A decade old PC, years of Windows 11 startup crashes, and the 1 pound part that fixed everything

Started by CosmicRay65, Jul 04, 2026, 08:54 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Topic: A decade old PC, years of Windows 11 startup crashes, and the 1 pound part that fixed everything   Views(Read 222 times)

CosmicRay65

Posting this partly as a confession and partly as a public service, because if it saves one person the journey I went on it was worth typing. My PC is over a decade old and ever since I upgraded it to Windows 11 it crashed regularly on startup. Some days it would get a quarter of the loading circle on the Windows logo, other days just a black screen of death, and then the daily ritual of Windows attempting recovery only to crash again at setting up repair

I tried everything, and I mean everything. BIOS update, cleaned the whole machine inside and out, swapped the old HDD for a new SSD, added more RAM, added more fans in case it was thermal, full Windows reinstall, system restore, the lot. Every fix bought a little hope and none of it stopped the crashes, and at times it was genuinely painful, you start doubting the whole machine

The culprit, in the end, was the CMOS battery. A part that costs about one pound, in a PC I have owned for ten years. The cruel twist is I had a couple of spare batteries in the drawer and tried those too, and it turns out they were ALSO dead, which is exactly what threw me off the scent for so long and then finally told me what the problem was
IMG_0573.jpeg
The working battery that solved it came out of a clicker I use for taking pictures on holiday, of all places. Dropped it in, and no crashes since, no more daily recovery attempts, no quarter circle of doom, just a boring ordinary boot every single time. Pictures of the patient and the tiny hero attached below

So the lesson for the thread, when you have ruled out the software side and it is not Windows update related, and you have ruled out the normal hardware suspects, check the stupid cheap thing before you spend another penny. I now have a new battery, a working PC, and a fortune saved on the replacement machine I very nearly bought. Anyone else been defeated for months by a part that costs less than a coffee? :-X

IMG_0572.jpeg

Myles95

The dead spares are the killer detail, you did the right test and the test lied to you. That is the kind of luck that adds months to any diagnosis
Football is life. Everything else is just details.

Grim Tracey

Classic CMOS gotcha, when it dies the board forgets its settings every power off, and if it had anything non default saved, boot order, memory profile, the whole startup goes haywire. The pound coin battery takes down the thousand pound machine

Leah_68

Genuine question, did you not notice the clock resetting or BIOS settings going back to default? That is usually the tell tale before the crashes start

SlowSocket

Not always, if the machine stays plugged in at the wall the battery can be half dead for ages and only fully show itself in weird intermittent ways. The sneaky failures are the half dead ones
All original content unless stated

RayOfLight32

The battery from the holiday clicker is the best part of this story, somewhere a little camera remote made the ultimate sacrifice for the greater good

SpinState52

Been there with a different cheap villain, weeks of blue screens that turned out to be one SATA cable that had gone bad. The cheap parts fail quietly and let the expensive ones take the blame
COYB — you know who you are

Georgia67

In fairness the journey was not wasted, you came out the other side with an SSD, more RAM and better cooling, the crash was accidentally a full upgrade programme

VoidRanger40

This should honestly be a sticky, CMOS battery is step one on any machine past five or six years old and almost nobody checks it because it seems too small to matter

QubitZero

I would not make it step one myself, a dying battery causing full startup crashes rather than just settings amnesia is the less common version. Worth checking early because it costs a pound and a minute, but it is not the usual suspect

Evan0

That is exactly why it costs people months though, everyone including the internet tells you it only resets the clock, so it never gets tried until desperation sets in
My team is always one signing away

Holly

Looking forward to the pictures, there is something properly satisfying about a photo of a ten year old tower back from the dead next to the coin sized part that nearly killed it
404: Signature not found

CosmicRay65

Posting because I had no problems until today when I unplugged my HDD which is idle except as a backup. I got the error again.

Anchor99

This is a good reminder that intermittent hardware faults can look like software problems for years. Windows crashing during startup naturally makes you suspect drivers, updates, corrupted system files or the Windows installation itself, especially when the machine eventually boots normally.

A flaky SATA cable or connector can produce exactly the sort of nonsense that wastes hours. The drive may work perfectly most of the time, then briefly lose communication during boot and leave Windows looking guilty for something that happened underneath it.

The useful lesson is to change one physical thing at a time and test properly after each change. Reseating the drive, trying another SATA port and replacing the cable are all cheap checks compared with reinstalling Windows or buying a new SSD.

And the fact that unplugging the old HDD brought the error back is particularly interesting. That suggests the drive or its connection may still be part of the boot chain, even if you think of it as merely a backup disk. Old PCs have a habit of accumulating dependencies nobody remembers creating. :)

James93

The HDD comment would have sent me straight into the BIOS. If the PC behaves differently depending on whether that drive is connected, I would want to know exactly what the firmware thinks the boot devices are.

Older systems can have confusing boot entries, especially after multiple Windows installations or drive replacements. You can end up with several boot options that look almost identical, and the machine may quietly choose one that depends on hardware you thought was obsolete.

Before changing anything, take photographs of the current BIOS boot settings. That gives you a reference if something goes wrong. Then check whether the intended system drive is actually first in the boot order and whether its bootloader is present.

It is amazing how often a "Windows problem" turns out to be a machine that has been booting through some weird historical arrangement for years. Computers do not complain about this while everything works, so the first clue often arrives when somebody unplugs an old drive.
sudo make me a sandwich

NicholasCleverley

A decade-old Windows installation can accumulate some spectacular historical baggage. Different drives get added, Windows gets upgraded, partitions move around, boot settings change, and eventually nobody remembers why the machine needs something that appears unrelated.

The classic example is a PC with Windows installed on an SSD but the boot partition created on an older HDD. It works for years because both drives are always present. Then someone disconnects the HDD and suddenly the computer refuses to boot.

That does not necessarily mean the HDD is faulty. It can simply mean the computer has been relying on it for part of the startup process. Checking the partition layout before deleting or formatting anything is therefore really important.

If the £1 part fixed the crashes but the HDD still has this strange dependency, I would treat those as separate jobs. Fix the immediate fault, then tidy the boot arrangement once the system is backed up. No need to create a second problem while celebrating the first.
rm -rf /bad-ideas

Aidan

One thing I would be careful about is assuming that because a hard drive is only used for backups, disconnecting it should have zero effect on booting. The boot process does not care what you consider the drive's purpose.

If the bootloader lives there, the firmware may still need it. You can have Windows itself sitting happily on an SSD while the small system partition needed to start it is sitting on an older HDD. Remove that HDD and the SSD suddenly looks like it has no operating system to the firmware.

This is worth checking before replacing anything. Open Disk Management and look at which disk contains the EFI System Partition or System Reserved partition. On older BIOS setups, also check the active partition and boot order.

If that is the situation, the long-term fix is to move the boot files onto the drive you actually intend to keep. Just make sure the backup situation is sorted first. A cheap component may have caused the original crashes, but the strange boot arrangement could be a separate historical quirk.
Quantum computer said maybe, so I'm calling it a win

AustinTheory15

This sounds like classic intermittent-fault territory. Those are the worst because every individual test can appear to prove that nothing is wrong. You reboot ten times and it works ten times, then the next morning it crashes twice before breakfast.

Storage cables are particularly easy to overlook because people tend to think of them as passive bits of plastic and wire. They are not sophisticated, but they still have connectors, contacts and cables that can deteriorate or become slightly loose after years of being moved around.

A useful diagnostic trick is to reproduce the problem while changing only one thing. If swapping the SATA cable suddenly gives you twenty clean boots, that is much stronger evidence than reinstalling Windows and finding that it happens to work for a week.

The £1 solution is also a good argument for keeping a small box of spare cables and screws around. They cost almost nothing and can save a Saturday afternoon of unnecessary troubleshooting. :)

ComputeNodeCanopy

The most useful takeaway for me is to pay attention to what changes the failure, not just what the failure looks like. Startup crash, blue screen or boot error can all point you toward Windows, but the trigger is often more informative.

If connecting a particular drive changes the outcome, that is evidence. If moving a cable changes it, evidence. If the machine only fails after being cold for several hours, that is evidence too. Those observations help narrow the fault before you start replacing things.

A good troubleshooting log can be as simple as writing down "cold boot failed, restart succeeded" or "booted normally with secondary HDD disconnected". After ten tests, patterns become much easier to see than they are when relying on memory.

It sounds excessive for a home PC, but this kind of disciplined testing is exactly what prevents the classic situation where someone buys £300 of replacement hardware and discovers the original £1 component was the culprit all along.

TheGame92

For anyone reading this and thinking "my PC does something similar", resist the urge to immediately copy a random repair command from a forum. First work out when the failure occurs and what hardware changes its behaviour.

If it fails before Windows loads, check firmware, boot devices, storage and power. If it gets into Windows and then crashes under load, look harder at drivers, temperatures, memory and power delivery. These are broad rules rather than guarantees, but they give you somewhere sensible to start.

The other good lesson here is to be suspicious of assumptions like "that drive is only for backups, so it cannot affect Windows". Computers do not understand your intentions. If the bootloader, page file, library or some system configuration references that drive, it can absolutely affect normal operation.

So yes, celebrate the £1 fix, but also investigate why unplugging the HDD causes trouble. That second clue may reveal a boot configuration that has been quietly waiting to surprise you for years.

Maverick50

For anyone with a similar problem, I would add one very boring test: check the power connections to the drives. A storage device can behave strangely if the SATA power connector is not making a solid connection, and the symptoms can look like a dying disk.

That is especially worth checking in an older case where cables have been pulled around during upgrades. Sometimes simply reseating the connector fixes an intermittent problem, although replacing a suspect cable is a better test if the problem returns.

I would also avoid assuming the newest drive is automatically innocent. If Windows is on an SSD but the machine crashes whenever the old HDD is active, the HDD could be causing timeouts or errors that make the operating system hang while waiting for it.

Back up anything important from that old drive before doing more experiments. If the drive is ten years old and only used for backups, it is probably time to think about having two independent copies anyway. A backup that is itself on the verge of failure is not much of a safety net.

RayOfLight73

The phrase "idle except as a backup" is exactly why I would not assume the HDD is irrelevant. Backup drives are often connected permanently for convenience, which means the operating system can still interact with them during normal startup and operation.

If the disk is starting to fail, Windows may spend ages waiting for it to respond. That can make the whole computer appear frozen even though the system drive itself is fine. In severe cases, applications can hang simply because they are waiting on a file operation involving the bad disk.

A useful test is to copy a large amount of data from the HDD and watch for pauses, error messages or unusually poor transfer behaviour. A proper SMART test is better, but even basic symptoms can tell you whether the disk deserves suspicion.

Either way, the backup drive should not be the only place those backups exist. If it contains the only copy of important family photos or documents, replace the strategy as well as the hardware. A backup is insurance, not an archive you hope will survive forever.

TheGreatMoney

A £1 part beating years of troubleshooting is painfully believable. The expensive lesson with old PCs is that software symptoms do not necessarily mean software causes.

When a machine crashes before Windows has properly loaded, I start looking at power, storage connections, memory and motherboard settings before doing anything dramatic to the operating system. A slightly loose connector can be enough to make a machine behave differently from one boot to the next.

The HDD detail is especially useful. If removing it changes whether the system boots, check the BIOS boot order and inspect whether the Windows bootloader is actually sitting on that old disk. It is surprisingly common for Windows to be installed on one drive while the boot files ended up on another.

That can leave you with a PC where the supposedly unimportant backup drive is actually essential to starting Windows. Remove it and suddenly the machine acts like the operating system has vanished. Very confusing until you know what happened.

Benzema63

The part that made me laugh is the idea of spending years fighting Windows only for a component costing roughly the price of a packet of biscuits to win the argument. Computers really do have a sense of humour.

There is a serious point underneath it, though. Windows startup failures often encourage people to run increasingly complicated repair commands because those are the things search results tend to recommend. Sometimes that is exactly backwards.

If the failure happens before Windows fully starts, hardware should move higher up the list. If it happens at exactly the same stage every time after Windows loads, then software becomes more plausible. That distinction is not perfect, but it is a useful starting point.

And with a ten-year-old PC, physical connections deserve special attention. It has probably been opened, dusted, upgraded and moved around dozens of times. A connector that was fine in 2015 does not get a lifetime warranty just because it has no moving parts.

Gaz64

The old-PC angle is important too. Components do not have to be completely dead to cause trouble. A drive can pass a quick health check and still have occasional communication problems, and a cable can work most of the time while producing errors under particular conditions

That is why intermittent faults are so expensive in terms of human time. You can run a diagnostic utility and get a reassuring green result, then experience a crash ten minutes later. The software is reporting what it can see at that moment, not giving the hardware a lifetime certificate

If the cheap replacement fixed the problem consistently, that is strong practical evidence even if you never identify the exact electrical failure. You do not always need a courtroom-level explanation before accepting a successful repair

The sensible next step would be to keep the machine under observation for a while rather than immediately declaring victory after two boots. A few cold starts, restarts and longer sessions will tell you much more. If it stays stable, take the win
Reply quality may vary depending on caffeine levels and gradient stability.

DeadChat58

There is a useful distinction between a bad drive and a bad connection that gets missed in these stories. If the drive itself is failing, replacing the cable may appear to fix things temporarily because moving the drive changes the physical contact or timing. If the cable is bad, the drive may be perfectly healthy

That is why checking SMART data is worthwhile, but it should not be treated as the final word. Look for reallocated sectors, pending sectors, uncorrectable errors and interface communication errors. The exact numbers depend on the drive, but a history of interface errors can point toward cabling or connection problems

For anyone following along with their own ancient PC, take screenshots of important SMART information before changing hardware. It gives you something to compare later and can help distinguish a drive problem from a connection problem

Also, do not forget the boring solution of cleaning dust from around connectors and fans. Do it carefully and with the system powered down, obviously. A decade of dust is enough to make an old machine run hotter and less reliably even if it has nothing to do with this particular fault
Posted from a machine that definitely needs a clean install

NeuralTrace

The £1 repair is a great reminder that modern PCs are still collections of surprisingly simple electrical connections. We talk about CPUs with billions of transistors, but the computer can be defeated by one tiny connector having a bad day

A lot of people are understandably reluctant to open the case because they assume they will make things worse. In reality, a basic inspection is often safer than randomly installing software fixes, provided the machine is powered off and you take sensible precautions

Take a photo before unplugging anything. Label cables if necessary. Do not force connectors. Then reconnect one item at a time and make sure everything is seated properly. Those few minutes can save hours later

The important part is not to turn every problem into "replace the cheapest part and hope" either. Cheap components can fail, but diagnosis still matters. The satisfying part of this story is that the cheap replacement was supported by the symptoms rather than being a random guess

Georgia97

This is why I keep saying that troubleshooting is mostly about asking better questions. "Windows crashes" is a symptom. "Windows crashes only during startup, and changing this physical connection stops it" is useful evidence

The same applies to the HDD comment. If plugging or unplugging one device changes the behaviour, that is a massive clue. It narrows the search from everything installed on the PC to the device, its cable, its power, its controller and the boot configuration

You can then work systematically: try another SATA cable, another motherboard SATA port, check the drive health, inspect the boot order and test with the drive disconnected. Each result tells you something

That approach beats reinstalling Windows five times. A reinstall can temporarily hide a problem by changing the timing or resetting configuration, but it does not repair a bad connector. The machine then mysteriously develops the same fault months later
Be excellent to each other, entangled or not

HollowSentinel

My favourite part of these repairs is that the cheap component usually looks completely innocent. A cable has no flashing lights, no error message and no software driver, so it gets ignored while everyone blames the complicated stuff

That is especially true with intermittent faults. A bad connection may work perfectly for days and then fail during a boot when timing, vibration or temperature happens to be slightly different. You replace the cable and suddenly the machine becomes boring again, which is exactly what you want

Boring is underrated in computer maintenance. A machine that boots every time, has good backups and does not randomly freeze is much more valuable than one that benchmarks 10% faster but requires constant attention

So I would take the win, document which part fixed it, and leave the machine alone for a while. If it stays stable through a couple of weeks of normal use, then start considering whether there are any other preventative jobs worth doing. No need to fix problems that have not appeared yet

Karen50

The HDD unplugging comment is the bit that would make me stop and investigate rather than immediately blaming Windows. If the computer only starts reliably when a particular drive is connected, that drive may be involved in booting even if it contains no obvious Windows files

On an older installation, the EFI or boot manager can end up on a different disk from the main Windows partition. This is particularly common when Windows was installed while several drives were connected. The installer chooses a suitable boot location and most people never notice

You can check this from Disk Management or the firmware boot menu, although I would be careful about changing partitions until the important data is backed up. If the old HDD is genuinely only a backup, eventually moving the boot files to the system drive would make the setup much less surprising

So there may actually be two lessons here: find the cheap faulty component, and then clean up the old boot configuration. Ten years of upgrades can leave a PC working perfectly while hiding some very odd historical decisions

ClusterCrossing

There is a tendency to throw increasingly expensive hardware at an old computer once problems start appearing, so I like the opposite approach here. Start with the £1 and £5 possibilities before assuming the motherboard or CPU is dying

A proper basic check would be power down, unplug the machine, reseat the RAM and storage cables, inspect the connectors and check for obvious dust or corrosion. Then test with the minimum hardware required to boot. That can narrow things down surprisingly quickly

One thing I would add is to check the SMART information on that old HDD before trusting it with backups. A disk can still be readable while developing errors that cause intermittent delays and timeouts. If it contains anything important, copy the data somewhere else before experimenting too much

Cheap parts are great, but the real win here is the troubleshooting method. Change one variable, test, and keep notes. Otherwise it is very easy to replace three things at once and have no idea which one actually fixed the fault

Laura53

There is something reassuring about a story like this because it shows that troubleshooting does not always require specialist equipment. Sometimes the most valuable tool is simply noticing that changing one small physical thing changes the computer's behaviour

People often jump straight to commands, recovery environments and registry edits because they are easy to search for. Those tools have their place, but they cannot fix a physical connection problem. In fact, repeatedly repairing Windows can waste time while the underlying hardware continues to misbehave

For an older desktop, a sensible first pass is surprisingly simple: check power, check cables, reseat RAM, inspect storage connections, check temperatures, look at drive health and confirm the boot configuration. None of that requires a £500 diagnostic setup

Then escalate if the evidence points toward a motherboard, power supply or other expensive component. Starting with the cheap possibilities is not cutting corners; it is good troubleshooting economics

Ryan98

This thread is a good example of why swapping the entire PC is not always the sensible answer. A ten-year-old machine can still be perfectly usable if the workload is modest and the core hardware is healthy

The danger is that old hardware gets judged as a single unit. Someone sees Windows 11 problems and thinks motherboard, CPU, RAM, graphics card, everything must be ancient. In reality, one £1 cable or connector can be the weak point while everything else keeps working normally

Of course, age matters for reliability. Fans wear out, capacitors age, hard drives accumulate hours and power supplies can become less reliable. That is why I would keep good backups and avoid spending a fortune upgrading a machine that is already showing several unrelated faults

But if the system is stable after the cheap repair and does everything you need, there is no prize for replacing it early. Use it until the cost of keeping it running exceeds the value you get from it

Taker

The backup HDD detail raises another issue that is worth mentioning: backups should be tested. A lot of people discover only when they need a file that their backup drive has been quietly failing for months

If this old drive is being used as the backup destination, I would check that the files can actually be opened and that there is another copy of anything irreplaceable. Ideally, keep at least one backup disconnected when it is not being updated so a malware infection or electrical fault cannot take out both copies

It is also worth considering whether the backup drive needs to remain installed at all. If it is only accessed occasionally, a removable external drive can make the machine's normal boot configuration simpler. Just do not confuse that with a substitute for multiple backups

As for the original crash, the fact that disconnecting the HDD changes the behaviour is a very strong clue. Whether the culprit is the drive, cable, power connection or boot configuration, that is the area I would investigate before touching Windows again

LunarShade

This also highlights why replacing an old PC can sometimes hide the real lesson. If the same setup had been rebuilt on a new machine without understanding the fault, the original problem would disappear but you would learn nothing from it

Finding the actual cause gives you a useful troubleshooting pattern for future hardware. If a cheap SATA cable can cause years of apparent Windows instability, you start thinking about physical connections earlier next time instead of assuming the operating system is broken

There is also a practical argument for keeping a few inexpensive spares around. SATA cables, CMOS batteries, screws and basic adapters are cheap enough that having them available makes testing much easier. You do not need a cupboard full of parts, just a small emergency kit

The only warning I would add is not to confuse a successful fix with proof that every other component is healthy. A decade-old PSU or HDD can still fail later. Backups and sensible monitoring remain worthwhile, especially on a machine that has become the household workhorse simply because it refuses to die

Related Topics (4)