You are called to a unit you have never seen: the heat will not run, and the wiresheet on the controller is live in front of you. What is the fastest first move?

Find the heat output on the sheet and step upstream from it, one wire at a time.. A broken sheet is still mostly healthy — one block is lying and the rest are doing their jobs. Reading forward from the inputs inspects the whole healthy majority before touching the fault; starting at the wrong output and tracing upstream prunes everything that is not on the fault path. Forward is how you learn a sheet; backward is how you fix one. (Forcing the output has its place as an experiment — but as a first move it proves the equipment, not the logic, and an undisciplined force on an occupied building is its own hazard.)

Tracing backward from an OFF heat output, you reach the gate AND. Its permit leg reads TRUE; its call leg reads FALSE. Where does the trace go next?

Up the FALSE call leg — the TRUE permit leg just retired its whole upstream chain in one glance.. The AND is honest — given TRUE and FALSE, FALSE is the right output — so the fault is upstream of one of its legs. A TRUE leg makes sense, which retires the entire subtree that built it: the permissive chain is cleared without being walked. The FALSE leg is the one that fails to match what the sequence needs (there is a call for heat), so the trace follows it. Backward tracing is a pruning strategy — the sheet gets smaller at every hop.

A live sheet shows an SR latch with Q reading TRUE while its S input reads FALSE. Is the latch broken?

No — Q is state: S only needed to be TRUE once, possibly thousands of scans ago, and the latch has held since.. A latch remembering is the idiom working, not a fault. Q is state, held across scans; S is a momentary cause that may be long gone. At a stateful block the hop question shifts tense: not "given these inputs, is this output right?" but "given these inputs and a legitimate past, is it right?" — so you ask what could have fired S, not just what S says now. A latched trip whose cause has cleared looks exactly like this, on purpose, until someone resets it.

On a sheet with a feedback loop — a wire carrying a block's output back into its own chain — the value on that feedback wire is how many scans old?

1 scan(s). Exactly one. The controller evaluates blocks in dependency order within each scan, and a wire that loops back cannot be fresh and be an input at the same time — so it carries the previous scan's value, one scan stale, by design. That one-scan memory is why a latch can hold, why a loop resolves instead of chasing itself, and why a feedback wire's display can legitimately disagree with the value you would compute from this scan's inputs.

On a live wiresheet you can watch a signal work its way across the sheet — the comparator updates, then the AND a moment later, then the output.

False. False — a whole combinational chain settles within one scan, in dependency order, and the controller scans far faster than the display repaints. Live values move together between paints; you never see a value in flight. The diagnostic corollary: a chain that disagrees with itself — TRUE and TRUE into an AND showing FALSE — is not propagation delay, because there is none to see. That block is your NO.

You suspect a fault somewhere between a temperature input and the heat output, several blocks deep. What is the fastest experiment a live sheet offers?

Force the temperature input to a value that should demand heat and watch the ripple — every downstream block confesses within a scan.. Because the whole sheet settles every scan, a forced input ripples through everything downstream before the next repaint — comparators, latch, gate, and output all answer at once. A yes moves the fault upstream of your force; a no keeps it downstream — a binary search that converges in a few forces. The discipline that keeps it safe is commissioning's override rulebook: log every force, clear every force. A forgotten force is a live hazard wearing a disguise.

Point-to-point checkout and live-wiresheet troubleshooting are the same activity under two names.

False. False — they answer different questions. Commissioning proves a CORRECT program was installed: point-to-point checkout walks every physical point against the sequence, with override discipline and a signed record. Troubleshooting a live sheet assumes the install and asks which block is lying now. They share tools — forces, live values, the sequence of operations — but checkout is systematic proof over every point, and a trace is targeted pursuit of one fault. Knowing which job you are doing tells you whether to walk everything or prune everything.

An unfamiliar sheet shows the outdoor-air temperature feeding a MULTIPLY with a negative constant, then an ADD, then a LIMIT clamped between two values, and out to a hot-water setpoint. What are you looking at?

A reset chain — a setpoint computed from a condition: y = mx + b with a clamp.. MUL → ADD → LIMIT fed by a condition is the reset chain: the multiply is the slope, the add is the intercept, and the limit clamps the line to a sane band — a hot-water setpoint that slides down as the day warms. Recognize it and the trace collapses: if the setpoint looks wrong, check the input, the two constants, and the clamp values — four numbers, not four blocks of mystery. A setpoint that appears typed-in but keeps moving is almost always one of these.

On an unfamiliar sheet you find: an ADD and a SUBTRACT building setpoint-plus and setpoint-minus values, a greater-than and a less-than comparing the space temperature against them, and an SR latch fed by both comparators. What are you looking at?

A band-edge cluster — a deadband thermostat that switches once per crossing instead of chattering.. SP ± DB → GT/LT → SR latch is the band-edge cluster: math builds the two edges, the comparators watch them, and the latch holds between crossings — hysteresis built from stock blocks. Naming it collapses five blocks into one word, and the word tells you what every pin should be doing: set fires past one edge, reset past the other, and anything else is your NO. Reading idioms first means the trace hops between patterns, not between blocks.

A mixed-air low-limit should trip when the mixed-air temperature falls below 45 °F (7.2 °C). The live sheet reads as below — and the unit is tripped on a healthy 55 °F (12.8 °C) morning. Which block is lying? MA TEMP (analog in): 55 LIMIT SP (constant): 45 comparator A > B: TRUE LOW LIMIT TRIP (BO): ON

The comparator — every value on the sheet is honest arithmetic, but its sentence is backwards: the low-limit needs A 45 = TRUE. The first hop where an output stops matching what the sequence needs is the comparator — a low-limit must answer "has the temperature fallen BELOW the limit?", which is A < B. Someone configured a greater-than where a less-than belongs, so the trip is on whenever the air is comfortably warm and would clear in a freeze — the exact backwards behavior a swapped comparison always produces. The config is the lie; the arithmetic never was.

Practice · Reading a Wiresheet

Reading a Wiresheet Quiz Logic

Read the Reading a Wiresheet lesson first, then test yourself — ten questions on the method: the first move when an output is wrong, which leg the trace follows, why a latch showing Q TRUE over a FALSE S is healthy, what the scan means for the values you watch, and naming a five-block idiom from its description. Each reveal links back to the section it came from.

← All practice