Education · Controller Wiring

Controller Wiring Wiring

Before a DDC controller can run a single line of sequence, every point on it has to be landed — power on its terminals, a sensor on each input, an actuator or a contactor on each output. Get the program perfect and one reversed wire still leaves you with a temperature that reads −40 °F or a damper that never moves. This page walks the four things that hang off a generic controller — power, inputs, outputs, and the common that ties them together — and the handful of wiring mistakes that catch everyone at least once. When you want to try the landings hands-on, the Controller Wiring Simulator is the same controller with live feedback and faults you can trip on purpose.

Power and the shared common

A typical field controller runs on 24 VAC from a small control transformer — line voltage (120 or 277 V) on the primary, 24 V on the secondary. The two secondary leads land on the controller's power terminals: the hot leg on 24V~ and the other leg on 24COM. 24 VAC is alternating current, so neither leg is "positive" — but one of them is the controller's reference, the 0 V that every other measurement is taken against, and that leg is COM.

This is the single most important idea in the whole panel: COM is the shared reference for everything. The return side of every sensor, the common of every analog output, the bottom of every input — they all come back to the same common as the controller's own power. An input measures the voltage (or resistance, or current) between its terminal and COM; if the sensor's return never reaches COM, the controller has no reference for it and the reading is meaningless. Most controllers give you a row of COM terminals near the inputs precisely so you have somewhere to land all those returns. They're internally bonded to 24COM — same node, different screws.

Where this turns into a field gotcha is sharing a transformer. Several controllers, or a controller and its powered actuators, often run off one transformer to save space and VA. That's fine — as long as every device references the same secondary leg as its common. Many controllers half-wave rectify their 24 VAC internally, which makes them polarity-sensitive: tie one device's COM to one leg and the next device's COM to the other, and the two commons sit a full 24 V apart. Now any wire that's supposed to be "common" on both — a shared sensor, a shield, a jumper — bridges the two legs and you get roughly 48 V across the gap, a blown fuse, or a fried input. The rule is simple and absolute: on a shared transformer, every COM lands on the same leg. Phase them, or fuse and separate them — never cross them.

A dead short across the secondary (hot tied straight to COM with no load between) blows the transformer's fuse instantly, which is the panel telling you, loudly, that the circuit is wrong. That's by design — the fuse is cheaper than the controller.

One transformer feeding two controllers on a shared common A 24 VAC control transformer on the left has two secondary leads: a hot leg drawn in red and a common leg drawn in blue. The hot leg runs to the 24V~ terminal of two controllers stacked on the right; the common leg runs to the 24COM terminal of both. Both controllers take their common from the same secondary leg — the correct, phased arrangement. A callout notes that crossing the legs between the two controllers would put 48 volts across the gap. XFMR 24 VAC 40 VA 24~ com CONTROLLER A half-wave 24 VAC CONTROLLER B half-wave 24 VAC both COMs on the same leg = same 0 V reference

24 VAC hot common / return

Inputs — what the controller reads

Most field controllers give you universal inputs (UI) that can be configured per point to read a few different electrical quantities, plus dedicated binary inputs (BI) for on/off contacts. The wiring is mostly about getting the right two or three conductors to the right terminals — and, every time, getting the return back to COM. What the input reads depends on how you configure it and what kind of device is on the other end.

A 10K thermistor is the simplest: two leads, no polarity, no power. One lands on the UI, the other on COM. The controller, configured for resistance, pushes a tiny known current through it and measures the voltage drop — that's the resistance, and the type curve turns resistance into temperature. Cut one lead (or land the return somewhere that isn't COM) and the controller sees infinite resistance, which the curve reads as full-scale cold. That's why a clean break shows up as a wildly low temperature, not a blank: the input is doing exactly what it's told with the open circuit it's given. (Get the type wrong — Type II curve on a Type III sensor — and it reads, but it reads several degrees off; configuration matters as much as wiring.)

A 0–10 V transmitter is a three-wire device: it needs power (often 24 V from the same control transformer), it puts out a 0–10 V signal on a third wire, and its own common ties back to COM. The controller, configured for voltage, reads the signal against COM. No power, no signal — the point reads 0 and sits there looking like a sensor pegged at the bottom of its range.

A 4–20 mA transmitter is where the loop-power idea trips people up. A two-wire transmitter doesn't have a separate signal wire — it regulates the current flowing through the loop it sits in, between 4 and 20 mA, in proportion to what it measures. That means the loop has to be powered: the 24 V hot leg drives current out to the transmitter's +, through the transmitter, back into the controller's input (configured for current) on the − side, where a small internal resistor turns the current into a voltage the controller can read, with the bottom of that resistor on COM. Forget to power the loop and there's no current to regulate — the point reads 0 mA, dead. The payoff is that 4 mA at the bottom of the range is a live, non-zero signal, so a broken wire (0 mA) is distinguishable from a real low reading; that's the whole reason the standard starts at 4, not 0.

A binary input watches a dry contact — a relay, a switch, a proof-of-flow paddle. The controller supplies its own low-voltage "wetting" current out the BI terminal; the field contact either completes the circuit back to COM (closed) or doesn't (open). Land one leg on the BI and the other on COM and you're done — but only if the return actually reaches COM. A contact wired to the BI with its other leg floating can never make the circuit, so the input reads OPEN no matter what the equipment does. Some contacts are instead externally powered ("wet"), already carrying a voltage from another panel; those don't want the controller's wetting voltage and have to be landed as the controller's manual calls out, or you back-feed one panel from another. Know which kind you have before you land it.

One thing this page deliberately skips: shielding and grounding of analog cable. The short version is that you land the cable's drain/shield at the panel end only, never both ends, or you create a ground loop that adds noise to the very signals you're trying to read cleanly. The Field Wiring — Sensors drill works through the why.

Four input landings on a controller A controller terminal column on the left lists 24V~, UI1, UI2, UI3, a COM terminal, and BI1. A red hot bus runs down from 24V~ and a blue common bus runs from COM. To the right, four devices land on them: a 10K thermistor on UI1 with its return to the common bus (signal wire dim); a three-wire 0 to 10 volt transmitter on UI2 taking 24 volt power off the hot bus, output to UI2, and common to the common bus; a two-wire 4 to 20 milliamp transmitter whose loop is powered off the hot bus and returns into UI3; and a dry contact on BI1 returning to the common bus. Every return lands on the shared common. CONTROLLER 24V~ UI1 UI2 UI3 COM BI1 10K THERMISTOR resistance · 2-wire 0–10 V XMTR 3-wire · pwr/out/com 4–20 mA XMTR 2-wire · loop-powered DRY CONTACT return to COM hot bus feeds the transmitters · every return lands on COM

24 VAC hot / loop power common / return signal

Outputs — what the controller drives

Outputs come in the same two flavors as inputs: analog outputs (AO) that modulate — a 0–10 V or 4–20 mA signal to a damper or valve actuator — and binary outputs (BO) that switch something on or off. The recurring mistake on the output side is forgetting that the controller's signal is not the load's power.

A 0–10 V actuator needs three things, and a tech in a hurry lands two of them. It needs its own power — 24 VAC to run the motor that swings the damper. It needs the control signal from the AO terminal. And it needs its signal common tied back to the controller's COM, so the actuator and the controller measure that 0–10 V against the same reference. Land the signal and skip the power and the actuator sits dead while the controller cheerfully commands 60% — the AO is an output, not a power source, and 10 V at a few milliamps won't swing a damper. Land the power and signal but float the common and the command has no shared reference and drifts. All three, every time.

A binary output is usually a dry relay contact inside the controller — two terminals that the controller closes together on command, switching nothing of its own. To make that relay do something, you pass a circuit through it: bring the 24 VAC hot leg to the relay's common terminal (often labeled BO-C or a per-output source pin), run the switched side out to the load — a fan contactor coil, a small relay, a solenoid — and return the load to 24COM. Close the relay and the load gets its 24 VAC; open it and the load drops out. The classic head-scratcher: everything's wired, the controller commands the BO on, the relay clicks — and nothing happens, because BO-C was never fed from the hot leg. The relay is switching a leg that has no power on it.

Some binary outputs are triacs instead of relays — solid-state switches that handle 24 VAC only (no dry-contact flexibility, and a small leakage current even when "off" that can hold a sensitive load partly energized), but with no moving contacts to wear out. They're common for driving floating/tristate actuators, where two triac outputs (open and close) inch an actuator in each direction. The wiring intent is the same — switch a 24 VAC leg to a load — but a triac isn't a dry contact: it won't switch a separate voltage, and that leakage means it's the wrong choice for, say, proving a circuit truly de-energized.

An analog and a binary output landing A 0 to 10 volt actuator takes three connections from the controller: 24 volt power (red) and a signal common back to 24COM (blue) to run its motor, and a control signal from the AO1 terminal (dim). Below it, a binary output drives a fan contactor: a jumper feeds the hot leg to the BO-C source terminal (red), the BO1 contact switches out to the contactor coil, and the coil returns to 24COM (blue). A note reads: feed BO-C from the hot leg, or the relay clicks but nothing runs. CONTROLLER 24V~ 24COM AO1 BO-C BO1 0–10 V ACTUATOR power + signal + shared common FAN CONTACTOR 24 VAC coil feed BO-C from the hot leg, or the relay clicks but nothing runs

24 VAC hot common / return signal

The whole controller, landed

Put it together and a single generic controller carries all of it at once: power in from the transformer, a thermistor and a couple of transmitters on the inputs, a dry contact on a BI, an actuator on an AO, and a fan contactor on a BO. The diagram below traces one of each on a DDC-8 — the same controller as the simulator. The amber paths are the ones with current actually flowing: the 24 VAC hot bus feeding everything, and the 4–20 mA loop circulating from the hot leg, through the transmitter, and back into the input. Blue is the common every return shares; dim is the signal wiring; red is hot at rest.

A fully landed DDC-8 controller A generic DDC-8 controller terminal strip down the left side, powered from a 24 VAC transformer at top right. The hot leg (amber, energized) feeds the 24V~ terminal and runs down a vertical hot bus that powers the loop transmitter, the actuator, and the BO-C source. The common leg (blue) feeds 24COM and runs down a vertical common bus that every device return lands on. On the right, one device of each kind: a 10K thermistor on UI1; a 4 to 20 milliamp transmitter on UI2 whose loop circulates from the hot bus through the transmitter and back into the input (amber); a dry contact on BI1; a 0 to 10 volt actuator on AO1 taking power, signal, and common; and a fan contactor switched by BO1. Direction arrows mark current flow on the energized amber paths. GENERIC DDC-8 24V~ 24COM COM UI1 UI2 BI1 AO1 BO-C BO1 XFMR 24 VAC 10K THERMISTOR UI1 + COM 4–20 mA XMTR loop-powered DRY CONTACT BI1 + COM 0–10 V ACTUATOR pwr + sig + com FAN CONTACTOR switched by BO1

energized (current flows) 24 VAC hot common / return signal

That's the whole job: power referenced to a shared common, each input fed the conductors it needs to read, each output given both a signal and the power to act on it. The Controller Wiring Simulator lets you land every one of these yourself and watch what happens when you get one wrong — an open thermistor, an unpowered loop, a BO-C left off the hot leg.

One piece sits just outside this page: the network. A field controller almost always has an RS-485 / BACnet MS/TP trunk or a BACnet/IP port so the BMS can talk to it — its own small world of daisy-chain topology, +/− polarity, end-of-line termination, and station addressing. That's a wiring story of its own, and a dedicated bus simulator is the right place to walk it; for now the controller's network terminals are the seam where it plugs in. For the protocol side of what rides that trunk, the BACnet Networking explainer picks it up.

← All explainers