The Art of Troubleshooting

Here is a truth no montage diagram will tell you: on a real night, things go wrong. Electrodes pop, patients sweat, signals vanish, and a channel that looked perfect at lights-out turns to garbage at 2 a.m. What separates a competent technologist from a great one is not that their setups never fail — it's how calmly and systematically they fix failures when they happen. Troubleshooting is arguably the most important active skill in the job, because the equipment records blindly; only a thinking human notices that a signal is lying and corrects it before hours of data are lost.

This chapter teaches you to think like a troubleshooter. The specific problems matter, and we'll cover them, but the deeper lesson is a method: a calm, logical way of moving from "that looks wrong" to "here's why, and here's the fix," ideally without waking your patient.

In this chapter you will learn to:

• Approach a bad signal systematically instead of reactively.

• Verify signal quality before lights-out and troubleshoot in real time after.

• Recognize and correct the common artifacts and sensor failures.

• Tell the difference between a signal problem and a real patient event.

• Decide when to enter the room and when to observe, minimizing patient disruption.

• Document troubleshooting so it protects the study and the scorer.

Why troubleshooting is so important

Every other skill in this book produces its value only if the signal is clean. Perfect electrode placement, a beautifully prepped patient, flawless staging knowledge — all of it is wasted if a loose lead turns your frontal channel to noise for three hours and no one catches it. The recording system will happily save hours of meaningless data; it has no idea anything is wrong. You are the safeguard.

Troubleshooting also protects the patient's diagnosis. A study riddled with unaddressed artifact is hard or impossible to score accurately, which can mean a repeat study (costly and burdensome to the patient) or, worse, a wrong interpretation. When you fix a failing signal at 2 a.m., you are protecting the truth of a medical decision made days later. That is why labs value a strong troubleshooter so highly, and why this skill is worth practicing deliberately.

Think systematically: a method, not a panic

The single biggest mistake new technologists make is reacting to a bad signal randomly — yanking at wires, changing settings, waking the patient — instead of reasoning. Build the habit of a systematic approach. When a signal looks wrong, work from the simplest and least invasive explanation outward:

1. Confirm it's real. Is the signal actually wrong, or are you misreading a normal physiological change? (More on this below.)

1. Localize it. Is one channel affected, or many? One bad channel points to that electrode or its connection; many channels point to a shared cause (a reference, a ground, environmental interference, the patient).

1. Check the least invasive causes first. Settings, cabling, and connections at the headbox before anything that touches the patient.

1. Then the patient side. The electrode, the prep, the skin — addressed with minimal disturbance.

1. Verify the fix and document it.

The pattern "one channel vs. many channels" is your most powerful diagnostic shortcut. A problem in a single derivation is almost always that electrode or lead. A problem appearing across many channels at once is almost always something shared: the reference electrode, the ground, environmental noise, or the patient themselves (sweat, movement). Train yourself to ask "one or many?" first, every time.

Signal quality before lights-out

The best troubleshooting happens before the patient is asleep, when you can freely touch, re-prep, and reposition. Never rush the pre-sleep signal check. At lights-out you should have verified every channel is present and clean, impedances are within target (Chapter 6), and biocalibration confirmed each channel reads what it should (Chapter 8). A channel that is marginal at lights-out will almost certainly fail later, when fixing it means disturbing a sleeping patient. Fix it now. The disciplined technologist is strict at lights-out precisely to buy a quiet night.

Real-time troubleshooting during the study

Once the lights are out, the game changes: now every intervention risks waking the patient, so your judgment about whether and how to act becomes as important as the fix itself. You are constantly weighing "how badly is this signal degraded and how important is it?" against "how much will fixing it disturb the patient and the study?" Sometimes the right move is to correct a setting remotely; sometimes it's to wait for the patient to shift naturally; sometimes it's to enter quietly and re-seat an electrode. We'll return to that judgment call. First, the specific problems.

The common artifacts and how to handle them

60 Hz interference (electrical/line noise) shows up as a fast, uniform, regular oscillation riding on the signal, often across multiple channels. It comes from electrical sources and, importantly, frequently signals a grounding or impedance problem — a poorly connected or high-impedance electrode is more susceptible to it. Address it by checking the ground and reference, checking and improving impedances, identifying nearby electrical sources, and confirming connections. Because it often affects many channels, "60 Hz everywhere" points you to shared causes rather than one electrode.

Figure — Simplified waveform sketch showing a clean trace on the left transitioning to a rapid, evenly-spaced, uniform high-frequency oscillation on the right, labeled as 60 Hz interference.

Figure 7.1 An original, simplified illustration of the fast, uniform oscillation characteristic of 60 Hz (line) interference riding on a signal.

EKG (cardiac) crossover appears when the heartbeat's electrical signal bleeds into other channels (like EEG or EMG), showing up as regular, heartbeat-timed spikes. Because it's tied to the pulse, you can recognize it by its rhythmic timing that matches the EKG channel. It often relates to electrode placement and impedance; addressing connection quality and, per your lab's methods, electrode positioning helps reduce it.

Sweat artifact shows as slow, rolling, wandering baseline swings, usually when a patient is warm. It can obscure the underlying signal with a low, drifting wave. The real fix is often environmental and patient-comfort based: cooling the room or the patient reduces sweating. Because it tends to affect the channels where sweat is pooling, it can look regional.

Figure — Simplified waveform sketch showing the trace slowly wandering up and down in long, low, rolling waves, labeled as sweat artifact.

Figure 7.2 An original, simplified illustration of the slow, rolling baseline drift characteristic of sweat artifact.

Popping leads / electrode pop appears as sudden, sharp, isolated spikes or step-like jumps in a single channel, caused by an unstable electrode connection (drying gel, loosening contact). Because it's typically one channel, it points you straight to that electrode — re-prep and re-seat it.

Figure — Simplified waveform sketch showing an otherwise steady trace interrupted by sudden sharp vertical spikes and step-like jumps, labeled as popping lead artifact.

Figure 7.3 An original, simplified illustration of the sudden sharp spikes of a popping (unstable) electrode.

Loose electrodes and poor impedance produce noisy, unstable, high-amplitude, or wandering signals in the affected channel(s). They are the root cause behind much artifact (including susceptibility to 60 Hz). The fix is to re-prep the skin and re-seat or replace the electrode, restoring a low-impedance connection.

Now the sensor-specific failures:

Respiratory belt failure — flat, absent, or erratic effort signal from a chest or abdomen belt. Check that the belt is properly positioned, snug, and connected; a slipped or loosened belt is common. Because effort signals are central to classifying breathing events, restore them promptly.

Nasal pressure signal loss — the airflow trace from the nasal pressure transducer goes flat or weak. Causes include a displaced or blocked cannula, mouth breathing, or a disconnection. Check the cannula position and connection. (Note that a genuinely absent airflow may be a real apnea, not artifact — see "signal vs. event" below.)

Thermistor problems — the oronasal temperature airflow signal degrades or disappears; check position at the nose/mouth and the connection. Since the thermistor and nasal pressure sensor complement each other, comparing them helps you tell a sensor failure from a real breathing change.

Pulse oximeter dropout — the oxygen/pulse signal drops out or reads erratically, often from patient movement, a poorly positioned probe, or poor perfusion. Reposition the probe for a good trace. Because oximetry is a key safety and scoring signal, don't let it stay dropped.

Body position sensor problems — the position reading is stuck or clearly wrong relative to what the patient is doing (or the video). Check the sensor's attachment and orientation on the torso.

Leg EMG artifact — noise or instability in the leg channels, or movement that could be artifact versus a real limb movement. Check the leg electrode connections; distinguish true periodic movements from electrical artifact by their character and correlation with other signals.

Video and audio issues — loss or poor quality of the synchronized video or audio used to correlate events. Verify the equipment is running and capturing; video/audio is valuable context for interpreting events and correlating position and behavior.

PAP mask leak (during titration/split-night, Chapter 11) — an audible or measured leak degrades therapy and signals. Address mask fit and seal, adjust the mask or straps, and reassure the patient. A good pre-sleep mask fit (Chapter 6) prevents much of this.

Patient movement — gross body movement produces broad, high-amplitude artifact across many channels simultaneously. It's usually self-limiting; often the right response is to wait for the patient to settle rather than intervene, then confirm signals recovered.

Figure 7.4 A quick-reference guide to common artifacts and sensor failures: how each looks, its likely cause, and the first fix.

Signal problem versus patient event

This is one of the most important distinctions in the entire job, and getting it wrong cuts both ways. A signal problem (artifact) is the equipment misrepresenting reality — a loose lead, interference, a dropped probe. A patient event is the body genuinely doing something clinically meaningful — an apnea, a limb movement, an arousal, an arrhythmia. Your task in real time is to tell them apart, because the responses are opposite: you fix an artifact, but you observe, protect, and document a real event (you don't "fix" an apnea by adjusting a wire).

The danger runs both directions. Mistake a real apnea for a "sensor problem" and you might dismiss a genuine, even dangerous, finding. Mistake a bit of artifact for a real event and you corrupt the scoring with false positives. So how do you tell them apart?

• Correlate across channels. Real physiological events usually show a coherent story across multiple signals. A true apnea shows absent airflow plus continued or changing effort plus often an oxygen drop plus eventually an arousal. A sensor failure typically affects just its own channel without that coordinated pattern.

• Use the video and behavior. Does what you see on camera match the signal? A "leg movement" with a visible leg kick is real; flat oximetry while the patient is clearly moving is a probe issue.

• Check plausibility and timing. EKG crossover is timed to the pulse; 60 Hz is machine-fast and unphysiological; sweat drift is slow and rolling. Real events have physiological shapes and follow physiological logic.

• When unsure, don't destroy the evidence. If you can't immediately tell, note it, keep observing, and preserve the data rather than blindly "correcting" something that might be a real event. As a trainee especially, ask for help rather than guess when a patient's safety or the study's validity is at stake.

The habit that keeps you safe is always reading signals together, never in isolation. The whole picture almost always reveals whether you're looking at a body or a broken wire.

When to enter the room, and avoiding unnecessary waking

Every real-time intervention is a trade-off, and mature judgment here is a hallmark of an experienced technologist. Ask yourself: how important is the affected signal, how badly is it degraded, is it likely to self-correct, and how much will fixing it disturb the patient?

• Observe / wait when the artifact is likely self-limiting (gross movement settling), when the patient is transitioning and may re-seat a lead by moving, or when the degraded signal is non-critical and can wait for a natural awakening.

• Fix remotely when the problem is a setting, montage, or something you can address at the equipment without touching the patient — always your first choice.

• Enter the room when a critical signal (like oximetry or key EEG/airflow) is lost or badly degraded and won't recover on its own, or when patient safety or comfort requires it.

When you do enter, minimize disruption: move quietly, use minimal light, work efficiently, touch only what you must, and if the patient wakes, reassure them calmly and briefly (recall the communication skills from Chapter 6). The goal is a fix so smooth the patient barely stirs. Waking a patient unnecessarily costs sleep data and trust; but never let fear of waking them stop you from acting on a genuine safety concern. Safety always outranks a tidy sleep record.

Figure — Decision-tree flowchart starting with "Signal looks wrong" and branching through: is it a real patient event or artifact; one channel or many; can it be fixed remotely; is it self-limiting; is a critical signal lost; leading to actions of observe, fix remotely, or enter quietly, and always document.

Figure 7.5 A systematic decision path from noticing a bad signal to fixing it with minimal patient disruption.

Documenting troubleshooting

Undocumented troubleshooting is almost as bad as none, because the scorer and physician need to know what happened to the signal and why. Whenever you address a problem, record what you observed, what you believe caused it, what you did, when, and whether it resolved. This does three things: it lets the scorer correctly interpret a stretch of odd data, it explains any gap or switch (for example, moving to a backup electrode), and it protects both the patient and you. Tie this into the overall study record from Chapter 8.

Example tech notes:

"02:14 — O2 channel dropped to erratic values; patient moving in bed. Waited to settle; repositioned oximeter probe on left index finger at 02:19. Good trace restored. No true desaturation seen; artifact only."

"00:47 — F4 developed intermittent popping (single channel). Entered quietly, re-prepped and re-seated F4 at 00:52 with minimal patient disturbance. Impedance improved; channel clean. Scoring switched to backup F3–M2 for 00:47–00:52."

"03:30 — Diffuse slow baseline drift across multiple channels; patient warm/diaphoretic. Lowered room temperature; drift gradually resolved by ~03:55. Consistent with sweat artifact, not physiological."

Notice each note distinguishes artifact from real events, states the action and time, and confirms the outcome — exactly what a scorer needs.

How troubleshooting affects scoring and interpretation

Everything in this chapter feeds forward into Part III. A stretch of uncorrected artifact can be unscorable, forcing the scorer to skip data or guess. Mislabeled artifact can create false events that inflate an index; missed real events (dismissed as "sensor noise") can hide a genuine diagnosis. Clean signals, promptly corrected problems, and clear documentation are what make accurate scoring possible. In a very real sense, the quality of the diagnosis is capped by the quality of your troubleshooting. That is why this "invisible" skill quietly determines how good a technologist you are.

Clinical Takeaways

• Troubleshooting protects the entire study; the equipment records blindly, so you are the safeguard.

• Work systematically: confirm it's real, ask "one channel or many," check least-invasive causes first, then the patient side.

• Fix marginal signals before lights-out; it's far easier than fixing them over a sleeping patient.

• Always read signals together — a real event tells a coherent multi-channel story; artifact usually doesn't.

• Weigh every real-time fix against patient disruption: observe, fix remotely, or enter quietly — but never let a real safety concern go unaddressed.

• Document every troubleshooting action; it makes the data scorable and protects everyone.

Study Questions

1. Why is troubleshooting one of the most important active skills in the job?

1. Describe the systematic method for approaching a bad signal, including the "one vs. many" rule.

1. How do you tell 60 Hz interference, sweat artifact, and a popping lead apart by appearance?

1. Explain how to distinguish a real apnea from a nasal-pressure sensor failure.

1. What factors decide whether you observe, fix remotely, or enter the room?

1. Why must troubleshooting be documented, and what should a tech note include?

1. How can uncorrected or mislabeled artifact distort scoring in both directions?

Lab Reality Check

Troubleshooting is the skill that most makes you feel like a real technologist, because it's where you stop following a checklist and start thinking on your feet. Early on, a bad signal at 2 a.m. will spike your heart rate. With experience, you'll glance at it, ask "one channel or many?", and know within seconds whether you're looking at a loose F4, a sweaty patient, or an actual apnea you need to leave alone and document. You'll also develop the veteran's restraint: the wisdom to not charge into the room every time a signal wobbles, because half of them fix themselves when the patient rolls over. The technologists who struggle are the ones who either panic and wake patients constantly, or the opposite — who ignore a dying signal for three hours and hand the scorer an unscorable mess. Aim for the middle: calm, systematic, minimally disruptive, and always documenting. And never forget the one rule that overrides tidiness: if it's a real safety concern, act — a clean sleep record is never worth a patient's safety.

Explain Like I Am 10

Imagine you're recording a song, but sometimes the microphone acts up — it buzzes, or crackles, or goes quiet — and sometimes the singer actually does something different, like coughing. A sleep technologist has the same challenge all night: they have to tell the difference between the equipment messing up and the person's body actually doing something. Fixing those problems is called troubleshooting, and it's like being a detective for wiggly lines on a screen.

The detective has a few smart tricks. Trick one: ask "is one line messed up, or lots of lines at once?" If just one line is bad, the problem is probably that one sensor — maybe it came loose. If lots of lines go bad together, it's usually something they all share, like the room being too warm and making the person sweaty. Trick two: look at all the lines together like a team. If the person really stops breathing for a moment, several lines change in a way that makes sense together — that's a real body event, and you don't "fix" it, you write it down and keep watching. But if only one line looks weird and the rest are fine, it's probably just a wire being silly.

The trickiest part is being gentle. The person is asleep, and you don't want to wake them just to fix a tiny problem. So a good technologist fixes what they can from their computer without touching the person, waits to see if a problem fixes itself (a lot do when the person rolls over), and only tiptoes into the room when they really need to. But there's one big rule that beats all the others: if something is about the person's safety, you always help right away, even if it means waking them. A neat recording is never more important than a person being okay.

Remember This

• Troubleshooting means figuring out if the equipment is broken or the body is doing something real.

• Ask "one line or many?" — one bad line = one sensor; many bad lines = something shared.

• Look at all the signals as a team; a real body event makes sense across several lines at once.

• Try to fix problems without waking the sleeping person, and let some fix themselves.

• Safety always wins — if the person could be in danger, help right away.

Quick Review Questions

1. What's the difference between the equipment messing up and the body doing something real?

1. If only one line on the screen looks bad, where's the problem probably coming from?

1. Why do you look at all the signals together instead of just one?

1. Why does a technologist try not to wake the sleeping person for small problems?

1. What's the one thing that's always more important than a neat recording?

Common Mistakes

• Reacting randomly instead of systematically. Yanking wires and waking patients before reasoning.

• Skipping a strict pre-lights-out check. Marginal channels always fail at the worst time.

• Reading a channel in isolation. It's how you mistake artifact for events and vice versa.

• "Fixing" a real event. Adjusting a wire because a true apnea looked like a sensor problem.

• Waking the patient unnecessarily. Intervening when the artifact would have self-corrected.

• Failing to document. Leaving the scorer to misread a stretch of odd data.

• As a trainee, guessing instead of asking when safety or study validity is at stake.