talking about models made after 2005
had a PSU die on me in 2002 or some thing but it was from 1995 and in a overloaded system and I don't think it damaged any thing soundcard was fine and still use it.
I see Silverstone's contemporary PSUs are rated for 12years of 24/7 use.
if its a good-quality PSU then its fine, it'll probably last forever. if its a cheap chink shit PSU then its not fine, it will eventually die, but you should preemptively replace it with something decent anyway.
reason I ask is people with old retro computers seem to use them until they die and then just replace them and system is fine.
I would understand if older PSUs from the 90s where shit but I would imagine modern ones would shut off before they fail? just wondering I have like 2-3 old ones from 2008ish decent brands like antec was thinking of using 2 or 3 of them because they only 450-550w in a SLI system one for each GPU and one for CPU and wondering if its safe to just keep using them like that for ages...
sure I could just get a 1000w one for 100$ or some thing but I don't like to be wasteful
was going to cable tie one of them into the 5inch CDrom bays of my case and use a adaptor to run them all off one wall plug so its neat at the back?
might cut a hole in motherboard IO shroud and run cable from main one back inside case and split them out there and have some thing like this at back of case down the bottom or cut a hole in IO shroud for it as well.
idknow that 12 years 24/7 use is average failure time.
sure others die in 6 years and some last 20years or more.
just wondering what actually happens when they die.
always hear horror stories of them dying and taking all your components with them but is that actually the PSU or surge or mains power problem?
New caps all use water based electrolyte meaning they live only few years and in most cases PSU environment is absolutely brutal to them so they don't last even that long.
Usually they don't blow anything up when they die. Usually isn't always, granted, but its not anything I'd stay up at night worrying about.Well, unless its one of those "700W" PSUs thats lighter than your mouse and cost $20, or anything made by Deer. Those are firetraps.
if you want to run a retro system you can use a new ATX PSU in it. New PSUs don't have a -5v rail but there's adapters you can get on ebay that include a little voltage converter that provides it. (These are not a hazard because they're so low power)
after the capacitor plague around the turn of the millennium PSU makers (and mobo makers) started not only buying Japanese parts rated for 105C but advertising that fact. A lot of modern PSUs use quite a lot of polymer caps, too. And modern PSUs, being more efficient, give them a less thermally-demanding environment, in all but the super-high-wattage units. Yeah, they won't last until the heat death of the universe, but they'll last quite a long time.
>New caps all use water based electrolyte
Hello the mid-2000's called they want their caps back
Other than caps and fans, there isn't anything else to go bad, barring some catastrophic event. A cap going bad can take out other components though. If you want to get the most out of a supply, then all electrolytics should be replaced after 5 years or so. Sooner if your the super paranoid type.
I have see two psu's die and take the motherboard with it. So i have started replacing my psu every 4 years or so.
my caps are rated at 85w Silverstone 800w SFX-L going in a tiny dan a4 style case.
think it will get too hot? I can replace the fan on it with a thicker better one as its out of warranty.
885 deg or maybe 82deg cant remember.
I think case temp in these tiney cases can get to like maybe 60deg as hardline tubing melts in them that's rated at 50c.. but not sure if it will hit 80c ;/
it's not about the total watt rating or efficiency, it's likely your PSU died because you overloaded a rail, Older systems relied heavily on the 5v rail of the PSU, the 12v, not so much, this is why if you wanted to power on an old 266mhz bad boi, you could kill your PSU or at least suffer brown outs.
the inverse is true, a PSU made for that type of system trying to load up everything on a mincey 12v rail would yield similar results. its not that the PSU was dying, but rather the new generation of components you plugged it to were too demanding for its amps per rail.
its equally likely it just outright failed on you, or you were drawing too much power over pci(e)
right that's a good point.
I guess ill check what the amps of the 12v rail is on my late 2000s psus im thinking of using.
>if you want to run a retro system you can use a new ATX PSU in it. New PSUs don't have a -5v rail but
It depends on what retro system it is. Not all of them need -5V, it's mostly if you have shit like a Commodore PET because -5V is needed for the RAM, which will actually self-destruct if that voltage is missing.
A lot of early chips were like that, needing +/- 5v and +12. CPUs had gotten away from it by the early 80s (needing only +5 was a selling point, it made logic design simpler) but lots of minor glue-logic chips kept using old processes for some time. As did some expansion cards, modems used to have a reputation for it.
Actually as late as 2008 MSI put out a motherboard with a sound chip that used -5v. PSUs without it were showing up by then and people were surprised when their onboard sound quit working when they got one.
Dust buildup is a huge problem with these.
Over temp might fuck a regulator sending high voltage down the rails.
A proper psu has fuses and thermal shutoff, these 10year old likely not.
So keep it clean or maybe even install a more strong fan.
More heat also kills the caps significant faster
>CPUs had gotten away from it by the early 80s (needing only +5 was a selling point, it made logic design simpler)
Eh? The 6502 and Z80 came out in the mid-70s and ran off a single +5V line.
any way to tell if the PSU is damaged or worn out beside owning it and knowing how long you used it and in what conditions?
prob used these 3psus for like 3years 24/7 in hot Australian summers but in cases with sides off and good fans etc.
also how safe is changing fan in PSU obviously not talking about touching electrical shit that's bad but like could a dif fan fuck up its power curve and make it not cool its self enough?
>>any way to tell if the PSU is damaged or worn out beside owning it and knowing how long you used it and in what conditions?
Well actually you can't even tell then. Some PSUs will be fine when used in severe conditions, some will be failing. You have to stick a multimeter and an oscilloscope into em to really have any idea. Hope you also did that when the thing was new, so you know what the baseline is, so you can tell whether a PSU that has a certain amount of ripple or what have you is slowly deteriorating, of if the design is just like that.
>also how safe is changing fan in PSU obviously not talking about touching electrical shit that's bad but like could a dif fan fuck up its power curve and make it not cool its self enough?
use your common sense. Get a fan similar to the original if ytou're gonna hook it up to the same output the original fan used. Doesn't have to be identical but the rated current and airflow should be similar. If you're willing to run some wires outside the PSU and hook it up to the main 12v rail then you can use anything that's at least as powerful as the original.
You might be able to get away with something weaker but really quiet if you know for certain you'll never be putting anywhere close to the PSUs nominal rated load on it. I did that with an old Antec, replaced the fan with a really slow one, and it was fine because it was a 650W PSU powering a GPU-less system with a 65W CPU and two HDDs. It could never suck more than about 15% load out of the thing. PSUs are a bit less efficient at low load (unless you splurge on an 80+ titanium one) but the absolute numbers are so small that it doesn't matter. It doesn't take much of a fan to get rid of fifteen watts of heat.
then again you might want to be a bit more conservative if you have very high ambient temps.
old prebuilt computers can have really bad psus they usually used the cheapest noname chinkshit they could find
>could a dif fan fuck up its power curve and make it not cool its self enough?
psu fans are bulk bought generic chinese oems as long as it fits (they are often odd sizes) you are good
you could also buy a fanless psu just be sure to install it in the correction orientation so heat can rise, anyway my recommendation on replacing psus is
generic psu: inspect visually when cleaning for bulging/exploded caps
name brand psu: just use it until any kind of power event (brown out/black out/surge) then inspect
evga/seasonic: fuckin plug it in and forget about it until the heat death of the universe since they put in every practicable safeguard
!!! don't mix modular cables even within the same brand !!!
I got a 80 platinum PSU for 100$ shipped.
custom cables are retarded use cheap shit from ebay and don't care about asthetics
ive seen multiple youtubers get cable mod cables where the wires are done wrong and they don't work and that's to youtubers the general population prob get even more shit from them.
they prob just bribe customers with free shit to stop them talking on forums about how bad their Vietnamize or chink cable weavers are.
nothing is safe, OP.
Embrace the horror.
im just sick of dumb people on here that say PSUs die every like 4years and to buy a new case and hole system that often.
I keep my cpus for like 10years and power supplies for even longer
I feel there is lots of people here who spent 1/2k on gpus and are paranoid they will get hurt buy their PSUs but the gpu will be obsolete and literally not have a driver or run games before your PSU dies even a cheap one.
>implying I meant just PSUs
life is horror, user. Every step you take. Death is always right in front of you, shaking its skinny butt for you to nut inside it.
lyl
>If you're willing to run some wires outside the PSU and hook it up to the main 12v rail then you can use anything that's at least as powerful as the original.
He could also just solder it to to the PCB.
At some point, someone at Commodore figured out that their PSUs were shit so they outsourced them. The C128 PSU was made by Mitsumi, and they also used outside suppliers for the Amiga PSUs.
Shitty OEM PSUs are about the fastest way for a company to go bankrupt, so most tend to be at least mildly decent (albeit 250w), or the company would be infamous for mass failures and the like.
Keeping it cool is the key thing, I believe a 10C increase results in halving the lifespan or thereabouts. Fans are easy to replace, just don't touch anything inside the PSU; I suggest working with one hand behind your back or in your pocket for extra safety (minimizes the possibility of electricity traveling across your heart). Leaving the PSU sit unplugged for a while is also wise. I'd take apart the old one first to see what kind of termination they used, or just solder the new fan's leads to the old fan's.
>At some point, someone at Commodore figured out that their PSUs were shit
If I had to guess, I assume it had to do with Captain Jack's retardation.
honestly if its new build and you have some really old psu like 10+years old better just get a new one, they cost like 50$ for decent one these days anyway...
It was. He believed in as much vertical integration as possible, including making their PSUs in-house.
You'll also want some form of relay for each secondary to turn them on slaved to the primary mobo one, and don't mix PSUs together;
Ok, probably:
>PSU A: Mobo + GPU 1
>PSU B: GPU 2
Bad:
>PSU C: Half of anything (such as GPU 3)
>PSU D: Other half of thing
Note that some PSUs have multiple rails for some retarded reason, which have limits which must be accounted for.
I'd also have some case fans running off of each slave PSU lest they get stuck switched on, as the GPU fans and power regulation are driven by the mobo PSU. I once had a similar setup, until the system shutdown while I was away, but the second PSU kept feeding 12v to one of the cards. No clue how hot it got, but the temp has high enough to discolor copper heatpipes to a nice gold color, and I burnt the shit out of my hand when I touched one.
Tldr; just buy the PSU, it's probably cheaper anyway
you just cross the pins on the psu not plugged into the motherboard leaves it on all the time but ill just get a kettle cable with a switch on it to turn all power off the entire case.
I don't understand the setup your describing sounds like you don't understand to do it and are larping worse case dream situation.
I use a PSU to power my vinyl pre-amp this way for instance. good point about the multiple rails thou ill check that out.
Cross the pins on the motherboard 16plug or what ever look on youtube how to do it just use a paperclip and cover it with some tape its easy.
also a PSU doesn't send power to a GPU when the GPU is not engaged from the motherboard the situation your describing is fanciful also even a 100deg core will not make a heatpipe burn your hand idiot the heatpipe will be like 60deg.
No. PSU should be changed every 2 years for maximal safeness.
your random LARPING does make me think of some thing thou. the GPU gets some power from the motherboard to like turn on and power some of it and then some from the extra 6/8pins the idea of a GPU being 30% powered by the motherboard and 70% powered by the 2nd PSU does worry me but maybe its convient and will leave it turned on and off when the system is. however you can get pcie extenders I assume for bitcoin mining that divert the motherboard power and power it via a molex instead of the motherboard that might be safer for the card but would make you dreamed up situation actually possible.
dreaming is a way of preparing for conflict
I swap out a 80 platinum every 6months just to be sure. if you going to spend 1grand on a gpu you should spend at least 500$ on psus a year.
>is it safe to use PSU until it dies?
no you need to check its lifebar regularly
just select the PSU, rightclick and pick "HP" from the context menu
Jumpers aren't going to shut your GPUs off if your system shuts down
Pic related: OEM copper red
Pic related: I-fucked-it brass
what am i looking at here
This card was being powered on pci risers as you mentioned. I thought system was off because all fans were off (2nd PSU was a seasonic that automatically shuts its fan off under low load). No fans & no lights = everything's off, right? Wrong
The copper pipes are no longer copper colored
Perhaps this picture illustrates it better.
Then by all means, plug a second PSU into a high power GPU (I nuked a 280x) and shut your computer down while forgetting to switch of or un-jumper the second PSU, because nothing will happen, as the board designers obviously accounted for this failure mode and have taken precautions accordingly. Just don't be a nigger and try to RMA a board you knowingly killed
i dont think that they are that bad. i have an audio amp with almost 40y old caps and it works fine.
This thread makes me nervous since my PSU is a mid tier Corsair that's 10 years old. On the flip side, getting a new high efficiency modular one would same a shitload of case real estate.
I replace my PSU every 3,000 miles for preventative maint.
Anyone knows how loud Cooler Master Elite V3 is? Youtube people only review stupidly expensive PSUs.
yeah
They are switching mode power supplies/regulators.
Unless it's capacitor related, some boost inductor, transformer or regulator/sensor IC failing, the usual case of failure from old age is just the switching mode transistors dieing. (Unless it's a fuse.)
There are no problems using several PSUs in one system for GPUs for example, just don't forget to also connect the grounds of both PSUs externally and not just over the GPU.
I used to run a SLI setup off two PSUs, one for the board, drives and CPU, the other for two GPUs. For the longest time I didn't even switch the second PSU automatically, but powered on the GPU power before powering on the computer itself (just had the 24-pin header shorted to always turn on when the physical power switch was flipped) and nothing bad ever happened. Later I just connected both green wires of both PSU 24-pin connectors together so it auto switched from one switch.
Hope this is helpful.
you can get a little board with a 24-pin connector and a relay that will automatically start a secondary PSU without splicing wires, add2psu I think it was called.
Yeah I've seen them too, good advice. This was years ago though as just a dirty hack in the first place.
>reason I ask is people with old retro computers seem to use them until they die and then just replace them and system is fine.
Unless you fry your ICs, like RAM because before the PSU failed, it also spiked a rail. I.e. 5V on C64.
It has more to do with not letting the caps go unused for too long and drying out.
Why is this guy surprised that Korean RAM died on him? Anything Korean from that era was junk.
That thread has a lot of surprisingly uninformed posts.
no your dumb copper comes in dif colors you idiot. a gpu can not be powered on or run by external cables only it needs power from motherboard and data to turn on your a LARPING faggot.
Someone on Youtube had a 486 PC where all the RAM sticks had died. Turns out they were Goldstars.
Stupid question, but why doesn't he just get one of those Lo-Tech XT cards? They have a modern SRAM in them.
IIRC the board was only designed in 2014 and didn't come to sale pre-assembled before 2016.
using psus that old is a waste of money.
They have an efficiency of like 75% at best if it's a good brand, and possibly even 50% if it's a no-name shit.
This adds up, losing 100W to heat compared to a new psu is $40-$70/year depending on energy prices and daily hours.
don't pay power bills and putting a 2-3 hunks of steal and copper and electronics into landfill is betting than buying a new one for the environment.
>putting a 2-3 hunks of steal and copper and electronics into landfill
they don't end up on a landfill, copper alone is too valuable.
You can buy seasonic titanium psu (96% efficiency at 50% load and 230V) and the savings will pay for itself in two years.
Most stuff from the 80s can be modded to use modern SRAMs. The amount of work required to do so varies depending on the hardware. Someone on Lemon64 did an SRAM mod on a C64 just for kicks, not one of the easier systems to do because of its relative complexity.
Mainly, you need a circuit to demux the data and address lines (DRAM has them multiplexed, SRAM doesn't) and leave the RAS/CAS lines disconnected as SRAM doesn't use those. It would be easier on a Z80 system like the ZX Spectrum because it already has no RAM refresh circuit (the CPU does that on-chip).
Holy god, why do tantalum caps blow out on IBM XTs so much?
IBM used understrength ones because it was cheap. It's probably the single biggest failure point in PCs/XTs.
can we stop talking about retro shit I don't really care if you 24mhz computer dies. im talking about how safe psu's from antec and in the 2006-2010 time period where and if I can use them for infinity safely.
Ray Carlsen sells power savers that will shut down your C64 if the brick starts going hellofamonkey. As far as C64 PSUs go, the European cheese wedge is worse than the US black brick for reliability.
What C64s had what PSUs?
Most European breadbins had the cheese wedge. The black brick was standard on US model breadbins while both PAL and NTSC C64Cs had a white brick. The cheese wedge overheats more readily than the black brick which at least has some vent holes in it.
The black brick is reasonably reliable if you take proper care of it, which includes giving it proper ventilation and not overloading it with power-hungry peripherals (in particular certain cartridges). Having said that, it still lacks any overvoltage protection for when it does fail, and that's why ultimately any stock C64 PSU is untrustworthy.
I notice that post-Tramiel Commodore outsourced all of their PSUs instead of making them in-house and never had any problems with them again.
im not using a PSU from 2006 when I was 4 years old if I was that young I would just get my parents to buy me a 1600w you retard.
im obviously a older neet eco warrior.
If you dont trust it dont use it, get a enermax enermax.com
not entirely ibm's fault, tantalums will go up like a matchhead for literally no raisin. bastard things can go to hell right alongside selenium rectifiers.
Tantalums made in the 80s weren't as good as modern ones because manufacturing processes have improved, but IBM were still at fault for using caps that were understrength to handle a 12v rail.
vcfed.org
This guy got lucky. Fired up like a champ after more than 20 years of not being powered on. Though it could also depend on the storage conditions it was in.
my shitty corsair vs with capxon caps died.
it didn't blow up or anything.
shit just crashes my pc when substantial load goes through it.
must be the ripple going nuts.
That would never happen on a C64. Many are the grim stories of people who took theirs out of the attic and found them dead.
Yeh if it does work, you will want to run it at least an hour. That's to ensure the caps are properly juiced up. Tantalums and electrolytics tend to not like going unused for a long time and a residual charge is a bit important to keeping them healthy. But again, like said it also does depend on the storage conditions and the large electrolytics used in PSUs are expensive, so those tend to be of higher quality than the small ones you would find on a motherboard.
yer I got one have a 800w SFX-L Silverstone that's made by enermax uses 85deg caps thou rather than 105deg Japanese ones thou but hopefully its ok.
>enermax.com
sadest thing thou enermax mad this same psu at the same time as the Silverstone 800w but kept the better capacitors and heatspreaders for it. guess the cheaped out on Silverstone because who cares if it dies :(
Just build a new PSU, you monkey. You'd be a retard to use any old PSU.
The 12V AC is mostly fine for use though, it's the 5V that you NEED to replace in order to even power on one, otherwise you're just a dummy.
It's 9VAC actually which is used to generate 12VDC for the VIC-II and SID (but HMOS versions only use 9VDC).