Computer Repair · October 28, 2026 · ComputerKing Team

Second-Hand Laptop Passed Inspection but Fails at Home? Hidden Faults Your On-Site Check Cannot Catch, Plus a First-Week Stress Test

Passing an on-site inspection only means the machine was fine at that moment, in that place, under that light load: intermittent, load, time, environment, and data blind spots mean many faults only appear at home. In week one, record a baseline, then test charging, sustained load, sleep and wake, external devices, and real daily use day by day, and decide on day seven from the records, moving important data only after the condition is confirmed.

Key Summary

• Passing an on-site inspection only means the machine was fine "at that moment, in that place, under that light load." It does not mean it will stay fine once you use it at home every day.
• The faults an on-site check misses usually have five causes: they appear intermittently, they need load to appear, they need time to appear, they need a different environment to appear, or they only show up in data.
• Do not move all your important files onto the machine in week one: record a baseline first, then test charging, sustained load, sleep and wake, external devices, and real daily use, day by day.
• Keep evidence for every test result: battery reports before and after, photos, error screenshots, and your message history with the seller. Without evidence, it is hard to say clearly when a problem appeared.
• The earlier article #93 covered the on-site inspection checklist (appearance, battery, screen, accounts) and is not repeated here. This article covers the gap: what the on-site check cannot catch. All repair and inspection prices are quoted individually.

Introduction

You checked everything in front of the seller: it powers on, the screen has no dead pixels, every key works, the battery report looks fine, the accounts are signed out, and the whole checklist is ticked before you pay and take it home. Then, after two or three days at home, problems start appearing one by one: it shuts down suddenly at half battery, it slows down badly when you open several programs, Wi-Fi does not reconnect after waking from sleep, and an external drive is recognised sometimes and not others. Many people's first reaction is that they missed something during inspection, or that they were cheated.

Very often, neither is true. An on-site inspection and long-term use at home are two different tests. That half hour in front of the seller usually means mains power, a light load, a stable room temperature, and a few power cycles. At home, you run on battery all day, keep dozens of browser tabs open, wake the machine from sleep dozens of times, and use your own network and external devices. Many problems only appear under the second set of conditions, so no amount of care during the on-site check can catch them. That is not a failure of your inspection. It is a blind spot built into on-site inspection itself.

The earlier article already explained how to inspect a machine before buying (site article #93, "Second-hand laptop pre-purchase inspection checklist"), covering appearance, hinges, ports, keyboard, trackpad, battery health, screen dead pixels, and account locks. This article does not rewrite that checklist. It covers one thing only: which problems an on-site check cannot catch, why it cannot catch them, how a first-week stress test at home can force them out one by one, and how to read the results before deciding whether to accept the machine, keep watching it, get it inspected, or contact the seller.

1. Why an On-Site Check Misses Things: Five Blind Spots

Blind Spot 1: Intermittent Problems

Some faults are not always present. A loose display cable may only break contact when the lid is opened to a certain angle. A loose port may only disconnect when the cable is nudged. An ageing Wi-Fi module may only start dropping out after several hours of continuous use. A check lasting a few minutes may simply never hit that angle or that moment. The awkward part of intermittent faults is that they are not "present or absent" but "how often they appear." The shorter the test, the lower the chance of catching one.

Blind Spot 2: Problems That Need Load

Under a light load and under a heavy load, the same machine behaves like two different machines. Typing a document barely uses the processor or graphics, the fan turns slowly, and temperatures stay normal. Add a video call with screen sharing, a large spreadsheet, dozens of browser tabs, or a video export, and the processor runs at full speed for a long time. Only then do weak cooling, an ageing fan, or dried thermal paste show themselves: a hot body, a fan running at full speed and getting loud, and the system slowing itself down. An on-site check that only powers on the machine and types a few words never reaches that load level.

Blind Spot 3: Problems That Need Time

Battery faults are the clearest example. An on-site battery report may show a small gap between design capacity and full charge capacity, and a cycle count that does not look high, so the battery looks healthy. But battery ageing is not always even wear. One weak cell can fail early: on mains power everything looks normal, but on battery the charge can fall off a cliff at a certain percentage, or the machine can shut down outright. That only shows up after real full charge and discharge cycles, and half an hour on site is not enough to complete even one.

Blind Spot 4: Problems That Need a Different Environment

Your home network, external devices, power strip, and daily habits are all different from the seller's. Wi-Fi that worked perfectly on the seller's network may prove unstable on your own router, with your own channel settings, Bluetooth earphones, and mouse. A charger that sat firmly in the seller's socket may prove loose on your own power strip or USB hub. When the environment changes, every variable changes with it, so "normal on site" cannot simply be carried home.

Blind Spot 5: Problems That Only Show in Data

Some problems cannot be seen with the naked eye and only appear in system data. The health of a storage drive is not proven by the fact that it powers on and opens files. A battery report is not a single number: its capacity history matters, and whether capacity fell gradually or suddenly matters. A lack of error pop-ups does not mean the system log is clean. On-site checks mostly look at surface results, with no time and no reason to compare data line by line, so gradual decline stays hidden until it surfaces at home.

Table 1: Five Blind Spots Compared

2. The Five Most Common Hidden Faults

1. Battery: Fine on Power, Fails on Battery

The signature is unmistakable: perfectly normal on mains power, trouble on battery. The charge display may fall suddenly from 40 percent to 10 percent, or shut the machine down at 20 percent, then jump back to a higher number once plugged in again. That does not mean the battery report lied. The report reflects static capacity, not real performance under load. The fix is several full charge and discharge cycles, watching whether the charge falls smoothly or drops off a cliff at a certain point. How to read a MacBook cycle count and status, and how to generate a Windows battery report with `powercfg /batteryreport`, were both covered in article #93, section 4, and are not repeated here. The point here is "before and after": generate one report on the day you bring it home, and another at the end of week one, and read the two side by side.

2. Cooling and Throttling: Slow Only When Hot

Cooling problems show up as "fine at first, slow later." The first half hour is smooth, then after two hours of continuous use the fan gets louder and louder, the body gets hot, programs respond more slowly, and in severe cases the machine shuts down to protect itself. That is not "a slow machine." It is the system reducing speed because temperature has reached a critical point. The test is not typing a few words after powering on, but a planned sustained session using the work combination you actually use, for a long stretch, watching whether fan noise, heat, and slowdowns appear together. A loud fan with no slowdown is a different situation from a loud fan with a clear slowdown: the first may only mean the cooling system is working hard, while the second is worth further inspection. Any work that involves opening the machine, cleaning dust, or replacing thermal paste must be left to a professional. Prices are quoted individually.

3. Storage: Gradual Decline

Storage problems rarely mean the drive is completely unrecognised on day one. More often they are gradual: a file opens slowly now and then, copying a large file pauses now and then, a read error pops up once and then goes away. Those signals are easily dismissed as "slow network" or "a program problem." In week one, watch for the patterns: the same file opening fast today and slowly tomorrow, copy speeds that stall repeatedly, and any storage-related warnings from the system. The most important rule is: do not put your only copy of important data on this machine in week one. Keep a backup first. If storage shows suspicious signals during testing, stop moving important data onto it at once. Data recovery and storage inspection prices are quoted individually.

4. Intermittent Connections: Angle, Time, and Sleep and Wake

This covers three common cases. First, cables and ports: opening the lid to a certain angle makes the screen flicker once, or nudging a plugged-in device breaks the connection. Second, Wi-Fi and Bluetooth: the connection stays fine for hours and then starts dropping, or fails to reconnect after waking from sleep until you toggle it off and on. Third, sleep and wake itself: the machine fails to wake, wakes to a black screen, or wakes with the fan at full speed but never reaches the desktop. The test method is repetition: open and close several times, plug and unplug several times, sleep and wake several times, and record each result. One success does not mean every attempt succeeds. Intermittent faults are caught by repetition.

5. Software and Accounts: Only Visible After Reinstall

On site, the machine may still be signed into the seller's account, or running a system the seller set up, and everything looks perfectly normal. Once you sign the seller out, sign yourself in, or reinstall the system, things that were invisible before can appear: a management setting that was never removed, an account lock that was never cleared, or management software left in a hidden partition. Article #93, section 6, already explained that the seller must sign out and remove all locks in front of you before the sale. This article adds the at-home recheck: after reinstalling or signing in with your own account, check account status and management settings again to confirm nothing is left behind. If the previous owner's account still appears after a reinstall, that is a stop signal. Do not try to work around it yourself. Ask the seller, and get it inspected if needed.

3. The First-Week Stress Test: A Seven-Day Plan

The principle is "baseline first, then load, then daily life." Day one is not heavy load; it is record-keeping. The middle days cover charging and sustained load. The last days return to real daily use, to see whether problems appear in your real combination of work. Record the result each day. Do not rely on memory.

Table 2: First-Week Test Schedule

Day 1: Baseline Record. Do Not Skip It

On the first day, before installing software, record the baseline. Photograph the serial number, generate a battery report, note the system version and storage capacity, and photograph the appearance. These records serve two purposes. First, if a problem appears later, you can compare whether it was already there when you brought the machine home or only appeared during week one. Second, when talking to the seller, you can lay out a clear timeline instead of arguing from memory. A baseline does not need to be complicated: a few photos plus one report is enough, but it must be done on day one. Done late, it is no longer a baseline.

Day 2: Charge Cycle. Watch the Curve

Run the battery from full charge to low charge, then charge it back to full: one complete cycle. What matters is not the single number of "how many hours it lasted" but whether the charge falls smoothly. A smooth fall is normal ageing. A sudden cliff at a certain percentage is a suspicious signal. Do not run a heavy load during this test, or you will not be able to tell a battery problem from a load problem. While charging, watch temperature: slightly warm is normal, too hot to touch or any unusual smell is an immediate stop signal (see the stop section below).

Day 3: Sustained Load. Force Cooling Problems Out

Use the work combination you actually use, continuously, for a long stretch. Do not run benchmark software you would never normally use just to "test" the machine. Benchmark numbers do not represent daily experience; your real combination represents what you bought the machine for. Watch whether fan noise, body heat, and slowdowns appear together, and note how long it takes before slowdowns begin. If it is always smooth for the first half hour and clearly slow after two hours, that pattern itself is important evidence.

Day 4: Sleep and Wake. Repetition Catches Intermittent Faults

Sleep and wake problems often cannot be caught in a single try. Sleep, wake, sleep again, wake again, several times. Each time, check whether it wakes immediately, whether the screen comes back at once, whether Wi-Fi reconnects by itself, and whether the fan runs at a normal speed. Record which attempt failed and what you did before it. That repeated record is the evidence for an intermittent fault. Without a record, it is very hard to explain.

Day 5: External Devices and Network. One at a Time

Do not swap everything at once on this day. Connect one thing at a time: first your home router, then your external drive, then your Bluetooth earphones and mouse, testing each one for a while before connecting the next. Whichever item triggers the problem as soon as it is connected is the suspect. Swap everything at once and you will not know which one caused it, and you will have to disconnect everything and test again, wasting time and risking damage to ports from repeated plugging.

Day 6: Real Daily Work. Do Not Work Around the Machine

No deliberate tests on this day. Use the machine for a full day of real work. Deliberately working around the machine (for example, deliberately opening fewer programs, or deliberately staying on mains power) hides problems. Real daily work shows the true experience. This day's record is the most valuable of all, because it reflects not a test environment but the real purpose you bought the machine for.

Day 7: Summary Decision. Do Not Delay

On day seven, compare the baseline with the week-one record and decide using the decision table in the next section. The longer you delay, the vaguer the timeline becomes, and the harder it is to explain things clearly to the seller. Even if the decision is "accept," write down why you accepted and what minor issues you already know about. That becomes the starting point for any future comparison.

4. How to Record Evidence That Actually Helps

Evidence is not about quantity. It is about timeline and conditions. A useful record answers four questions: when did it appear, what were you doing when it appeared, what did you do just before it appeared, and what made it go back to normal. For example: "Wednesday at 3pm, after two hours of a video call with screen sharing, the fan ran at full speed and programs slowed noticeably, and ten minutes after ending the call it returned to normal." That record is far more useful than "the machine is slow."

The method is simple. Every time something happens, photograph the error message, note the time and what you were doing, and note whether restarting or closing a program helped. Keep the battery report from day one and the one from the end of week one: only the pair, compared, has meaning. Keep the receipt, the message history, and the serial-number photo from the sale. Article #93, section 9, already explained why keeping documents matters. This article adds that the test record itself is a document. Do not rely on memory.

One point must be stated clearly: test records are for explaining a problem, not for fixing the machine. After finding a problem, do not open the machine yourself and do not try fixes found online. Opening the machine, replacing parts, and deep inspection must be left to a professional, with prices quoted individually. Once you have opened it yourself, it becomes very hard to separate who is responsible for what.

5. Reading the Results: A Four-Way Decision Table

Table 3: Test Result Decision Table

When deciding, separate "a problem exists" from "a problem needs immediate action." Minor ageing is normal for a second-hand laptop. Not every sign of ageing means a return or a repair. The test should be: does it keep happening, does it affect your real use, and is it getting worse. A one-off signal that never repeats and does not affect work can be watched. A repeated signal that affects work and shows a worsening trend should not be delayed.

When talking to the seller, present the timeline and the test records, not feelings and guesses. For example: "The battery report on the day I brought it home looked like this. On day three of week one, during a sustained-load test, this happened, and the record is here." That is far easier to handle than "it was already broken when you sold it to me." Whether the seller offers a return or exchange, and who provides any warranty, must be asked before the sale. Article #93, section 9, already covered this. The point of the week-one test is to find problems while those arrangements are still in force, not after they expire.

6. Common Mistakes: Five Errors in Week One

The first mistake is moving all your important data onto the machine on day one. During testing, the machine's condition is not yet confirmed. Putting your only copy of important data on an unconfirmed machine is a large risk. The right order is: test with non-important files in week one, and only move important data after the condition is confirmed, with a backup made first.

The second mistake is testing for a whole week on light load only. Typing and browsing all day will never reveal cooling or battery-load problems. Problems only appear once real use begins, by which time the timeline has already stretched. Week one must include one planned sustained load. Do not leave it to chance.

The third mistake is noticing a warning signal and not recording it. Intermittent problems are most afraid of missing records. It appears once today, looks normal tomorrow, and a few days later you no longer remember what you were doing. Record it immediately, every time. Do not think "I will record it next time it appears." The next time may be a long time away.

The fourth mistake is changing every environmental variable at once. When something goes wrong, you cannot tell which change caused it, and you have to disconnect everything and test again, wasting time and risking port damage from repeated plugging. Change one thing at a time and test each one.

The fifth mistake is opening the machine yourself to inspect it. Seeing a cooling problem and opening the machine to clean dust, or seeing a battery problem and opening the battery, carries safety risks and makes responsibility hard to separate afterwards. Opening the machine, replacing parts, and deep inspection must be left to a professional, with prices quoted individually.

7. When to Stop Immediately

Stop testing, stop charging, and stop using the machine at once, and get it handled professionally, in any of these cases: a swollen battery (bulging body, raised trackpad, deformed keyboard), an unusual smell or burning smell, smoke, a damaged power cable or plug, a body that stays too hot to touch, or persistent storage noise with important data not yet backed up. Never poke or press a swollen battery, and never keep charging it while testing. Article #93 stressed this repeatedly, and it is repeated here because week-one testing is exactly when people are most tempted to ignore a safety signal in order to "test it properly." In front of a safety signal, the test plan stops immediately. No test result is worth a safety risk.

Frequently Asked Questions

Q: If the on-site inspection passed, does that mean the machine definitely has no problems?

A: No. An on-site inspection only covers that moment, that place, and that light load. Intermittent problems, load problems, time problems, environment problems, and data problems can all wait until you get home to appear, which is why a first-week stress test is needed as a second layer of checking.

Q: Does the first-week test need a lot of paid test software?

A: No. Your real work combination is enough to force cooling and battery problems out. Benchmark numbers do not represent daily experience. Your daily combination represents the real purpose you bought the machine for. Do not test for the sake of testing.

Q: The battery report looked normal on day one. Why does it shut down suddenly on battery at home?

A: Because the report reflects static capacity, not real performance under load. When one cell ages early, mains power looks completely normal, but on battery the charge can fall off a cliff at a certain percentage. Several full charge and discharge cycles, watching whether the charge curve is smooth, are needed to reveal this kind of problem.

Q: How long should a sustained-load test run?

A: There is no single number. The standard is "until your real work combination shows a change." If you normally spend two hours in video calls, the test should run for two hours, not ten minutes. Watch whether fan noise, body heat, and slowdowns appear together. That pattern itself is evidence.

Q: If a sleep and wake test passes once, does that mean there is no problem?

A: No. Intermittent problems are caught by repetition. One success does not mean every attempt succeeds. Sleep, wake, sleep again, wake again, several times, recording each result, is the only way to say clearly whether a problem exists.

Q: When a test finds a problem, should I contact the seller first or get it inspected first?

A: Both, in the right order. First stop using the machine if there is any safety risk. Then get it inspected, bringing the test records with you, and contact the seller at the same time with a clear timeline. Inspection prices are quoted individually. Whether the seller offers a return or exchange must be asked before the sale.

Q: If the first-week test passes, does that settle everything for good?

A: No. The first-week test is a second layer of checking, not a promise that lasts forever. A second-hand laptop is still a used machine. Keep watching for changes afterwards, and check the battery report from time to time. The week-one baseline record is the starting point for all future comparisons.

Q: Which situations mean stopping immediately, with no further testing?

A: A swollen battery, an unusual smell, smoke, a burning smell, a damaged power cable or plug, a body that stays too hot to touch, or persistent storage noise with important data not yet backed up. In these cases, stop using the machine at once, do not charge it again, and do not keep testing. Get it handled professionally. Never poke or press a swollen battery.

Conclusion

Passing an on-site inspection is worth celebrating, but not worth complete confidence on its own. The on-site check tests "that moment." The first-week stress test tests "after you bring it home." Together, the two layers make a complete second-hand laptop check. Remember the order: day one records the baseline, day two runs a charge cycle, day three runs a sustained load, day four repeats sleep and wake, day five connects your home devices one at a time, day six is a full day of real work, and day seven compares the records and decides. Record the timeline and conditions for every result. Minor ageing can be watched. Repeated problems that affect work should not be delayed. In front of a safety signal, stop at once. Buying a second-hand laptop need not be frightening. The key is not to rush all your important data onto it, not to mind the trouble of one extra week of testing, and to have records when a problem appears, so it can be explained clearly and handled quickly.

Sources

• Site article: Second-hand laptop pre-purchase inspection checklist - https://www.computerking.hk/blog/second-hand-laptop-buying/
• Site article: MacBook Battery Health: How to Read Cycle Count, Maximum Capacity and Status Together - https://www.computerking.hk/blog/macbook-battery-health-guide/
• Apple official support document, "Determine battery cycle count for Mac notebooks" (as cited in article #93)
• Microsoft Support, "Caring for your battery in Windows": https://support.microsoft.com/en-us/windows/experience/power-battery/caring-for-your-battery-in-windows
• Consumer Council press release on post-discontinuation repair service periods: https://www.consumer.org.hk/en/press-release/20120116-3

*This article is general information only. Actual steps depend on the individual model, and the manual for your own machine takes precedence. Anything involving heat, live power, or opening the machine should be left to a professional. Repair and inspection prices are quoted individually.*