6.4 Detector Amplifier Setup and Detection Troubleshooting
Key Takeaways
- Detector sensitivity determines the smallest inductance change that produces a call; raising it detects motorcycles and bicycles but also increases splashover and crosstalk.
- The frequency-select switch exists to separate adjacent channels electrically and is the correct remedy for crosstalk, not a sensitivity adjustment.
- Detector cards flag open loop, shorted loop, and excessive inductance change as loop faults, and most units are configurable to output either a constant call or no call when a fault is latched.
- Substituting a known-good detector card is the fastest way to divide the problem between the cabinet electronics and everything outside the cabinet.
- A stuck-on detector produces continuous max-out on its phase, while a dropped detector produces premature gap-out — the two symptoms point in opposite directions.
6.4 Detector Amplifier Setup and Detection Troubleshooting
The detector card — the electronics unit — is where the loop's physics becomes a controller input. It is also the easiest component in the chain to swap, which makes it the natural first move in a diagnostic sequence and, for undisciplined technicians, the reason so many good cards end up in the shop.
What the Settings Actually Do
Sensitivity
Sensitivity sets the threshold inductance change required to declare a detection, usually expressed as a percentage change and selected on the card by a rotary switch or menu, commonly across a range of settings numbered roughly 1 through 7 or 0 through 9.
| Setting | Behavior | Failure Mode |
|---|---|---|
| Too low | Only large vehicles produce enough change | Motorcycles, bicycles, and small cars are not detected; drivers wait indefinitely for a phase that is never called |
| Correct | Detects the design vehicle reliably in its own lane | — |
| Too high | Detects very small inductance changes | Splashover from the adjacent lane, crosstalk with nearby loops, and false calls from pavement moisture changes |
The field procedure is empirical: raise sensitivity until the smallest design vehicle (a bicycle where bicycle detection is required, otherwise a motorcycle) actuates consistently from the marked position, then confirm that a vehicle in the adjacent lane does not actuate the channel. If both conditions cannot be met, the problem is loop geometry, not sensitivity — the loop is too large or badly placed and needs to be replaced with a smaller loop or a quadrupole.
Frequency Select
Most cards offer two to four selectable operating frequencies. Adjacent loops that share a frequency couple electrically and each reports the other's traffic. Setting neighboring channels to different frequencies breaks that coupling. This is the correct and only real fix for crosstalk; reducing sensitivity merely masks it and cripples the loop.
Presence vs. Pulse Mode
- Presence mode holds the call for as long as the detection zone is occupied. Standard for stop-bar and turn-bay detection.
- Pulse mode issues a short fixed-duration call when the vehicle arrives and then releases, regardless of dwell. Used for counting and for setback detectors where a vehicle stopped over the loop should not hold a permanent call.
A card left in pulse mode on a stop-bar detector produces the most confusing symptom in the trade: the phase is called correctly but will not hold, so a single waiting vehicle gets minimum green and nothing more.
Card-Level Timing
- Delay — requires continuous occupancy for a set period before a call is passed. Suppresses right-turn-on-red calls.
- Extend (carryover) — holds the call after the vehicle clears. Bridges gaps between short zones or between an advance and stop-bar loop.
These timers exist both on the detector card and in the controller. Setting the same delay in both places doubles it, which is a routine source of "the left turn takes forever to come up" complaints after a card replacement.
Loop Fault Detection
Modern detector cards continuously supervise the loop circuit and latch a fault for:
| Fault | Detected Condition | Usual Field Cause |
|---|---|---|
| Open loop | Inductance out of range high, or no resonance | Cut loop wire, failed splice, disconnected terminal |
| Shorted loop | Inductance out of range low | Insulation failure conductor-to-conductor |
| Excessive inductance change | Inductance shifted beyond a set percentage (often around 25%) from the reference | Loop degrading, water intrusion, pavement movement |
When a fault latches, the card's configured fail-safe behavior takes over. Two options are common:
- Fail to constant call — the channel outputs a permanent call so the phase is served every cycle. Safe for traffic, and it makes the failure visible as a phase that maxes out continuously.
- Fail to no call — the channel outputs nothing, and the phase is simply never served unless recall is set.
Agencies overwhelmingly configure critical phases to fail to constant call, because a phase that silently stops being served can strand a driver indefinitely. A technician replacing a card must confirm the fail-safe configuration matches the agency standard rather than the card's factory default.
Reading the Card's Indicators
A typical card front panel shows a DETECT indicator (illuminated while a call is output) and a FAULT indicator (illuminated or flashing on a latched loop fault, often with a flash pattern that encodes open versus shorted). Watching the DETECT indicator while a vehicle drives the loop is the single fastest confirmation that the entire outside-the-cabinet chain is intact.
A Repeatable Troubleshooting Sequence
The goal of the sequence is to divide the system in half at each step rather than replacing parts in order of convenience.
- Reproduce and characterize. Is the phase never called, called constantly, or called intermittently? Constant call points at a short or a stuck card; no call points at an open or a dead card; intermittent points at water, a marginal splice, or a mechanical problem.
- Read the controller's detector status. Modern controllers display live detector status and log detector diagnostics — failed detectors, no-activity alarms, and max-presence alarms. This tells you what the controller is receiving without a meter.
- Read the card's indicators. A latched fault indicator immediately moves the investigation outside the cabinet.
- Substitute a known-good card. If the symptom follows the card, the card was bad. If the symptom stays, the fault is outside the cabinet. This one step splits the problem cleanly and takes under a minute.
- Measure at the cabinet terminals. Disconnect the pair and take series resistance, insulation resistance, and inductance. Compare against design values.
- Measure at the pull box. Same three readings on the loop side of the splice. If the loop reads correctly here but not at the cabinet, the fault is the splice or the home-run cable. This is the highest-value measurement in the sequence and the one most often skipped.
- Inspect physically. Open the pull box and look for standing water, a floated splice, rodent damage, or a home-run cable pinched at the conduit entry. Look at the sawcut for missing sealant or exposed wire.
Symptom-to-Cause Reference
| Field Symptom | Most Likely Causes |
|---|---|
| Phase maxes out continuously with no traffic | Shorted loop; card failed to constant call; card stuck; water in pull box |
| Phase is never called | Open loop; card set to fail-to-no-call with a latched fault; loop pair disconnected; sensitivity too low |
| Motorcycles and bicycles not detected | Sensitivity too low; loop geometry wrong for small vehicles; wrong loop type for a bicycle lane |
| Calls appear from the adjacent lane | Splashover (loop too large or sensitivity too high) or crosstalk (identical frequencies on adjacent channels) |
| Detector works when dry, fails after rain | Water intrusion at a splice or in the sawcut; low insulation resistance |
| Phase called but green will not hold | Card in pulse mode where presence is required; passage time too short |
| Left turn takes far too long to come up | Delay timer set both on the card and in the controller |
| Detection failed immediately after nearby construction | Loop or home-run cable cut by an excavation; verify with continuity and insulation readings before assuming anything else |
Documentation
Every detector repair produces two records: the work order (what failed, what was measured, what was replaced) and the cabinet timing and configuration sheet if any setting changed. Recording the actual measured resistance, insulation resistance, and inductance values matters more than recording the conclusion — the next technician needs the trend, because a loop that read 60 megohms last year and 12 megohms today is failing whether or not it is calling correctly right now.
A technician raises detector sensitivity to detect motorcycles at a stop bar, and the channel now also calls when vehicles pass in the adjacent lane. What does this indicate?
A detector card latches a loop fault and its fail-safe behavior is configured to output a constant call. What will the field symptom be?
In a detection troubleshooting sequence, what is the value of measuring loop resistance, insulation resistance, and inductance at the pull box after already measuring them at the cabinet?
After a detector card is replaced, drivers report that the left-turn phase now takes noticeably longer to appear than it did before. Everything else operates normally. What should be checked?