I don't actually have any problems with my rig at the moment, this thread is purely for educational purposes, but say someone has a computer that is rebooting/crashing a lot, is there a reliable way of finding out if the PSU is faulty or inadequate without either a) opening up the case and using a multimeter or b) buying a new PSU.
Diagnosing power supply problems
- Thread starter decimator
- Start date
i think that is the only 100% accurate way of testing if the psu is to blame for a problem.
you could always jump start it to make sure it's staying powered up for more than a few secs.
Yes I was wondering if they would be any use, I've got one called PC Probe that came with my Asus mobo which can apparently measure all these voltages accurately to two decimal places.
I'm not sure what you mean by jump start, but I was thinking more about PSU's that were working OK until you launch a really demanding appliacation or something
PSU diagnosis is pretty tough, the tools provided by the motherboard are too crude really ,and tell you the wrong info most of the time.
A multimeter will give you a clue as to whats going on, if underload you see large voltage shifts, then the PSU is struggling to regulate its output and is probably being pushed over the limit. However minor damage to a PSU's regulators can cause serious ripple even though the voltages may appear ok. Ripple is very bad for the motherboard.
PSU's can also suffer from burst 'Bad caps', which will normally cause extremely poor quality outputs, and will also eventually fry the PSU's regulators completely.
As Philly said, swapping the PSU for a 'known' good unit is normally the 'easiest' solution. And of course if the replacement PSU doesnt solve the problem.. then you have probably proven that the original PSU is still usable too
A multimeter will give you a clue as to whats going on, if underload you see large voltage shifts, then the PSU is struggling to regulate its output and is probably being pushed over the limit. However minor damage to a PSU's regulators can cause serious ripple even though the voltages may appear ok. Ripple is very bad for the motherboard.
PSU's can also suffer from burst 'Bad caps', which will normally cause extremely poor quality outputs, and will also eventually fry the PSU's regulators completely.
As Philly said, swapping the PSU for a 'known' good unit is normally the 'easiest' solution. And of course if the replacement PSU doesnt solve the problem.. then you have probably proven that the original PSU is still usable too
Would you not need to look at the voltages on an oscilloscope (rather than a multimeter) to detect any ripple? also what sort of thing could damage the PSU's regulators? shock? overloading? or just ageing?
Absolutely, thats the problem I was trying to say... A motherboard isnt even good at voltages. A multimeter is an improvement, and at least it will show voltage change under load... but as you say a Scope is required for ripple tests, and other 'noise' conditions.
A very common cause of PSU failer (apart from being overloaded), is simply bad capacitors. A sealed cap can last for years without serious degrading, but if the cap vents its history, and after that the circuit is no longer running to the original design spec. Loads of people have heard of bad caps on motherboards, but unfortunatly PSU caps are just as prone to be 'bad'.
All down to cutting costs really, I've got a 30 year old 'Hi fi' amplifier, with some whopping large caps on it, and it still runs perfectly. These days a lot of componants will fail in just a few years.
Most commonly this leads to ripple and eventually very commonly the regulators will be fried.
A burst capacitor can show up in many ways. Most dramatically the entire end of the cap can be blown apart, spewing paper and traces of chemicals around the computer. A tiny bit of liquid, but mostly its blown off as a nasty smelling gas.
However it can be a lot more 'subtle' a pin sized hole in the metal top of a cap, or sometimes the entire cap can appear swollen. Sometimes they burst at the bottom instead, in which case they may just appear to be leaning over (Of course to make it worse sometimes they are soldered with a bit of 'spare' leg so just because a cap is leaning doesnt automatically mean its failed.)
Sometimes on a cap that has burst the metal top might appear rusty at the seams, its not rust of course, but a reaction as the gasses escape.
Bad caps have found their way into some fairly high end products unfortunatly, but yes in general better brands will use higher quality 'japanese' parts which are much less prone to the badcap issue than say Caps from China etc. (Not saying all chinese caps are bad.. just those from certain factories produced in certain years..)
Its possibly an urban legend, but the story goes that some scientists stole a forumla for a new liquid to go inside caps, and sold it to another manufacturer. Unfortunatly the forumla was faulty, but this was not found before millions of caps had been distributed around the world.
Whether or not this is true, or if it was simply cap companies trying to reduce costs, and experimenting with different electrolytes we'll probably never know for sure. But what is for sure... cheap caps dont last very long. Many motherboards now use solid caps instead which are immune to this problem (and hopefully have a much greater lifespan).
However it can be a lot more 'subtle' a pin sized hole in the metal top of a cap, or sometimes the entire cap can appear swollen. Sometimes they burst at the bottom instead, in which case they may just appear to be leaning over (Of course to make it worse sometimes they are soldered with a bit of 'spare' leg so just because a cap is leaning doesnt automatically mean its failed.)
Sometimes on a cap that has burst the metal top might appear rusty at the seams, its not rust of course, but a reaction as the gasses escape.
Bad caps have found their way into some fairly high end products unfortunatly, but yes in general better brands will use higher quality 'japanese' parts which are much less prone to the badcap issue than say Caps from China etc. (Not saying all chinese caps are bad.. just those from certain factories produced in certain years..)
Its possibly an urban legend, but the story goes that some scientists stole a forumla for a new liquid to go inside caps, and sold it to another manufacturer. Unfortunatly the forumla was faulty, but this was not found before millions of caps had been distributed around the world.
Whether or not this is true, or if it was simply cap companies trying to reduce costs, and experimenting with different electrolytes we'll probably never know for sure. But what is for sure... cheap caps dont last very long. Many motherboards now use solid caps instead which are immune to this problem (and hopefully have a much greater lifespan).
Capacitors can also fail in another way. Their ESR (Equivalent Series Resistance) can rise. This is very common. Many of the capacitors in a PC power supply are required to be low ESR designs and cause supply failure if their ESR rises too much.
A capacitor that has burst or leaked will very likely have a raised ESR. However, a capacitor can have a raised ESR but look in perfect condition. You need an ESR meter to measure it.
A capacitor that has burst or leaked will very likely have a raised ESR. However, a capacitor can have a raised ESR but look in perfect condition. You need an ESR meter to measure it.