Building a test kit, and knowing what each tool proves

Why this matters

This is the last lesson of the course and the capstone of its largest domain. Everything before it described a tool or a fault; this one is about carrying the right things and knowing, before you reach for any of them, what the result will mean.

That second part is the point. A technician with a bag of instruments and no model of what each proves runs tests and collects results. A technician who can say "this test will eliminate the cable and the port, and nothing else" is doing diagnosis. The exam's scenario questions are written to test the second kind of thinking: given a symptom, which tool, and what does the result tell you.

It is also the lesson that turns the course's recurring theme into a working habit. Every check has a boundary; a pass answers its own narrow question and no other; and knowing the boundary is what stops a clean result being mistaken for a clean machine.

The lesson

The kit that fits in a bag, and what each item is there to rule out

Measurement

• Multimeter — voltages, continuity, fuses, switches. Rules out power delivery and broken conductors. Lesson 42.
• Power supply tester — rails plus power good, in seconds. Rules out a dead supply. Lesson 43.
• Cable tester — wire map. Rules out termination faults. Lesson 44.
• Tone generator and probe — identifies which cable is which. Not a fault finder; an identifier.
• Loopback plug — rules out the network adapter.

Hand tools

• Magnetic screwdriver set including Torx and small Phillips.
• Plastic prying tools and spudgers, for anything with clips.
• Tweezers, a torch, and a small brush.
• Cable ties and labels, which are the preventive half of the job.

Consumables and spares

• Thermal paste, isopropyl alcohol and lint-free cloths.
• Compressed air or a low-static blower.
• Spare screws, standoffs and a coin cell.
• A short known-good patch lead, and one display cable of each common type.

Protection

• Anti-static strap and mat, used everywhere except near mains.
• Gloves and eye protection for toner and for anything being cut.

Software on a drive

• A bootable operating system for testing and recovery.
• Memory test, drive health, and a stress-test tool.
• A partition and imaging tool.

The organising principle: each item is there to eliminate a branch of the diagnosis, and anything that eliminates nothing is weight.

Spares as diagnostic tools: known-good parts and how to keep them known-good

Substitution is the most reliable test available, and it requires something to substitute with. A small set of known-good parts turns many multi-visit faults into single-visit ones:

• A power supply, which is the highest-value spare because it is the most likely failure and the hardest to test properly.
• A memory module, which tests both a suspect module and a suspect slot.
• Patch leads and display cables, which are cheap and account for a large share of faults.
• A known-good drive, useful for proving a machine works with a clean system.
• A spare mouse and keyboard, which eliminate the input device in seconds.

The discipline is in the words known good, and it is easy to lose:

• Test a spare before it goes in the bag, and again after it has been used in a failed machine.
• Label it and keep it separate from parts pulled out of faulty machines. A "spare" that is actually a part removed from a dead machine is worse than no spare at all, because it contaminates every test it takes part in.
• Retire a spare that has ever behaved oddly. Certainty is the entire value; a doubtful spare has none.

This is the physical form of the rule that runs through this repository: an instrument must be shown to work before its results are believed. A spare that has not been proven is not a test.

Software tools that belong on a technician's drive alongside the hardware ones

A single bootable USB drive covers most of what the hardware kit cannot.

• A bootable operating system — a live environment that runs without touching the installed system. Proves whether a fault is hardware or software in one boot, which is one of the highest-value tests available.
• Memory diagnostics, run for hours rather than minutes, from lesson 19.
• Drive health and SMART reading, from lesson 37.
• A data copying tool that skips errors and logs what it could not read.
• Stress tests for processor and graphics, which force thermal and power faults into the open — lesson 36's forcing technique.
• Hardware inventory tools that report the exact board, processor, memory configuration and firmware version, so an order is placed against facts.
• Partition and imaging tools, for capturing a drive before working on it.

Two cautions. Keep the drive updated, because a two-year-old live environment may not support recent hardware. And know what each tool writes — a diagnostic that modifies a disk is not appropriate on a machine whose data has not yet been copied, which is lesson 37's ordering rule applied to the toolkit.

Recording results on site so a second visit starts where the first ended

Lesson 34 made documentation a step. Here is the field version, written while standing at the machine.

Record, as you go:

• The symptom as observed, with the exact error text or a photograph.
• Every measurement, with its conditions and the expected value.
• Every substitution, and whether it changed anything.
• What was eliminated, which is the part with the most value to the next person.
• Serial numbers and firmware versions of anything replaced or suspected.
• What is outstanding, including any temporary measure left in place.

Why it pays, restated concretely: a second visit that begins with "the supply measured 11.2 V under load, memory tested clean overnight, the fault persisted with a known-good graphics card" starts at step three of the method. A second visit that begins with "it was doing something odd" starts at step one, and the customer pays for the same hour twice.

The related habit is photographs. Cable routing before disassembly, the board layout before a component comes out, the screen at the moment of failure. They cost nothing and they answer questions nobody thought to ask.

Knowing which tool to reach for from the symptom, which is the examinable skill

The whole course, compressed into a mapping.

Two closing principles, and they are the ones worth carrying past the exam.

Prefer the test that eliminates the most for the least effort, which is frequently not the test of the most likely cause. Swapping a cable before opening a machine is not laziness; it is ordering by cost.

Know the boundary of every result. A cable that passes a wire map, a drive that passes SMART, a supply that passes a tester, a port that passes a loopback — each of those has answered one question. Treating any of them as a clean bill of health is how a fault survives a full investigation, and recognising it is the difference between running tests and diagnosing faults.