configures device behaviour. That is deliberately small. AS-i was never meant to carry process values; it was meant to carry the states of switches and the commands to actuators, and four bits covers a surprising number of real field devices.
Error detection and the fault lists
AS-i’s error handling is simple and effective for what it is. Each telegram carries a parity bit and defined start and end bits, and the master validates the structure of every response. A slave that fails to answer within the slave response time-out — 11 bit times plus a margin — causes the master to end the transaction or repeat it.
That single repetition is built into the cycle budget, which is why the standard’s cycle calculation counts the number of requests during data exchange and the inclusion phase including one repetition.
The four master lists then do the diagnostic work:
- LPS — slaves the configuration says should be there
- LDS — slaves actually detected
- LAS — slaves detected and active
- LPF — slaves reporting a peripheral fault
A device in LPS but not LDS is missing or dead. In LDS but not LAS means it answered but did not activate — usually a configuration mismatch between the device’s I/O and ID codes and what the master expects. In LPF means the device is fine but its process side is reporting a problem, such as a broken sensor on a module’s input.
That progression is the fastest AS-i troubleshooting path there is, and it needs no tools beyond the master’s diagnostic display.
Topology, distance, and the repeater
AS-i is topology-free in a way most buses are not. Line, tree, star, branch anywhere — because the electrical model is a shared two-wire supply rather than a transmission line, there are no terminators and no stub-length rules.
The distance limit is 100 m per segment. Beyond that, the standard defines the repeater: a device that regenerates the AS-i signal and provides galvanic separation between parts of the network, so network lengths of more than 100 m become possible.
For anyone coming from CAN or PROFIBUS, the absence of termination discipline is the single biggest practical difference. Most of what goes wrong on those networks simply cannot go wrong here.
Profiles: how devices declare themselves
The standard covers profiles, which combine a slave’s I/O configuration and identification codes into a defined device type. A profile tells the master what the device is and how its four bits should be interpreted, without device-specific configuration.
This is what makes AS-i genuinely plug-and-play in a way most fieldbuses aspire to. A profiled slave from any manufacturer behaves the same way on any compliant master, and the extended ID codes let profiles distinguish variants within a family.
The standard also defines combined transactions, which is how AS-i carries more than four bits when needed — analogue values transferred across multiple cycles, and the safety-related communication used by AS-i Safety at Work. The safety mechanism is worth a note: safety slaves send code sequences of eight code nibbles rather than static data, and a safety monitor watches the sequence. A frozen or incorrect sequence means a fault. That is how AS-i achieves safety over a bus that has no other integrity mechanism.
Where AS-i wins, and where it stops
AS-i wins where device count is high, data per device is tiny, and installation labour dominates cost. Conveyor lines with dozens of proximity switches. Bottling plants. Material handling. Anywhere the alternative is hundreds of individual cable runs back to a marshalling cabinet.
The economics are the argument: no shielded cable, no terminators, no connectors, no marshalling, and a cycle time you can rely on. Adding a device late in a project means clamping a module onto the existing cable.
AS-i stops when devices have anything to say. Four bits per cycle is nothing. Analogue values need combined transactions across multiple cycles. There is no parameter tree, no device configuration beyond a 4-bit parameter, no diagnostics from the device beyond a peripheral-fault flag.
That is not a shortcoming — it is the design. AS-i occupies the layer below the fieldbuses, and the two coexist naturally: an AS-i segment gathers the binary field level, a gateway presents it to PROFIBUS, PROFINET, or DeviceNet, and the PLC sees the AS-i inputs as a block of bits in its process image.
That gateway arrangement is how AS-i appears in most plants. It is rarely the plant network. It is what hangs off the bottom of it.
FAQ
How many devices can AS-Interface support?
31 slaves in standard addressing, or 62 using extended addressing, which splits each address into A and B sub-addresses. Extended addressing costs update rate — those slaves are served every other cycle.
What is the AS-i cycle time?
About 5 ms with a full network of 31 slaves. The standard bounds it: with 30 or fewer transactions per cycle the cycle time shall not exceed 5 ms including management and inclusion phases. Slaves in extended addressing mode update every second cycle, so roughly 10 ms.
Why is the AS-i cable yellow?
The standard requires it — the AS-i standard cable shall be yellow, specified against RAL 1012, with brown for + and blue for −. The colour and the asymmetric profile together make the cable unmistakable and impossible to connect backwards.
Does AS-i need terminating resistors?
No. Unlike CAN or PROFIBUS, AS-i has no termination requirement and no stub-length rules, and supports line, tree, and star topologies freely. That absence removes the most common source of fieldbus commissioning problems.
How far can an AS-i network reach?
100 m per segment. Repeaters extend that by regenerating the signal and providing galvanic separation between network parts.
How much data does each AS-i slave exchange?
Four bits of input and four bits of output per cycle in the base profile, plus a 4-bit parameter value. Analogue and larger transfers use combined transactions spread across multiple cycles.
Can AS-i be used for safety applications?
Yes, through AS-i Safety at Work. Safety slaves transmit code sequences of eight code nibbles rather than static values, and a safety monitor watches for the correct sequence — a frozen or incorrect sequence is treated as a fault.
How does AS-i connect to a PLC?
Usually through a gateway that presents the AS-i input and output data images to a higher-level network such as PROFIBUS, PROFINET, EtherNet/IP, or DeviceNet. The PLC then sees the AS-i field level as a block of bits in its process image.