Audio deep dive
4:11Before you change the flame-detector settings
A two-host troubleshooting discussion about optical interferents, field of view, detector technology, configuration and system-chain evidence.
Editorial note: This AI-produced discussion is an accessible companion to the written guide and may simplify or paraphrase technical and regulatory points. For precise terminology, standards references and engineering boundaries, refer to the written article and its cited sources. Always follow the applicable OEM instructions, approved site procedures and project-specific engineering requirements.
The first rule
Treat the indication as a potential real fire until the site’s authorised response establishes otherwise.
A repeated nuisance pattern can damage alarm confidence. It must never become permission to silence, cover, bypass or desensitise a detector outside controlled site procedures.
Classify before correcting
An unwanted alarm and a detector fault are not the same event
An optical unwanted alarm begins when the detector’s sensing and processing criteria declare fire in response to a non-hazardous source or an unintended real-flame path. A spurious system indication can originate later in the chain—from power, wiring, communication, input mapping or logic. A fault or trouble condition means the detector or system has identified an impairment.
Those paths need different corrective work. Start with the event class, local detector status and output evidence rather than the label used in a shift report.
Optical alarm
The detector’s fire algorithm declared a flame-like signature.
Fault / trouble
A diagnostic identified impaired optics, power, electronics or communications.
Spurious indication
The panel or system reported fire without evidence that the optical algorithm declared it.
Spectral bands plus signal behaviour
The technology determines which false-alarm explanation is plausible
Detector selection is not only about wavelength and not only about software. The sensing bands decide what energy reaches the electronics; the processing method decides whether the combination, level and behaviour satisfy the fire criteria. Exact thresholds and algorithms are proprietary and model-specific.
UV
Senses: Ultraviolet radiation produced by combustion.
Evaluates: The detector evaluates UV energy and, depending on the model, its pulse behaviour over time.
Watch: Arc welding, electrical discharge, lightning and certain artificial UV sources can enter the same sensing band.
Single IR
Senses: Infrared energy in a selected combustion-related band.
Evaluates: Signal level and flame-like time variation help distinguish a flame from a steady background.
Watch: Hot equipment, modulated thermal backgrounds and exhaust or flare conditions can challenge the discrimination method.
UV/IR
Senses: A UV channel and an IR channel in the same detector.
Evaluates: Coincidence logic requires both channels to satisfy the model’s fire criteria.
Watch: It rejects many single-band interferents, but intense sources that stimulate both bands and contamination that weakens one channel still matter.
IR3 / MSIR
Senses: Several infrared bands rather than one band alone.
Evaluates: Band relationships, signal dynamics and model-specific algorithms are compared before an alarm is declared.
Watch: Immunity is relative—not absolute. Fuel, flare, exhaust, welding, geometry and sensitivity remain application-specific.
Do not generalise a model table: false-alarm immunity, fuel response, field of view, welding distance and range are test-condition and configuration claims. Use the manual and approval data for the installed detector.
01 · External optical stimulus
Interfering radiation or a real flame signature enters the field of view
Hot work, electrical arcs, hot process surfaces, engine or turbine exhaust, flares, lightning and reflected radiation do not affect every detector in the same way. The first useful question is therefore not “what is the universal cause?” but “what technology and model saw what source, from which direction, at that time?”
Evidence that points this way
- Event aligns with welding, grinding, startup, flare or vehicle activity
- Repeat occurs at a similar time of day or operating state
- One detector alarms while others with a different view or technology remain normal
Check before changing
- Preserve the detector and panel event records before changing settings
- Correlate the timestamp with permits, process logs and weather
- Walk the field of view from the detector position and look for direct and reflected paths
Safety boundary: Do not assume “welding” makes the event harmless. Gas cutting and other hot work can create a genuine fire, and site emergency and permit procedures take priority.
02 · Optical path condition
Contaminated optics, weather or an obstruction changes what the detector can see
Dust, oil film, salt, paint overspray, condensation, ice, smoke, fog and physical obstructions usually create a sensitivity or fault problem before they create a true optical false alarm. Uneven contamination, changing shadows or interaction with another radiation source can nevertheless complicate the signal and the diagnosis.
Evidence that points this way
- Dirty-window, optical-integrity or general fault appears before or after the event
- The problem follows rain, wash-down, dust, coating or construction work
- Scaffolding, sheeting, pipework or equipment has changed the line of sight
Check before changing
- Inspect the viewing window and the complete external path—not only the detector face
- Use only the cleaning method and materials in the exact model manual
- Check weather shields, heaters, seals, drains and enclosure condition where fitted
Safety boundary: A passing internal optical check is not proof that the protected hazard remains visible through the external scene or that rated range is still available.
03 · Geometry and mechanical condition
Placement, orientation or mounting exposes the detector to the wrong scene
A detector can be healthy yet poorly applied. A field of view that includes a welding bay, hot exhaust, flare reflection, roadway or bright rotating equipment can repeatedly present the algorithm with unwanted signals. A loose or vibrating mount can also move the scene relative to a background source.
Evidence that points this way
- The alarm history began after plant modification or detector replacement
- The mounting arm is loose, damaged or exposed to vibration
- The detector covers the hazard only near the edge of its directional response
Check before changing
- Compare the as-found aim with current drawings and the coverage study
- Inspect the mount, support and cable entries without altering the position first
- Review new obstructions and interference sources across the full cone of vision
Safety boundary: Re-aiming, adding a shield or moving a detector can create a blind spot. Treat it as an engineering change and re-check coverage.
04 · Performance settings and cause-and-effect
Sensitivity, verification time or system logic does not match the hazard and environment
Sensitivity, processing modes, alarm delays, latching and voting influence both unwanted-alarm exposure and real-fire response. The correct values are model- and project-specific. A setting that appears to solve one nuisance event can silently reduce detection performance or delay executive action.
Evidence that points this way
- The issue follows commissioning, firmware, configuration or panel-logic work
- Installed settings do not match the approved configuration record
- A single detector event causes a shutdown even though the design basis expected voting
Check before changing
- Capture as-found detector and panel configuration before alteration
- Compare settings with the approved fire-and-gas philosophy and cause-and-effect
- Confirm the exact model, mode, sensitivity and output mapping
Safety boundary: Reducing sensitivity, extending delay or changing voting is not a maintenance shortcut. It requires documented hazard review, approval and confirmation of the resulting coverage and response.
05 · Detection chain rather than flame algorithm
Power, wiring, communication or enclosure faults are reported as an alarm event
Not every “false alarm” originated in the optical decision. Loose terminations, moisture ingress, power disturbance, grounding or shielding problems, communications faults and incorrect input mapping can create a spurious panel indication or confuse alarm and fault states.
Evidence that points this way
- Detector local status and panel indication do not agree
- Several devices fault together or the event follows a power disturbance
- The event appears after enclosure, cable, junction-box or control-system work
Check before changing
- Compare local LED, analogue output, relay state and panel event class
- Check power and wiring using the exact OEM and site electrical procedures
- Inspect glands, seals, drains, junction boxes and documented termination changes
Safety boundary: Do not use generic insulation-resistance or live-circuit tests on detector wiring. Some OEM manuals prohibit particular test methods; the installed model manual governs.
Illustrative—not a ranking
Four model families show why “multispectrum” is not one universal algorithm
These current manufacturer examples make the functional differences tangible. They do not establish equivalence, preferred supply, universal immunity or suitability for a particular facility.
Det-Tronics X3301
Multispectrum IR
Three IR sensors, event logging, selectable sensitivity and automatic optical-integrity checking. Its manual separates fire, fault and high-background conditions and gives model-specific controls for testing and maintenance.
MSA FL5000
Multispectrum IR
Multiple IR inputs are classified by an advanced algorithm. MSA describes the family as using neural-network-based processing to improve rejection of lightning, reflected sunlight, arc welding and hot objects.
Honeywell FS24X
Triple IR + visible
WideBand IR, a combustion-related IR band, visible sensing and dual processors are used for signal analysis, diagnostics, event records and model-specific false-alarm rejection.
Dräger Flame 1500
IR3
Three IR wavelengths are compared with combustion-related CO₂ information and flame flicker. A heated window and optical verification support environmental robustness and diagnostics.
Evidence with limits
An optical self-test is valuable—but it does not prove site coverage
Built-in tests differ. The Det-Tronics X3301, for example, performs a calibrated automatic optical-integrity check once per minute and distinguishes the automatic check from manual checks that actuate alarm output. Dräger describes an optical verification test for the Flame 1500. The exact test path, pass threshold and output behaviour must come from the installed model manual.
Can support
Power, electronics and defined internal optical-path checks are functioning at the time of the test.
Can support
A model-specific contamination threshold or diagnostic condition has—or has not—been reached.
Cannot prove
The detector still sees the entire intended hazard through the external field of view.
Cannot prove
The selected model, sensitivity, placement and logic are correct for the site hazard.
Cannot prove
A test lamp or internal check reproduces every fuel, distance, weather and background condition in the design basis.
External scene still matters: an internal check cannot see a new scaffold outside the detector, prove the aim, confirm the whole cause-and-effect path or reproduce every design-basis fire.
Preserve evidence before adjustment
A safe investigation sequence moves from event response to controlled restoration
Respond as a potential real event
Follow the site emergency response. Confirm the location and condition through authorised means before classifying the indication as unwanted.
Preserve evidence
Record detector tag, panel event class, exact times, duration, local indication and any alarms, faults or inhibits around the event.
Correlate the plant context
Check hot-work permits, process transitions, flare or exhaust activity, maintenance, vehicle movement, weather and recent modifications.
Inspect the scene and hardware
Under the required permit and access controls, inspect optics, line of sight, reflections, obstructions, mounting, enclosure, glands and junction boxes.
Separate optical from electrical
Compare the detector’s local state and diagnostics with its analogue, relay or communications output and the control-system event record.
Clean or correct only approved defects
Use the exact OEM procedure. Do not re-aim, desensitise, bypass or alter voting as an informal experiment.
Perform the model-approved verification
Control extinguishing or shutdown outputs under the site procedure before a test that can actuate them. Use only the approved built-in test or compatible simulator method.
Escalate unresolved patterns
Provide the OEM or competent F&G engineer with model, firmware, settings, event history, photographs, environmental context and as-found test results.
Restore and close out
Confirm all protection, outputs and inhibits are restored; record as-found/as-left status and route any design change through management of change.
Stop conditions are part of the method.
Stop and escalate if the event cannot be safely classified, the procedure or manual is unavailable, the test may trigger suppression or shutdown, hazardous-area integrity is uncertain, or a proposed change affects detection coverage or cause-and-effect.
Condition-led and model-specific
The maintenance record should make the next alarm easier to diagnose
- 01Review detector and panel event histories for recurring time, weather or process patterns.
- 02Inspect the window, weather shield, mount, support, conduit entries, drains and enclosure condition.
- 03Confirm the external field of view remains clear after scaffolding, construction and plant changes.
- 04Clean only with the exact model’s approved materials and method; avoid residues and abrasive damage.
- 05Verify built-in diagnostics and carry out the approved functional test at the site-defined interval.
- 06Record detector tag, model, configuration, as-found condition, work performed, result and as-left status.
- 07Review sensitivity, delays and system logic against controlled documents—without changing them during routine inspection.
- 08Increase inspection frequency only from site evidence, environment, OEM instructions and the maintenance strategy—not a generic web interval.
Useful history: trend the event against plant activity, weather, condition, diagnostics, cleaning and configuration. “Tested OK” without as-found evidence rarely explains a recurring event.
One framework, different dominant sources
The five categories stay stable; their local ranking does not
Oil, gas and chemicals
Flares, hot work, reflective plant, hydrocarbon films, corrosive atmospheres and changing process backgrounds can dominate the investigation.
Mining and materials handling
Dust, vibration, mobile equipment, exhaust, welding and frequent changes to physical obstructions can alter both optics and field of view.
Power generation
Turbine and engine exhaust, hot casings, boilers, steam and electrical work create technology-specific optical and environmental challenges.
Warehousing and manufacturing
Racking, temporary screens, high-bay lighting, vehicles, hot work and production-layout changes often affect the detector scene.
A facility’s own event history is the only defensible way to rank its leading contributors. Even then, separate false optical alarms, real but unintended flames, diagnostic faults and spurious system indications before trending.
Five misconceptions
The shortcuts that turn nuisance reduction into hidden risk
“A multispectrum detector cannot false-alarm.”
Multispectrum processing improves rejection, but immunity remains model-, source- and application-specific. No technology removes the need for correct selection, placement and maintenance.
“If the optical self-test passes, coverage is proven.”
A built-in check verifies defined internal functions. It does not prove the external field of view, hazard geometry, atmospheric path or complete alarm action.
“Dirty optics are mainly a false-alarm problem.”
Contamination is often the more serious missed-detection or fault problem. It can also interact with changing light and shadows, so the actual event evidence matters.
“Lower sensitivity or a longer delay is a harmless fix.”
Both can reduce or slow real-fire detection. Any change must remain consistent with the hazard basis, coverage and approved cause-and-effect.
“A panel alarm proves the detector saw a flame.”
The complete chain includes sensing, local processing, outputs, wiring, communications, input mapping and system logic. Compare local and remote evidence before assigning the cause.
Common questions
What plant teams should ask before they reach for a setting change
Should an optical flame-detector alarm ever be treated as false immediately?+
No. Follow the site emergency and verification procedure first. Only classify the event after authorised checks establish that no fire condition exists and the event evidence has been preserved.
Which detector technology has the fewest false alarms?+
There is no universal winner. Multispectrum IR technologies generally use more optical information to reject common interferents, but the best choice depends on fuel, distance, field of view, hot backgrounds, flare or exhaust conditions, weather, approvals and required response.
Can a dirty lens cause a false alarm?+
It can complicate the signal or interact with another source, but contamination more commonly reduces sensitivity or produces a fault. Inspect the event class, optics and model diagnostics rather than assuming a direct cause.
Can we use any flame-detector test lamp on any detector?+
No. Test sources and procedures are technology- and model-specific. Some models use approved simulators; some rely on calibrated internal optical tests. Follow the current manual and control any outputs the test can actuate.
How often should optical flame detectors be cleaned and functionally tested?+
Use the current OEM instructions, the site fire-and-gas performance and maintenance strategy, applicable project requirements and evidence from the actual environment. There is no defensible universal monthly, quarterly or annual interval for every installation.
Primary references
Sources, standards hierarchy and claim boundaries
Seven supplied research files were consolidated for this rewrite. Technical claims were then checked against current manufacturer pages and manuals plus official ISA and ISO catalogue records. The OEM manual and approved site procedure govern work on the installed detector.
Process facilities
ISA-TR84.00.07 provides process-industry guidance for evaluating fire and gas detection effectiveness, including coverage, response and reliability considerations.
Point-type detector standard
ISO 7240-10 covers point-type, resettable flame detectors for building fire-detection systems and says special-risk detectors fall outside its direct scope, though it may offer guidance.
NFPA 72 may be relevant where adopted by the authority having jurisdiction, contract, insurer or site standard. It is not presented here as South African law. Hazardous-area, fire, permit, impairment and change-control requirements must be established for the actual facility.
- Det-Tronics — X3301 product overview
- Det-Tronics — X3301 enhanced instruction manual
- MSA — FL5000 device overview
- Honeywell — FS24X product page and technical resources
- Dräger — Flame 1500 IR3 product page
- Emerson — industrial flame-detection technology overview
- ISA — ISA-TR84.00.07-2018 fire and gas system effectiveness guidance
- ISO — ISO 7240-10:2012 point-type flame-detector scope
- NFPA — NFPA 72, National Fire Alarm and Signaling Code
What this article does not promise
This article does not promise zero false alarms, universal detector immunity, a guaranteed detection distance, a generic cleaning or proof-test interval, or compliance with a particular code. It does not authorise bypassing, shielding, moving, desensitising or changing detector logic. It does not claim manufacturer, certification, approved-inspection or accreditation authority for Touch Teqniques Engineering. Exact work must be scoped against the installed models, approved fire-and-gas basis, site procedures and applicable legal and contractual requirements.
Research status
Source-led technical briefing
Manufacturer manuals and official standards records checked
This article was rebuilt from a consolidated technical research pack and checked against current manufacturer documentation plus official ISA and ISO catalogue records. It is general information and does not replace the installed detector manual, approved site procedure or a site-specific fire and gas engineering assessment.
Research updated: 11 August 2026
Relevant domains: Optical flame-detector technologies, False-alarm and fault investigation, OEM diagnostics and maintenance limits, Fire and gas system effectiveness.