Integrated Troubleshooting & Field Application

When there is more than one fault

Multi-fault calls break pattern matching completely, because the combined symptom matches nothing. The method handles them naturally: each stage isolates one system, you fix what that stage condemned, then you re-run the sequence and see what is still wrong.

The discipline is refusing to fix two things at once. Repair, confirm, re-run, then diagnose the remainder. If you change two variables, you learn nothing from the result and you cannot bill either repair with confidence.

• Never change two variables between confirmations
• Re-run the full sequence after every repair — masked faults surface there
• A machine with a history of neglect usually has more than one problem

Intermittents

Intermittents are condition-dependent faults: they need heat, vibration, load, humidity, or elapsed time. The winning move is to identify the condition rather than to keep testing at room temperature with the panels off. Ask when it happens, then reproduce that state — run the machine loaded, warm the suspect area, flex the harness, or let it sit through the interval that precedes the failure.

Thermal and connection intermittents dominate. A connection that reads fine cold will drop voltage under load once resistance rises with temperature, which is why measuring under load is worth ten resistance checks.

If you cannot reproduce it, say so, document what you tested and ruled out, and set the machine up so the next visit starts with evidence rather than from zero.

Working the method under pressure

Commercial calls come with a manager standing behind you and revenue stopping. The temptation is to skip to a part swap because it feels faster. It is not: a wrong part is a second visit, a second parts cost, and a customer who now doubts you.

What you can compress is communication and sequencing, not stages. Tell the customer what you are doing and what each result will mean, get the machine running in a limited safe mode where that is possible, and order parts as soon as the evidence names them rather than at the end of the visit.

The service note that earns the next job

A professional note states the complaint as verified, the measurements taken with conditions, the confirmed cause, the repair performed, and the post-repair confirmation. It should also flag deferred items: the ageing damper, the loaded water filter, the marginal supply voltage.

This is what separates a technician from a parts changer commercially. It gives the customer a reason to trust the invoice, gives the next technician a starting point, and gives you a record if the repair is ever questioned.

• Complaint verified, measurements with conditions, confirmed cause, repair, confirmation
• Flag deferred findings explicitly — they become the next approved job
• Photograph findings; a picture of a corroded terminal ends every argument

Faults that cross system boundaries

Single-system faults are what training material usually shows: one component, one measurement, one part. Real calls are frequently cross-system, where a fault in one domain expresses itself as a symptom in another. A restricted condenser raises head pressure, which raises current, which heats a marginal connection, which drops voltage to the control, which resets the machine mid-cycle. The customer reports a machine that keeps restarting. Every one of those links is real, and only one of them is the fix.

The way to keep the thread is to place each observation in a domain — electrical, sealed system, water and drainage, mechanical, control and sequence — and then ask which domain could produce the observations in the other domains. The domain that explains the most observations is the one to test first. A domain that explains only one observation is a consequence.

This is also where equipment age and site conditions earn their place in the diagnosis. Neglected machines rarely have one fault, and a site with marginal supply or poor water quality produces the same cross-system pattern across every machine on the premises.

• Sort observations by domain before you form a theory
• The correct cause explains the observations in the other domains too
• One symptom explained by a theory is a coincidence; three is a diagnosis

When the obvious symptom is not the root cause

The loudest symptom is usually the last link in the chain, because it is the one that finally stopped the machine. Working backwards from it is natural and frequently wrong. A burned relay contact is a symptom of a load drawing too much or a connection running hot. An open thermal cut-out is a symptom of restricted airflow. A failed compressor start capacitor on a machine less than two years old is a symptom of supply quality or of the compressor itself.

The professional test is simple to state and hard to apply under pressure: for every failed component you find, name the condition that caused it to fail, and measure that condition. If you cannot name the condition, you are not finished. If you can name it but choose not to measure it, you have decided to accept a return visit.

There is a second, quieter version of the same error: fixing the first fault you find and stopping there, when the machine has two. Re-running the full sequence after every repair is what surfaces the second one while you are still on site.

• Every failed component gets a named, measured cause
• Re-run the full cycle after each repair — masked faults surface there
• The loudest symptom is usually the last link, not the first

If the service note says what you replaced but not what damaged it, the repair is incomplete regardless of how the machine ran when you left.

Testing order when two systems both look guilty

When two domains are both plausible, choose the test that separates them, not the test that confirms your favourite. Between an electrical and a sealed-system theory on a refrigeration call, verify supply and current draw first: it is minutes, non-invasive, and it either implicates or clears the electrical side outright. Between a control theory and a load theory, command the output and read at the board terminal and at the load: one procedure splits the whole circuit.

Sequence the irreversible tests last, always. Gauges on a system that has not yet had its electrical and airflow sides verified is the classic order error on commercial refrigeration, and it converts a diagnosable machine into an argument about charge.

Under commercial pressure the compression you are allowed is in communication and parts logistics, not in stages. Order the part the moment the evidence names it rather than at the end of the visit, and tell the customer what each pending result will mean so the wait is not silent.

• Prefer the test that separates two theories over the test that confirms one
• Non-invasive before invasive; reversible before irreversible
• Gauges last on refrigeration, after electrical, airflow and controls

A worked example, end to end

Call: a reach-in cooler runs continuously and holds 8 °C instead of 3 °C. The obvious theory is low charge. Domain sort first: control and sequence — the unit never satisfies, so it never cycles off; electrical — supply measured 231 V and compressor current sat 18% above nameplate; sealed system — suction line warm, condenser hot to the hand; water and drainage — drain pan dry, no ice; mechanical — condenser fan running, evaporator fan running.

Current above nameplate with a hot condenser explains the sealed-system observations without requiring a charge fault, so the electrical reading is a consequence of high head, not a separate problem. Cleaning the condenser dropped current to nameplate and head pressure to a normal range. Re-run: the unit pulled down to 4.5 °C and cycled — better, but not right.

Second pass on the remaining deviation: the door gasket had a compressed section at the hinge side, and the box was gaining load continuously. Replacing the gasket brought the box to 3 °C with a healthy cycle rate. Two faults, both real, and gauges were never fitted. The charge was correct the entire time.

• Domain sort, then the theory that explains the most observations
• Repair, confirm, re-run — then diagnose whatever is left
• A partial improvement is evidence of a second fault, not of a wrong diagnosis

Deciding with incomplete information

Real calls give you a fraction of what you would like: a vague complaint, no history, a machine that has been reset, and a customer who needs an answer now. The method still works, because it does not require complete information — it requires the next test to be the one that eliminates the most.

When information is thin, buy it cheaply. Ask three questions that change the diagnosis: when did it start, what changed around then, and does it happen at a particular time of day or load. Then take the two measurements that are cheap and broad — supply under load, and stored fault history in order. Those five pieces of information routinely halve the search space before a panel comes off.

It is also professional to state the limits of what you know. 'I have eliminated the electrical side and the airflow; the remaining possibilities are these two, and this test distinguishes them' is a stronger position with a customer than a confident guess, and it holds up if the second visit is needed.

• When did it start, what changed, and when does it happen
• Supply under load and stored faults in order — cheap, broad, always worth taking
• Stating what you have eliminated is stronger than asserting what you suspect

Failure modes and what confirms them

Test procedures

Manufacturer differences

Safety and professional boundaries

• Lock out and tag out before working inside commercial equipment that a building control or remote schedule can energise without warning.
• Never defeat an interlock or safety limit to progress a diagnosis, including on a machine you intend to condemn.
• Sealed-system work stays within your certification. If the correct next test requires recovery or brazing and you are not qualified for it, stop and escalate.
• Where a site condition — supply, ventilation, water — is the root cause, document it and refer it to the responsible trade rather than working around it.
• If the machine cannot be tested safely in the state the fault requires, say so on the note and escalate. Improvised access is the leading cause of technician injury on integrated calls.

Expected readings and what they mean

Field scenarios

Knowledge check