Incompatibilities during the build
Although most combinations of PC components
slot together seamlessly, sometimes you’ll find a particular component that
just won’t play ball. A very common problem is where motherboard and case
designers clash in their decision on how many pins to make the power LED
connector. On most motherboards this is a three-pin-wide header with a bank
slot in the middle. In other motherboards, the blank pin is omitted and a
two-pin design is utilised instead, as with the HDD-activity LED. Fortunately,
modifying a three-pin header into a two-pin one is simple, and it can be done
with a hobby knife or simply using a precision flat-head screwdriver to lift
the plastic pin cover, remove the female metal section of the pin and install
it in the adjacent blank hole. It’s annoying that a problem we were dealing
with 15 years ago is still prevalent in 2012, but that’s just part of being a
PC owner!
PC components
Another very common hurdle people have to
overcome is an inability to get the add-in cards (graphics and PCI cars) to
properly seat in the relevant slots. In a perfect world, all motherboards,
cases and standoffs would be precision engineered to fractions of a millimetre.
In the real world, unfortunately, this is not the case, and it’s not uncommon
to have problem sliding the PCI backing plates into the relevant long
rectangular holes in the case. This is fortunately easy to remedy and just
requires bending of the PCI backing plates very gently either towards or to the
back of the card until it slots in.
Problem during OS installation
If your PC fails with a blue screen error
during the installation of Windows, there can (as usual) be a number of
potential things that have gone awry. One of the most common is that you have a
problem with your memory configuration. It may need more voltage, less voltage
or may need to be run at more relaxed latency timings. Go into the BIOS, set
your memory to the failsafe slowest speed and manually force the latencies to
9-9-9-24 timings for DDR3. If this does not resolve your issue, use a different
PC to create a memtest 86+ boot CD (www.memtest.org)
and see if it throws up any errors. If the memory is error free, the cause may
be as a result of something else, like a hardware incompatibility, faulty
motherboard or overheating CPU.
During installation of the OS, at some
point the operating system looks for hard disks and may throw up a message
saying it cannot find any. This is normally caused by the fact that the chipset
or controller you’re using requires drivers that are not natively known by
Windows. This problem is a lot less prevalent with Windows 7 than it was in
Windows XP, but some of the latest hardware may require a third-party driver,
which can be loaded from a pen drive or DVD.
Unexpected shutdown
Error notification
If everything seems to be running smoothly,
but then your system suddenly shuts down or reboots unexpectedly, then you have
a problem. It’s important to determine which of these two things is actually
happening, because a system turning itself off completely is invariably more
serious than one that occasionally reboots itself. The most common cause of the
former is overheating. Modern chips don’t struggle on regardless once they get
hot. Instead, they reduce the clock speed down to a safe speed to maximise the
time before which temperatures get out of control. This is commonly known as
throttling and serves to minimise CPU damage caused by incorrectly seated
coolers. A second level of protection that’s built into all current AMD and
Intel CPUs is a thermal cut-off that will cut power to the machine once a
certain temperature is reached.
If these symptoms sound familiar, check the
seating of the cooler by removing it and checking the thermal paste footprint.
You should be able to see remnants spread evenly across the surface of the heat
spreader. If they’re concentrated only on a small section of the CPU or, even
worse, are nowhere to be seen, you need to remove the cooler completely,
reapply paste if necessary and reinstall the cooler carefully. Enter the BIOS
and go to PC health status and watch the CPU temperatures. Most chips idle at
less than 50oC. if your temperatures are significantly higher than
this or continue to rise until either the reported clock speed drops or the
machine hits more than 100oC and shuts off again, you know you have
a cooler-mounting related issue.
Unexpected reboot
A license exception has occured
If your system is plagued with the dreaded
occasional instability, whereby everything seems to ben working fine but you
experience the occasional blue screen error or unexpected reboot, the first
stage is to try and find a way to reliably reproduce the error. After you have
first successfully installed the OS, it’s always prudent to run a soak test before
you start using the machine. One of the best tests for CPU and memory is Prime
95, a program that can be grabbed for free at www.mersenne.org/freesoft. Within
the Torture test menu you will find several options. ‘Blend’ mode is ideal for
stressing memory while ‘Small FTTS’ is better for generating the maximum load
on your CPU. The test will run perpetually until you tell it to stop, but
serious errors are normally thrown up within half an hour of testing.
Although Prime 95 cannot distinguish
between CPU and memory-related problems, you can use the built-in tests modes
to get a pretty good idea. If the test errors quickly in the blend test but
slowly or not at all with small FTTS, you can almost certainly rule out the CPU
as the root of any problems. In which case, it’s back to the BIOS so that you
can adjust the memory timings, voltage or clock speed until you obtain
stability. If you cannot get stability, it’s like one of your sticks is
unfortunately not hitting the advertised specification or is faulty.
If your system is rock solid when it’s
being used for light tasks, but seems to have problems when it’s being
stressed, the most likely cause is going to be the PSU. If you’re using an
underspecified PSU, it won’t be up to the job of supplying the required voltage
to your processor and video card when the going gets tough. If one or more of
the voltage rails dips below that required by your hardware, it will not get
the juice it needs to work correctly, resulting in an error, which can emulate
any other kind of instability. For this reason, instability due to an
insufficient or faulty PSU can be one of the trickiest to isolate and track
down.
Graphical instability, driver resets or corruption
If your system has noticeable graphical
errors, commonly referred to as artefacts, it’s very easy to point your finger
at the video card. Before you assume the worst, however, double-check that the
signal cable is plugged in fully at both the monitor and the video card ends of
the lead. If it’s not, you can get some very unusual behavior. If that still
doesn’t resolve the issue, and you have easy access to a spare lead, it’s worth
trying that as well. If your monitor has dual inputs, try it using the
different interface.
If you still have visual artefacts, it’s
time to do some testing. 3DMark is a great way to do this, because you can loop
identical graphical tests and look for signs of artefacts. Artefacts can
manifest themselves as tearing, geometry errors or speckling and are normally
caused by factors including insufficient power to the video card, overheating,
incompatible driver revisions or, worst of all, a faulty card.
Try running with the case side off to allow
for easier heat dissipation if you suspect excessive heat is to blame.
Insufficient power from your PSU can also often show up as visible errors, so
if you’re concerned your PSU isn’t up to the job, there’s a good chance it
isn’t. Update any driver software you installed from a provided DVD (or Windows
update) to the latest from either AMD or NVidia’s site. If you cannot get
stability, try a stronger PSU, or if this is unlikely to be a cause. Test the
card in another PC to see if it works correctly.
If your system is using integrated
graphics, then a problem that manifests as visual distortion could actually be
a problem with your system memory or motherboard. Try reducing the frequency of
your memory or set your motherboard to its factory defaults to see if that
solves the problem.
