Reliable power supplies?
Multithreaded
motherboard problems with; see another of my threads), it went bang! -
twice. The electrics in the house went out, too, on the second thump,
so I had to struggle to get the circuit breaker to reset. Luckily, as
I'm intending to upgrade to a Socket AM3 system, I bought a 750W
Corsair PSU. Fitting this and gingerly pressing the Go button produced
a nice POST beep and a booted machine - initially. I did all the usual
diagnostics from floppy, that took hours, to confirm the ultra
expensive SCSI main HDD, SATA HDDs and PATA HDD were OK. Memtest86+
reported that the CPU and RAM were unharmed, miraculously. Knoppix
seems fine when booted off the DVD rewriter. Only the PATA HDD failed
its diagnostic, but I put that down to an existing fault on the
mainboard as I've previously discussed, and am willing to bet money
that the disk will be fine in any of my other eight PCs.
Today, things are not so happy. Windows freezes, sound is stuttery,
LAN speed is way down. So now I'm running SCSI diagnostics as I type
this, hoping it is just the motherboard that has suffered a delayed
reaction to yesterday's trauma. Ho hum.
Anyhoo, my question. What do people on here regard as good, and above
all, reliable, high power PSUs? I like the Corsair because it is very
quiet, and hope that running it at nowhere near top capacity will make
it last longer than its predecessor (an Enermax 465W from 2004). I
just want reassurance that I haven't jumped from the frying pan into
the fire. I need a high power PSU as I run four HDDs, including the
current-hungry 10,000RPM SCSI unit; the CPU takes 95W allegedly, and I
will be running a GTS250 graphics adaptor.
--
Multithreaded.
><(((°>
Firstly I'm a little surprised that your Enermax has failed but at least
it sounds as though it's failed in a "safe" way and not taken loads of kit
out with it. Although not the best Enermax psu's are usually a good. Out
of curiosity I usually take failed psu's to bits and more often than not
find bulging electrolytic capacitors which I assume to be the cause of the
failure. Be careful though in taking psu's apart, lethal voltages can
exist internally so leave unplugged overnight before removing cover.
There are a lot of good makes and a lot of rubbish psu's out there. As is
often said a good indication of quality is the units weight, heavier the
better.
Makes I'd trust include: Thermaltake, Coolermaster, Antec, Hiper, OCZ
Technology, Corsair, Seasonic & yes Enermax.
In the main the usual rule of "you get what you pay for" comes into play.
Buy a Q-TEC PSU 650W Triple Fan (Gold) for about ￜ25 = expect trouble.
Trouble can be big too. I've seen many a Bestec psu failure kill
motherboard and more.
Worst I saw was a psu failure that killed every part in the PC except for
a memory stick and the floppy disc drive.
R Gower
" ><(((�>" <nos...@butfish.com> wrote in message
news:op.u4qtljeucnngb9@xp64...
Not when it isn't plugged in they can't - you don't know how a switching
mode power supply works.
> There are a lot of good makes and a lot of rubbish psu's out there.
What research have you done?
>As is often said a good indication of quality is the units weight, heavier
>the better.
Not true, you are becoming confused trying to compare a switching
mode power supply to other unrelated electronic goods.
> Makes I'd trust include: Thermaltake, Coolermaster, Antec, Hiper, OCZ
> Technology, Corsair, Seasonic & yes Enermax.
>
Why would you trust each one and what are the differences between
the way the circuits operate in each and what test equipment are you
using? I hate to tell you, the circuits are manufactured by a few companies
and badged like other electronic consumer goods. You will find that
if you take a few apart, they are identical inside, often using the same
components and layout to give the same performance. So you do not
always get what you pay for, you get a brand name for however much
you are prepared to pay. That figure is based on how well a company
promotes the brand and advertises. The ones you name probably all
advertise in magazines so are the only brand names you are familiar
with - without knowing which of those have the same "insides".
I used to find that people would buy a cheap power supply and go
for the wrong one due to lack of knowledge. When it didn't operate
the computer correctly they would claim that particular brand was
no good. When they bought the one to suit with the correct rating,
they would see the computer working and claim that make of
power supply was superior. That was not based on a technical
knowledge, but it suited them to say it as they didn't want everyone
to know they knew very little about computers!
> In the main the usual rule of "you get what you pay for" comes into play.
> Buy a Q-TEC PSU 650W Triple Fan (Gold) for about �25 = expect trouble.
> Trouble can be big too. I've seen many a Bestec psu failure kill
> motherboard and more.
Have you really or are you just making that up to support an argument
you can't really make? I have never had a power supply blow up, fail
or damage anything and I bought cheap ones. They were the same inside
as more expensive brands but far cheaper. I've used various makes
over the years. The only power supplies I saw breaking computers were
those in the DELL Optiplex PCs, they used to go like smoke machines!
> Worst I saw was a psu failure that killed every part in the PC except for
> a memory stick and the floppy disc drive.
What rubbish, you just seem to be inventing what you claim as you go along.
You need to learn a lot more about electronics and the design of switching
mode power supplies. That might stop you making mistakes in fitting
underrated supplies and causing damage through your lack of knowledge.
Jaimie Vandenbergh
wrote:
>You need to learn a lot more about electronics and the design of switching
>mode power supplies. That might stop you making mistakes in fitting
>underrated supplies and causing damage through your lack of knowledge.
Thank you for being so helpful.
Why not use some of your boundless knowledge to educate the original
poster, and suggest decent PSUs that might replace the problematic
ones he has?
Cheers - Jaimie
--
A mind stretched by an idea can never go back to its original dimensions.
- Conan Doyle
><(((°>
>
> " ><(((ￜ>" <nos...@butfish.com> wrote in message
>> Buy a Q-TEC PSU 650W Triple Fan (Gold) for about ￜ25 = expect trouble.
>> Trouble can be big too. I've seen many a Bestec psu failure kill
>> motherboard and more.
>
> Have you really or are you just making that up to support an argument
> you can't really make? I have never had a power supply blow up, fail
> or damage anything and I bought cheap ones. They were the same inside
> as more expensive brands but far cheaper. I've used various makes
> over the years. The only power supplies I saw breaking computers were
> those in the DELL Optiplex PCs, they used to go like smoke machines!
>
>> Worst I saw was a psu failure that killed every part in the PC except
>> for
>> a memory stick and the floppy disc drive.
>
> What rubbish, you just seem to be inventing what you claim as you go
> along.
> You need to learn a lot more about electronics and the design of
> switching
> mode power supplies. That might stop you making mistakes in fitting
> underrated supplies and causing damage through your lack of knowledge.
>
>
Sorry to disappoint you but I won't rise to the bait.
I know how long I have worked as an electronic design engineer.
I know how long I have worked within a team designing very specialised
switched mode power supplies.
I know for how long I've been repairing and building computers.
I know how many computers I've repaired with failed power supplies and the
results of those failures.
You know nothing of this and guess what - I'm not going to enlighten you.
As I've said before regarding what's posted here by various knowledgeable
and less knowledgeable folks:-
"Take in what you consider useful & ignore the rest"
PeterC
> There are a lot of good makes and a lot of rubbish psu's out there. As is
> often said a good indication of quality is the units weight, heavier the
> better.
> Makes I'd trust include: Thermaltake, Coolermaster, Antec, Hiper, OCZ
> Technology, Corsair, Seasonic & yes Enermax.
They are all 'OK' or good; Antec are OK but not the best.
Overall I'd go for Seasonic, based on [some] experience and reviews, as
they're quiet and cool. Also, Seasonic make PSUs for some of the other
well-known brands.
--
Peter.
The head of a pin will hold more angels if
it's been flattened with an angel-grinder.
Multithreaded
wrote:
>Firstly I'm a little surprised that your Enermax has failed but at least
>it sounds as though it's failed in a "safe" way and not taken loads of kit
>out with it. Although not the best Enermax psu's are usually a good. Out
>of curiosity I usually take failed psu's to bits and more often than not
>find bulging electrolytic capacitors which I assume to be the cause of the
>failure. Be careful though in taking psu's apart, lethal voltages can
>exist internally so leave unplugged overnight before removing cover.
It's ironic. I bought the Enermax back in '04 to replace a perfectly
good, working PSU. I had just heard - on here - that Enermax were a
known good make back then, and I just wanted peace of mind.
>There are a lot of good makes and a lot of rubbish psu's out there. As is
>often said a good indication of quality is the units weight, heavier the
>better.
>Makes I'd trust include: Thermaltake, Coolermaster, Antec, Hiper, OCZ
>Technology, Corsair, Seasonic & yes Enermax.
The Enermax replaced an Antec I was originally running. That developed
a noisy inductor that chirps like a bird now, but otherwise is still
doing good service in my lowliest machine.
I built a super (for the time) system a couple of years ago, for a
friend, and put a Thermaltake in it. I was impressed by its heftiness
- I always take that as at least a hint of goodness - and super-quiet
operation. I cursed my friend for being so lucky. The PSUs in the
other eight of my PCs were/are all noisy as hell.
I'm glad you put Corsair in among the good makes. The thing is
enormous, has a huge, quiet fan, and purrs like a contented kitten
when under a considerable load. The voltages are bang on, too. (I
think I'd better not risk the anger of the PSU gods and use the word
"bang" in vain.)
>In the main the usual rule of "you get what you pay for" comes into play.
>Buy a Q-TEC PSU 650W Triple Fan (Gold) for about �25 = expect trouble.
>Trouble can be big too. I've seen many a Bestec psu failure kill
>motherboard and more.
>Worst I saw was a psu failure that killed every part in the PC except for
>a memory stick and the floppy disc drive.
I have had for years, running in two of my non-critical PCs, Q-TEC
PSUs, one 450W nominal, the other, 550. Both are noisy but have not
failed so far. Both drive machines with current-hungry SCSI HDDs, as
well as an assortment of IDE ones. Like you, I don't trust them, and
their lack of weight makes me suspect their current ratings. As the
machines get light use, usually, I'm quite happy to just let them be.
I once had an old 486, rescued from a skip, go pop on switch-on. I
think the floppy survived, as in your experience; nothing else did. I
can't believe I was so lucky a couple of days ago. I was sure that I'd
lost at least the ultra-expensive SCSI disk.
I've heard terrible things about Bestecs many times on here. Do they
make PSUs for anyone else? I hope I never come across one.
Thanks for the helpful reply, as ever.
--
Multithreaded.
Multithreaded
>find bulging electrolytic capacitors which I assume to be the cause of the
>failure. Be careful though in taking psu's apart, lethal voltages can
>exist internally so leave unplugged overnight before removing cover.
Just to add: I looked inside the dead PSU but could find no sign of
bulging caps: my first suspicion, in fact. I could see nothing
obviously wrong at all, but I didn't dare take the cover off as I'm
wary of charged capacitors and high voltages, having had many bad
experiences of such in the 80s - I'm a radio amateur, after all.
--
Multithreaded.
><(((°>
> ><(((ￜ> wrote:
>
>> Sorry to disappoint you but I won't rise to the bait.
>
> Two words: Suspect Roger. The headers and the fact he's never seen a
> Bestec
> (fitted to Packard Bells when they decided Fortrons were lasting too long
> past the warranty) take out a complete subsystem with a 12V rail sitting
> pretty at 16V tells all. What is it about Albasani that attracts trolls?
Regarding Bestec PSU failures:
Although I've encountered some in Scrapyard Hell PC's, the vast majority
of those I've seen were in eMachines PC's.
Very rarely did the motherboard survive.
PS Another make of PSU I'd trust but forgot to mention before is PC Power
& Cooling.
If anyone wants a more comprehensive list of decent PSU makes and some
discussion with it, then this forum is worth a look:
http://www.eggxpert.com/forums/1/323050/ShowThread.aspx
Clint Sharp
<r.g...@nospam.ntl.com> writes
>Not when it isn't plugged in they can't - you don't know how a switching
>mode power supply works.
criticising. Please don't post potentially lethal misinformation like
this.
--
Clint Sharp
><(((°>
> On Thu, 10 Dec 2009 22:18:45 -0000, ><(((ￜ> <nos...@butfish.com>
>> Buy a Q-TEC PSU 650W Triple Fan (Gold) for about ￜ25 = expect trouble.
>> Trouble can be big too. I've seen many a Bestec psu failure kill
>> motherboard and more.
>> Worst I saw was a psu failure that killed every part in the PC except
>> for
>> a memory stick and the floppy disc drive.
>
> I have had for years, running in two of my non-critical PCs, Q-TEC
> PSUs, one 450W nominal, the other, 550. Both are noisy but have not
> failed so far. Both drive machines with current-hungry SCSI HDDs, as
> well as an assortment of IDE ones. Like you, I don't trust them, and
> their lack of weight makes me suspect their current ratings. As the
> machines get light use, usually, I'm quite happy to just let them be.
>
> I once had an old 486, rescued from a skip, go pop on switch-on. I
> think the floppy survived, as in your experience; nothing else did. I
> can't believe I was so lucky a couple of days ago. I was sure that I'd
> lost at least the ultra-expensive SCSI disk.
>
> I've heard terrible things about Bestecs many times on here. Do they
> make PSUs for anyone else? I hope I never come across one.
>
> Thanks for the helpful reply, as ever.
I'm not aware of Bestec rebranding their PSU's for the major PC makers but
that isn't to say they don't.
The makes more usually rebranded include Astec, FSP group and probably the
biggest of all, Delta.
><(((°>
> ><(((ￜ> wrote:
>
>> Regarding Bestec PSU failures:
>> Although I've encountered some in Scrapyard Hell PC's, the vast majority
>> of those I've seen were in eMachines PC's.
>> Very rarely did the motherboard survive.
>
> I've seen blackened lumps of encapsulation on hard disk drives that used
> to
> be motor controllers, too. I'd forgotten about the eMachines, probably
> because by the time they'd reached me they were usually beyond help
> anyway.
>
>> If anyone wants a more comprehensive list of decent PSU makes and some
>> discussion with it, then this forum is worth a look:
>> http://www.eggxpert.com/forums/1/323050/ShowThread.aspx
>
> Nice to see Fortron get a mention. I've never had much of a problem with
> Fortrons.
In ten years or so I've perhaps seen 2 failures of FSP Group aka Fortron
psu's so yes IMO another good make.
Haven't however seen them readily available to buy and AFAIK have never
bought one.
Me thinks most of what they produce is sold directly to PC manufacturers.
Nick Le Lievre
> ><(((�> wrote:
>
>> Sorry to disappoint you but I won't rise to the bait.
>
> (fitted to Packard Bells when they decided Fortrons were lasting too long
> past the warranty) take out a complete subsystem with a 12V rail sitting
> pretty at 16V tells all. What is it about Albasani that attracts trolls?
There are some good people on Albasani like me.
Mike Tomlinson
<r.g...@nospam.ntl.com> writes
oh look. Another morph by the Tiscali idiot. abuse@albasani notified.
--
Mike Tomlinson
Mike Tomlinson
<m...@privacy.net> writes
> it went bang! -
>twice.
Lucky you :o) you normally only get one bang. Have you looked inside to
see what let go? I'm also interested in knowing the make and model, if
you still have it.
>Memtest86+
>reported that the CPU and RAM were unharmed, miraculously.
IME, damage due to a psu going bang is highly unusual. You'd have to be
very unlucky if it did. There are (should be) crowbars across the
outputs, so that if the voltage rises above a certain point, they are
instantly shorted out to protect the connected equipment. Cheap PSUs
either won't have these at all, in which case everything takes the
brunt, or cheap crowbar devices are fitted which react too slowly, in
which case it's pot luck.
But as you're going for the SoAM3 Ghia upgrade anyway, everything'll be
replaced with new, won't it? :o) The reality is that you can't trust
that motherboard (the one the old psu went bang on) and its parts any
longer. Give them to somebody you don't like.
An annoying little turd named westom1 will be along any second now to
tell you that a whole-house surge protector would have prevented your
PSU going bang while simultaneously announcing the Second Coming in a
blaze of glory (which you missed 'cos of the smoke pouring out of your
PSU). Take no notice, he's an idiot.
>Anyhoo, my question. What do people on here regard as good, and above
>all, reliable, high power PSUs?
It's a difficult question (see my thoughts re. power below.) IMO,
quality is more important than power rating. I won't go into it again -
it's been done to death here in the past - but I'd take a high quality
350W PSU anytime over a cheapo 650W. The 350W may have a lower wattage
rating on the label but it'll be better built and able to withstand a
sustained power draw of, say, 300W than the cheapo.
There are other factors to consider. For example, You also need to look
at whether the unit you are considering has split rails and if so, how
they are split (the rating label will tell you.) You would be unwise,
for instance, to attach all your four hard drives to one 12v rail.
Like hard drives, even good PSU makers can make turkeys from time to
time. A decent, consistent maker is Fortron Source (FSP). They OEM
(make PSUs for) many suppliers who then badge them with their own name,
but often keep the FSP-xxx part number still appears on the rating
label, which is how you can tell. I'd take a FSP-derived 350W over a
"1000W" Jeantech any time.
If you can, try and buy sight seen rather than online. Eyeball the unit
- does it look well made?
* is the packaging "cool" and flashy? Is it aimed at the modder crowd?
* Cheap fans look cheap. If it jams up, the PSU will overheat and you
may not realise until it decides to leave this vale of tears.
* Is it heavy (a sign of quality)? Sniff it - does it smell burnt
(you'll get some odd looks)? If so, reject it.
* Are the output leads a decent length? Some mobos put the connector in
weird places. Cheap PSUs usually have short, thin leads.
* Do you need 20 pin? Is the 24 pin header on the PSU split so it'll
fit 20 or 24 pin mobos?
* Do you need SATA connections?
etc. etc. etc.
> I like the Corsair because it is very
>quiet, and hope that running it at nowhere near top capacity will make
>it last longer than its predecessor
The best thing to do is to try and make sure it'll fall within the
efficiency curve. Putting a 750W psu in a system that will only draw
200W most of the time is running the power supply very inefficiently.
You want to try and aim for the middle of the curve, then you have some
in reserve for peaks and possibly to carry forward to your next upgrade.
> (an Enermax 465W from 2004).
It's worked for 6 years. Don't think it owes you anything. I'd be
interested in seeing a photo of the insides if you can.
> I
>just want reassurance that I haven't jumped from the frying pan into
>the fire. I need a high power PSU as I run four HDDs, including the
>current-hungry 10,000RPM SCSI unit
good idea, but as above, if you get a split-rail supply, try and
distribute 12v equally across the rails.
I have a 10k rpm 147gb broken drive to hand. This wants 1.2A on 12v, a
fair amount. You don't want to put four of these on one 12v wire -
that's 4.8A when all the drives spin up, and the mobo and the optical
drive(s) want some of that too.
How do you stand the noise of the 10k? :o)
>; the CPU takes 95W allegedly, and I
>will be running a GTS250 graphics adaptor.
Only at peak. Based on your spec given earlier, adding in your 4 hard
drives, I guesstimate via the highly accurate finger-in-the-air method
that the PC would use at most 120W when doing stuff like running Word.
I think a 750W psu is overkill. For that spec I would choose a 450 or
500W. Depends who made the 750W unit - it won't have been made by
Corsair. All they have done is stick on their own label and stuff it in
a flashy box.
Have a read of
http://www.tomshardware.com/reviews/inadequate-deceptive-product-
labeling,536.html
This review of 20 PSUs at Anandtech is good. I'd expected to fall
asleep, but it's well written, has interesting photos, and they identify
the OEM underneath the flashy stickers.
http://www.anandtech.com/casecoolingpsus/showdoc.aspx?i=3487
Right away, you'll see how similar supplies from different makers look
inside.
HTH.
--
Mike Tomlinson
Jaimie Vandenbergh
wrote:
>* Is it heavy (a sign of quality)? Sniff it - does it smell burnt
>(you'll get some odd looks)? If so, reject it.
Although I'm with you with the rest, I'll quibble about this pair,
both of which are down to logical fallacy:
A light PSU almost certainly won't be any good; a good one (over 200W)
will be heavy; but a heavy PSU won't necessarily be good.
The burnt smell can just be lingering from the assembly, if the device
happened to be made very recently. If one PSU smells worrying, do try
another one of the same model. Burnt smells from failure in QA will be
per-unit, not per-model. Also, different people have different levels
of knowledge of electronics smells! Perfectly fine kit can have a very
odd whiff to the inexperienced.
Cheers - Jaimie
--
"What happens if a big asteroid hits Earth? Judging from realistic
simulations involving a sledgehammer and a common laboratory frog,
we can assume it will be pretty bad." - Dave Barry
Mike Tomlinson
<m...@privacy.net> writes
>Two words: Suspect Roger.
Oh, it's Woger alright.
> What is it about Albasani that attracts trolls?
The anonymity, last refuge of cowards like Woger.
--
Mike Tomlinson
Mike Tomlinson
Vandenbergh <jai...@sometimes.sessile.org> writes
>A light PSU almost certainly won't be any good; a good one (over 200W)
>will be heavy; but a heavy PSU won't necessarily be good.
It's a guide Jaimie, not a cast-iron rule. A maker which hasn't tried
to skimp on heatsink material usually produces a heavy PSU, as well as
those that use a wound component for the PFC.
>. Also, different people have different levels
>of knowledge of electronics smells!
Was just starting on a reply when I saw you'd made the very same point.
There is quite a difference in the whiff between a freshly made PSU and
one that's met its maker in smoky fashion.
We had a tape drive psu turn its toes up recently. Not only did that
stink out the building, the black marks it made on the inside of the
case were quite impressive. But what disturbed me was that it had burnt
up sometime during the weekend and was _still_ emitting noxious smoke
when we came in Monday morning. It was in a locked room with no smoke
detector.
--
Mike Tomlinson
Rob
"Multithreaded" <m...@privacy.net> wrote in message
news:14g2i5dvln2g93sfv...@4ax.com...
Over the years, I've seen most makes of PSU give up the ghost.
The better makes usually go out with a whimper and don't damage
anything. The cheapies sometimes go with a bang and take mobos,
gfx cards or HDs with them..
I've been running 4 Corsairs on various overclocked systems for
the last couple of years without problems, so although the sample
size is small, I can recommend and will be sticking with them unless
I have trouble in the future (at which point I'll go for another good
make.) The Corsair 7 year warranty means that'll still be in effect
long after the rest of the system hardware has become obsolete or
gone tits-up. ;)
HTH,
--
Rob
Jaimie Vandenbergh
wrote:
>In article <29g4i592qf4g7k27v...@4ax.com>, Jaimie
>Vandenbergh <jai...@sometimes.sessile.org> writes
>
>>A light PSU almost certainly won't be any good; a good one (over 200W)
>>will be heavy; but a heavy PSU won't necessarily be good.
>
>It's a guide Jaimie, not a cast-iron rule.
I know, but the rest of your post was so good I had to polish it!
>We had a tape drive psu turn its toes up recently. Not only did that
>stink out the building, the black marks it made on the inside of the
>case were quite impressive. But what disturbed me was that it had burnt
>up sometime during the weekend and was _still_ emitting noxious smoke
>when we came in Monday morning. It was in a locked room with no smoke
>detector.
Not a nice feeling going into a room like that...
I had something similar at university. I was in the postgrad computer
lab, and smelt something dubious but couldn't identify what it was
coming from. After a couple of hours of no further excitement, I left
and locked up. Next day it turned out that a (rather ancient and
crufty) MicroVAX had gone up in a sort of gentle smoulder, rather like
a charcoal burner's fire. It was completely gutted, but hardly any
external evidence. Very odd affair.
The more usual flames-out-the-back-of-the-monitor are much more fun.
Cheers - Jaimie
--
"I think man is the most interesting insect on earth, don't you?"
-- Marvin the Martian
John Jordan
>
> A light PSU almost certainly won't be any good; a good one (over 200W)
> will be heavy; but a heavy PSU won't necessarily be good.
There are two possibly-unrelated definitions of "good" here too. You can
have a expensive PSU with excellent regulation and high maximum power
output that still has a design fault likely to fry components when it
fails. And you'll never know until long after they've stopped selling
that PSU.
Other than the Bestecs (which were relatively heavy), PSU failures
causing component damage are rare, even with crap PSUs. I can't say I've
seen any correlation in practise between PSU cost or weight and
component death. I'm not even convinced that cost is a reliable
indicator of lifespan, although those FSPs do seem to go on forever.
--
John Jordan
Johnny B Good
from =?iso-8859-15?Q?_=3E=3C=28=28=28=B0=3E?= <nos...@butfish.com>
contains these words:
====snip====
> In the main the usual rule of "you get what you pay for" comes into play.
> Buy a Q-TEC PSU 650W Triple Fan (Gold) for about �25 = expect trouble.
> Trouble can be big too. I've seen many a Bestec psu failure kill
> motherboard and more.
The Bestec psu fault was in a league of its own. Only the MoBo would
fall victim to the infamous 5VSB 'stealth fault' (where the 5v would
start creeping upwards past the 5.5v maximum through 6 volt, sometimes
landing up as high as 9.3 volt in extreme cases (7volt for 10 seconds
being the maximum voltage stress rating for TTL chips <and possibly the
same for their cmos equivilents>)
> Worst I saw was a psu failure that killed every part in the PC except for
> a memory stick and the floppy disc drive.
That last is an extremely rare type of event with PSU failure. From a
damage limitation PoV, the best type of PSU failure mode is the totally
out of the blue "Flash Bang" mode where the HT switching transistor(s)
had decided not to switch off when they were supposed to.
The energy such a onetime 'switching event' can transfer is extremely
limited so doesn't create the expected 'spike' on the LV output windings
of the ferrite transformer which feeds the LV rectifier/smoothing
capacitor packs before the safety fuse gets blown to smithereens.
Out of all the repairs involving replacement of an ATX PSU (I never saw
this with AT PSUs) which had gone 'flash bang', I never came across a
situation where the PC didn't immediately work perfectly fine after
being rejuvenated by the new PSU (other than the need to run CHKDSK to
mitigate the effect of having the plug pulled on the OS).
Attempting to repair a blown PSU by replacing the safety fuse alone, is
a complete waste of time unless this was the result of the user
inadvertantly moving the voltage selector switch from the 230v setting
to the 115v setting.
Even so, you'd be very lucky indeed if the safety fuse had actually
managed to avoid overvolt induced damage to the HT module. I had one
such type of 'failure' where, much to my amazement, replacing the blown
fuse (and removing the voltage selector switch wires ;-) did actually
clear the fault. Testing with a DMM beforehand had suggested that it
would be worth trying a replacement fuse otherwise I wouldn't have
bothered.
--
Regards, John.
Please remove the "ohggcyht" before replying.
The address has been munged to reject Spam-bots.
Nick
"Johnny B Good" <jcs.comp...@plugzetnet.co.uk> wrote in message
news:3130303037373...@plugzetnet.co.uk...
> The message <op.u4qtljeucnngb9@xp64>
> from =?iso-8859-15?Q?_=3E=3C=28=28=28=B0=3E?= <nos...@butfish.com>
> contains these words:
>
> ====snip====
>
> Attempting to repair a blown PSU by replacing the safety fuse alone, is
> a complete waste of time unless this was the result of the user
> inadvertantly moving the voltage selector switch from the 230v setting
> to the 115v setting.
>
certainly well aged.
Only to see what might happen. PC was a P1 and all stripped bare on the
kitchen table IIRC.
Reconnected all and pushed the GO button.
I'm glad I wasn't weariing safety specs (prob. not invented then). The
indoor pyrotechnics were quite spectacular. Definitely and without doubt
released the magic smoke :)
Now experiments are restricted to my workshop. By strict order of the
management.
The lingering smell is long since gone but the scars remain.
On the kitchen table.
Multithreaded
>Yes, the anode cap of a 6146B does bite and stays charged for a surprisingly
>long time. I suspect that's why they made the holes in the '101ZD's PA cage
>just the right diameter for a 1kV insulated screwdriver to discharge the
>things first.
Wow! What a coincidence - I had a '101Z in the early to mid 80s, and
recently was trying to remember what the final valves were. What was
the driver valve? It began with 12, [1] and mine developed an internal
short, causing a click and momentary loss of power on key-down. I was
a CW freak in those days; the only thing I ever had real talent at,
literally.
[1] 12AT7? AQ7? Geez, I used to have a superb memory for numbers of
all sorts, particularly transistors, valves and telephone numbers.
Must be my age (46).
--
Multithreaded.
Multithreaded
wrote:
>Lucky you :o) you normally only get one bang. Have you looked inside to
>see what let go? I'm also interested in knowing the make and model, if
>you still have it.
I looked through the grill and found no obvious signs of damage. It's
an Enermax EG465AX-VE(G).
>>Memtest86+
>>reported that the CPU and RAM were unharmed, miraculously.
>
>IME, damage due to a psu going bang is highly unusual. You'd have to be
>very unlucky if it did. There are (should be) crowbars across the
>outputs, so that if the voltage rises above a certain point, they are
>instantly shorted out to protect the connected equipment. Cheap PSUs
>either won't have these at all, in which case everything takes the
>brunt, or cheap crowbar devices are fitted which react too slowly, in
>which case it's pot luck.
I agree. I've only had a PSU go bang on two occasions in fifteen years
of PC building, as previously stated. I have several of my PCs running
on generic PSUs for years - only with light use, though - and no
problems at all. This includes two of the dreaded Q-TECs.
>But as you're going for the SoAM3 Ghia upgrade anyway, everything'll be
>replaced with new, won't it? :o) The reality is that you can't trust
>that motherboard (the one the old psu went bang on) and its parts any
>longer. Give them to somebody you don't like.
Agreed, again. I'll retire the motherboard; was going to anyway - see
my other thread. I don't dislike anyone enough to inflict the thing on
her/him (except the person who assaulted me in '95 and ruined my
hearing, but I assume he wouldn't be into computers anyway, heh). I'll
hold onto the drives, as they're all backed up, and may buy a backup
147GB 10K SCSI HDD, too, when I have more money. I just hope it hasn't
had its life shortened. All disks (even the floppy!) have had thorough
diagnostics performed on them over the last two days.
>An annoying little turd named westom1 will be along any second now to
>tell you that a whole-house surge protector would have prevented your
>PSU going bang while simultaneously announcing the Second Coming in a
>blaze of glory (which you missed 'cos of the smoke pouring out of your
>PSU). Take no notice, he's an idiot.
I have many surge protectors in the house. What they have to do with
the price of fish is beyond me as far as this saga goes.
>>Anyhoo, my question. What do people on here regard as good, and above
>>all, reliable, high power PSUs?
>
>It's a difficult question (see my thoughts re. power below.) IMO,
>quality is more important than power rating. I won't go into it again -
>it's been done to death here in the past - but I'd take a high quality
>350W PSU anytime over a cheapo 650W. The 350W may have a lower wattage
>rating on the label but it'll be better built and able to withstand a
>sustained power draw of, say, 300W than the cheapo.
I never go by paper power rating in itself. Quality, weight, make are
what I try to go on, and even then I think it's mostly in the lap of
the PSU gods.
>There are other factors to consider. For example, You also need to look
>at whether the unit you are considering has split rails and if so, how
>they are split (the rating label will tell you.) You would be unwise,
>for instance, to attach all your four hard drives to one 12v rail.
My new Corsair is single rail - very desirable from my point of view,
given all the HDDs I have, graphics card, CPU, etc.
>Like hard drives, even good PSU makers can make turkeys from time to
>time. A decent, consistent maker is Fortron Source (FSP). They OEM
>(make PSUs for) many suppliers who then badge them with their own name,
>but often keep the FSP-xxx part number still appears on the rating
>label, which is how you can tell. I'd take a FSP-derived 350W over a
>"1000W" Jeantech any time.
I concur yet again. I've heard many good things about Fortron over the
years, reading this NG, but have never seen one for sale on any of the
usual sites I haunt.
>If you can, try and buy sight seen rather than online. Eyeball the unit
>- does it look well made?
I took a chance and went on the personal recommendation of a nice
young chap at Scan. Due to my condition, getting out of the house is
usually difficult, and PC stores are rather thin on the ground in
these 'ere parts of Darrkest Zomerrset. The Corsair does, however,
look and even smell, well-made.
>* is the packaging "cool" and flashy? Is it aimed at the modder crowd?
No, I don't think so. I hate those neon cases, and other modder stuff.
>* Cheap fans look cheap. If it jams up, the PSU will overheat and you
>may not realise until it decides to leave this vale of tears.
The fan - there is only one of them - is huge (120mm?) and super-quiet
- at least to my failing ears.
>* Is it heavy (a sign of quality)? Sniff it - does it smell burnt
>(you'll get some odd looks)? If so, reject it.
It smells of victory - no, read, technology. Mmm! I love that smell.
Can you get it bottled? If so, I'd wear it as aftershave.
>* Are the output leads a decent length? Some mobos put the connector in
>weird places. Cheap PSUs usually have short, thin leads.
They're a bit too long, if anything, and my case is a monster in size
terms.
>* Do you need 20 pin? Is the 24 pin header on the PSU split so it'll
>fit 20 or 24 pin mobos?
Existing motherboard is 20 pin. The ATX connector on the Corsair has a
detachable 4 pin block, so it copes with old and new equally well -
what a life-saver.
>* Do you need SATA connections?
No. I use removable SATA caddies which take Molexes, which I much
prefer. The Corsair comes with eight SATA connectors which are merely
a nuisance for me.
>> I like the Corsair because it is very
>>quiet, and hope that running it at nowhere near top capacity will make
>>it last longer than its predecessor
>
>The best thing to do is to try and make sure it'll fall within the
>efficiency curve. Putting a 750W psu in a system that will only draw
>200W most of the time is running the power supply very inefficiently.
>You want to try and aim for the middle of the curve, then you have some
>in reserve for peaks and possibly to carry forward to your next upgrade.
The packaging gives an efficiency curve, so I realise that I will be
running it at the low end of its efficiency graph. I don't care. At
worst it's 70% efficient, and I'd rather have a lower probability of a
pop than saving a few Euros per year in juice. The PC keeps my living
room warm, too, so I save a little on gas - I have a really
inefficient gas central heating system dating back to '76, so I don't
mind wasting a few ergs in waste electricity.
>> (an Enermax 465W from 2004).
>
>It's worked for 6 years. Don't think it owes you anything. I'd be
>interested in seeing a photo of the insides if you can.
I don't feel hard done by. I merely regret not rebuilding the primo
machine last week; if I had, the double pop would never have happened,
at least not on my number one PC.
I've just bought a digital camera and need an excuse to use it in
anger. I'll try and get you the photo you want, and put it on my
website, which I will email to you (I don't like the whole world
seeing how dreadful my PHP and HTML skills are). Or I could just sent
the pic as an attachment, but I want an excuse to put in some
unnecessary PHP and Javascript, heh.
>> I
>>just want reassurance that I haven't jumped from the frying pan into
>>the fire. I need a high power PSU as I run four HDDs, including the
>>current-hungry 10,000RPM SCSI unit
>
>good idea, but as above, if you get a split-rail supply, try and
>distribute 12v equally across the rails.
Due to the case geometry, I tend to distribute the HDDs across several
leads, so rail overload isn't an issue. I've never really given that
much thought, though, so thanks for raising my awareness. And the
Corsair is single rail, but I will still use several cables, for the
above reasons, and also to reduce any I-squared-R losses, and voltage
drops.
>I have a 10k rpm 147gb broken drive to hand. This wants 1.2A on 12v, a
>fair amount. You don't want to put four of these on one 12v wire -
>that's 4.8A when all the drives spin up, and the mobo and the optical
>drive(s) want some of that too.
See above.
>How do you stand the noise of the 10k? :o)
Two reasons. I'm going deaf, so noise is the least of my problems.
Also, the fans in all of the drive caddie coolers tend to dominate
over the high-pitched whine of the SCSI screamer, as I call it. I
also, perversely, find the continual head seeking chatter reassuring.
It reminds me that all is well with the world, and that I can still
hear at all.
>>; the CPU takes 95W allegedly, and I
>>will be running a GTS250 graphics adaptor.
>
>Only at peak. Based on your spec given earlier, adding in your 4 hard
>drives, I guesstimate via the highly accurate finger-in-the-air method
>that the PC would use at most 120W when doing stuff like running Word.
When I get cold, I just fire up Prime95 on both cores (is there a
Prime95-type prog that will do all four cores of the new AM3 system,
or do I simply run two instances of the old program?). I also defrag
all four drives simultaneously, and look at the voltages on the
motherboard utility. The voltages never drop above measurement noise,
which has always amazed me. I can dry my wet hands on the air that
comes gushing out of the case and PSU fans when I do this. Who needs
hair dryers?
>I think a 750W psu is overkill. For that spec I would choose a 450 or
>500W. Depends who made the 750W unit - it won't have been made by
>Corsair. All they have done is stick on their own label and stuff it in
>a flashy box.
I guess it's too late now. I've bought it, and have had it running
today on the old motherboard which appears to be working perfectly
running UBCD4Win and a 64 bit OpenSUSE Live DVD, but the native XP
install now has serious probs as I've already discussed. As a
corollary, the PATA disk that I originally thought was faulty, then
found out it was the motherboard's PATA controller, is defective after
all. Go figure. I just can't deal with the weirdness of it all.
>Have a read of
>
>http://www.tomshardware.com/reviews/inadequate-deceptive-product-
>labeling,536.html
>
>This review of 20 PSUs at Anandtech is good. I'd expected to fall
>asleep, but it's well written, has interesting photos, and they identify
>the OEM underneath the flashy stickers.
>
> http://www.anandtech.com/casecoolingpsus/showdoc.aspx?i=3487
>
>Right away, you'll see how similar supplies from different makers look
>inside.
I'll study those links when I'm in a calmer frame of mind. I really
need to know this stuff. Thanks.
>HTH.
I'm sure it will.
--
Multithreaded.
Multithreaded
wrote:
>Over the years, I've seen most makes of PSU give up the ghost.
>The better makes usually go out with a whimper and don't damage
>anything. The cheapies sometimes go with a bang and take mobos,
>gfx cards or HDs with them..
As I've found.
>I've been running 4 Corsairs on various overclocked systems for
>the last couple of years without problems, so although the sample
>size is small, I can recommend and will be sticking with them unless
>I have trouble in the future (at which point I'll go for another good
>make.) The Corsair 7 year warranty means that'll still be in effect
>long after the rest of the system hardware has become obsolete or
>gone tits-up. ;)
>HTH,
All good to know. I'm grateful for your post.
--
Multithreaded.
Multithreaded
>12BY7A medium signal pentode. A lot of the problems on the '101Z/ZD were
Ha! I just remembered the designation this second, then came on here
to find your confirmation of the fact. At least I'm not losing my
memory totally.
>caused by the relays or simple dry joints, although there was one
Yes, the antenna changeover relay developed a fault whereby I got a
farting noise on ten metres (my favourite band in those days) going
from send to receive - annoying when operating semi-break in on CW as
I always did.
>particularly nasty fault with the DC blocking capacitor between the driver
>and the grid of the finals that used to short out, pushing the finals into
>full conduction. It was then a race between the finals and the HV windings
>of the transformer burning out, which is why replacement transformers are
>like hen's teeth because the finals inevitably won this little test of
>strength. Another particularly annoying fault was the two halves of the
>band-switching shaft becoming misalgned. Hilarity then ensued, of course,
>and was the cause of many an otherwise good 101 being sold for scrap.
Wow. I'm glad I sold my rig in '89 before the faults you describe
could take effect.
>Harry, G3LLL was the expert on these in the day. I still have an FT-101ZD
>(MkIII with WARC) in the workshop - I sold most everything else about four
Mine was WARC - I loved 30m - but was only the Z version: I couldn't
afford the digital readout in those days. I was a poor struggling
physics student, after all.
>years ago but I'll not go into reasons on here. I never really got over the
>hybrid radios and still enjoy the odd QSO with it, the radio, as far as I am
>concerned, epitomising the glory days of the service. There's nothing quite
>like the smell of a glowing anode heating dusty glass... yes, I suppose I
>was pushing those 6146s a bit :o)
I pushed the 12BY7A - or its replacement that I had to put in - so
hard it glowed purple! I reckon I could have worked the world on CW
just using that valve as the final. Happy days. I also hated the solid
state rigs of the time: too sensitive to the poor VSWR that I
subjected the valved rigs to, routinely. They seemed to thrive on my
tune for max smoke strategy.
--
Multithreaded.
John Jordan
>
> When I get cold, I just fire up Prime95 on both cores (is there a
> Prime95-type prog that will do all four cores of the new AM3 system,
> or do I simply run two instances of the old program?)
Recent versions of Prime95 will load at least four cores. Maybe get a
heater instead though, as this really stresses the motherboard VRMs :-)
--
John Jordan
bill coombes
"Chronos" <m...@privacy.net> wrote in message
news:hfssrv$9n2$1...@news.eternal-september.org...
><(((�> wrote:
> Sorry to disappoint you but I won't rise to the bait.
.. Fortrons were lasting too long past the warranty..
Yeah, lol, that's a really unforgiveable manufacturing mistake.
Rob
"Johnny B Good" <jcs.comp...@plugzetnet.co.uk> wrote in message
news:3130303037373...@plugzetnet.co.uk...
Sometimes, fuses do 'just fail'. They're only thin bits of wire, just thick
enough to not burn out at the rated current, so run warm. The thermal
expansion/contraction can eventually break the wire, usually at the end
caps where the stress is greatest.
--
Rob
Johnny B Good
from "Rob" <no...@nowhere.noway.com> contains these words:
> "Johnny B Good" <jcs.comp...@plugzetnet.co.uk> wrote in message
> news:3130303037373...@plugzetnet.co.uk...
> > The message <op.u4qtljeucnngb9@xp64>
> > from =?iso-8859-15?Q?_=3E=3C=28=28=28=B0=3E?= <nos...@butfish.com>
> > contains these words:
> >
> > ====snip====
====yet another snip====
> > Attempting to repair a blown PSU by replacing the safety fuse alone, is
> > a complete waste of time unless this was the result of the user
> > inadvertantly moving the voltage selector switch from the 230v setting
> > to the 115v setting.
> >
> > Even so, you'd be very lucky indeed if the safety fuse had actually
> > managed to avoid overvolt induced damage to the HT module. I had one
> > such type of 'failure' where, much to my amazement, replacing the blown
> > fuse (and removing the voltage selector switch wires ;-) did actually
> > clear the fault. Testing with a DMM beforehand had suggested that it
> > would be worth trying a replacement fuse otherwise I wouldn't have
> > bothered.
> Sometimes, fuses do 'just fail'. They're only thin bits of wire, just thick
> enough to not burn out at the rated current, so run warm. The thermal
> expansion/contraction can eventually break the wire, usually at the end
> caps where the stress is greatest.
There might be an element of truth to that in the case of a single
mains supply voltage PSU where the fuse rating doesn't have to be
determined by the current demanded at the lower voltage. Even so, the
ends of the fuse link are cooled by thermal contact with the end caps
which concentrates the heating effect to the middle which might
routinely glow on switch on and so oxidise which would make it
effectively thinner, leading, eventually, to the 'tired fuse' syndrome
to which you're referring.
In the case of a dual mains voltage PSU (230/115v), the fuse is rated
at twice the current it would otherwise be for 230v only operation. In
the UK, the 'tired fuse' syndrome is virtually non-existant on a dual
voltage PSU. Even in the case of 230v only PSUs (which are identical to
the dual supply type but without the voltage selector switch circuitry)
there's no guarantee that they're not still fitting the larger fuse
value appropriate to the 115v option.
Unless you're dealing with a very expensive PSU (requiring an equally
expensive replacement) or else a weird form factor PSU, I wouldn't
normally waste the time on such a repair. IME, replacing such a fuse is
simply a waste of time (and a perfectly good fuse).
The only reason I bothered at all in the case of the voltage selector
switch accident was that I felt there was the slimmest of chances that
the fuse might have actually succeeded in that most unlikely of tasks,
protecting the HT module components.
The DMM testing had suggested that this might actually be the case and
worth feeding the PSU another 'sacrificial lamb' to see if my act of
appeasement in the form of a voltage selector-ectomy had been a
sufficient repair. That it was, quite frankly, rather surprised me.
HTH