• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Time to change this PSU?

Status
Not open for further replies.
i wouldnt worry about unless it started crashing, blue screens ect. those numbers are within spec my seasonic is around the same on both my riggs.
 
It's old and it has ripple sure, but for now it works.

Hook up a multimeter to it and measure the +12 rail when the computer is under full load, use synth benchmarking tools for this, run P95 Small FFT and FurMark at the same time, see what happens with the voltage. The most basic multimeters will only display current value, but if you happen to have something a bit better you can use the memory settings to record min/max values and compare.

There are *other* ways to stress a PSU but I reckon you don't have access to some dummy loads, doesn't has to be anything super complicated for a burn-in test tho, I work with strings of light bulbs, pretty simple but it works if all I need is to find out if the unit has gone bad.
 
seasonic dude!
 
1000W for 250ms is 250J

which would need a 3000 uF capacitor (at 400V) which is many times more than typical.
 
250ms would certainly be enough to handle almost any "fluctuation" short of an actual power outage. And I agree, it would not be too difficult from a circuit design standpoint. I doubt it would add significantly to the cost either.

However, as Shrek is (I think) alluding to, I am not sure, at least with the higher wattage supplies, if the physical size of the bigger caps might be a problem fitting inside the allowed dimensions of an ATX compliant supply. I am assuming they would be larger to increase the surface area for better cooling. But it would not be the first time I assumed wrong. Today's capacitor technologies already allow today's caps to be much smaller than 10 or 20 years ago and still have the same electrical characteristics.
 
That Seasonic Focus is an excellent unit.
If you're worried I'd say just go ahead and buy it. After your use for this system is done in a few years you can just port that unit over to a new rig and keep this rig going with the Thermaltake unit.
I ported my 7-year-old Seasonic 650W to my new rig (it's a good enough unit that I could keep it going for another 3) and the computer that had the Seasonic is now using an even older 430W unit. It's just for family use and has a GTX 970 for the occasional game - though I'd only be playing very old games on it that the 970 could EASILY handle.
 
but might add $40-50 to the cost.
I don't see it costing near that much more. If it became a requirement, the big PSU makers would be ordering 100s of 1000s at a time - maybe even significantly more.

Don't forget, they would no longer be buying what they are now.
 
There is a cheaper alternative, namely super-capacitors on the output.
 
Its the physical size of that amount of capacitance would push ATX power supply sizes massively.

Looking at the Chemi Con capacitors from Mouser your talking 2.5" x 4-4.5" depending on total capacitance. That is a lot larger than most current capacitors in use especially in the <1000watt PSUs.
There is a cheaper alternative, namely super-capacitors on the output.
I am surprised a little that this wasnt pushed as part of say the ATX12VO standard seeing as there is a lot less cable clutter/mounting points freeing up some PCB space on both main board/connector and if it was wired in line with the output of the MOSFETS it would cover all devices in use.
 
I am surprised a little that this wasnt pushed as part of say the ATX12VO standard seeing as there is a lot less cable clutter/mounting points freeing up some PCB space on both main board/connector and if it was wired in line with the output of the MOSFETS it would cover all devices in use.
Cost, the #1 reason why most sensible innovations take time to reach mass market.
 
Cost, the #1 reason why most sensible innovations take time to reach mass market.
In that light - finding an investor who believes in that idea and is willing to fund the R&D and then help market and distribute it. No biggie! :rolleyes:
 
In that light - finding an investor who believes in that idea and is willing to fund the R&D and then help market and distribute it. No biggie! :rolleyes:
With this kind of advance, R&D isn't really needed. Just someone willing to make it happen and actively promote it to everyone who will listen.
 
@izy
i know some of the seasonics (past units) had some issues, might wanna look up what model to avoid.

while it might not be the latest unit, the corsair RMx are pretty good, if you can get the 750w under 100$

@Bill_Bright
once you go with RTX2000 and up (havent looked at AMD), transient load can be an issue,
and requires "replacement" of a psu, without having anything to do with a defective unit.

took swapping 3 different gpus (2080S), then 3 different psus (after i got a gpu with diagnostic leds),
to find the cause (psu).

seasonic/gigabyte/thermaltake (single rail gold/plat units with 650-850w) all "failed"
for (gpu) power delivery, with the corsair able to handle transient load.
 
@Bill_Bright
once you go with RTX2000 and up (havent looked at AMD), transient load can be an issue,
and requires "replacement" of a psu, without having anything to do with a defective unit.

took swapping 3 different gpus (2080S), then 3 different psus (after i got a gpu with diagnostic leds),
to find the cause (psu).

seasonic/gigabyte/thermaltake (single rail gold/plat units with 650-850w) all "failed"
for (gpu) power delivery, with the corsair able to handle transient load.
As someone who has owned many models of every generation of RTX, I can safely say that what you suggest only applied to the top tier models. The mid tier and mainstream models do not experience such problems. In the situation of the OP, they very unlikely to need a new PSU.
 
It seems that i had a RAM problem (bad OC) which caused my system to reboot time to time, i went and stress tested everything including TM5 overnight and i didnt had any errors till i stress tested the RAM with OCCT and it was starting to show errors after 20 mins (all other tests in OCCT were fine) , i changed some timings (went from CL16 to CL18 and lowered to 1.4v) and so far so good , no reboots no errors, so if everything works fine ill just keep this PSU till my next upgrade, thanks for suggestions ill keep them in mind when ill upgrade at the start of the next year.
 
Last edited:
@Bill_Bright
once you go with RTX2000 and up (havent looked at AMD), transient load can be an issue,
and requires "replacement" of a psu, without having anything to do with a defective unit.

took swapping 3 different gpus (2080S), then 3 different psus (after i got a gpu with diagnostic leds),
to find the cause (psu).
Now wait! I think it time to remind everyone (and perhaps get the thread back on topic) that the OP's current PSU is working just fine. There is no "cause" to find.
 
@lexluthermiester
true.
i always "forget" not everyone goes with xx60ti or above.

@Bill_Bright
my point was to show even units that are operating within specs/arent defective, might warrant a swap (with a different brand unit).
just because something isnt physically defective, doesnt always mean its working properly...
 
First and foremost, I totally agree when buying a computer power supply, we should ALWAYS go with a quality model from a reputable maker. And you can go through 100s of my posts (or just look here) and see where, when troubleshooting computer problems, I regularly recommend users swap in a known good PSU since EVERYTHING inside the case depends on good, clean, stable power - just to ensure they are providing good power, and to see if the problem stays or goes away.

And you can see where I regularly recommend going with a "reputable" brand (like EVGA or Seasonic "Golds" - not that "Gold" suggests better quality, it does not but that's a different discussion). That said, TT, while not known for top quality, is not a no-name generic brand, either.

However,
my point was to show even units that are operating within specs/arent defective, might warrant a swap (with a different brand unit).
Different brand? Why? Sorry but that really does not make any sense - on multiple points.

A reputable brand name does NOT ensure, or even imply the supply is more capable of supporting the connected devices. Nor does going with an off-brand generic imply the supply is less capable. About the best we can truly assume is (1) going with a reputable brand suggests that most likely the device "will" meet "published" specs and (2) it will be less likely to fail prematurely. But those are NOT givens or absolutes!

We must ASSUME, of course, the design specs meet or exceed the requirements to perform the task at hand. If not, the user failed to buy the correct size for the job. For example, a lawnmower engine, operating within specs and not defective, is not designed to pull a travel trailer (caravan) around. So "of course" you would need a different vehicle and bigger motor with the capacity to pull that load.

Again, the OP has stated this PSU has been working fine - suggesting it has the capacity to do the job, and is NOT defective. So why spend money to replace it just because it is a TT (a brand we don't like) and not a EVGA or Seasonic (brands we do) when the TT appears to be working properly now? That makes no sense.

just because something isnt physically defective, doesnt always mean its working properly...
Not sure what you mean by "physically" or "working properly". What other kind of "damage" to inanimate objects is there? :confused:

If a device is not damaged (physically... or otherwise?), "AND" the device is the right tool/size for the job, "AND" the user/operator is using it properly, then yes, the device will be working properly.

What do I mean by user "using it properly?" I mean he/she selected the correct size for the job, is feeding it the right fuel, and is operating it in the correct operating environment (adequate cooling, not in the rain, etc.)

Again, as I have also said 100s of times, the power supply is NOT an area to trim the budget. So I do agree that buying a quality supply from a reputable brand is ALWAYS the wise choice.

I am just saying spending good money (or advising others to spend their good money) to replace something that apparently is working properly and is meeting their needs, just because we don't like the company name on the side, is not a wise use of money, or good advice.

I will also add this, then move on. For anyone who does their own tech support on their computers - or is the designated computer guy/gal for their family, friends or neighbors, then by all means, have a known good, spare PSU handy for testing purposes. And for the OP, if he fits that description and has no spare PSU, then perhaps buying a new, quality PSU from a reputable brand to put in this computer is a good idea, then he can use this TT as a suitable spare/test unit.
 
@Bill_Bright
my problem is with you stating that if there is no "defect" in any form to a psu, it will work fine,
which i have seen more than once not to be the case.

fact is, it took me swapping 3 different units before having one that didnt get my gpu to trigger low power (boot/use),
and having changed only the psu, one time even reusing (matching) cable, so i can exclude anything else.

those units are now working with similar or even "bigger" hw (gpu), showing they have no "defect" in any form,
yet wouldnt work (as expected) on my rig.

transient loads (gpu) can be an issue, not just me and not rare, and has NOTHING to do with a defective unit,
or that it is working out of spec, or overloaded etc.
so stating the psu cant be the cause, because its working, is not always correct.
especially if not swapped for different one/or physical check with a psu tester, to confirm 100%.
 
My Seasonic Prime GX-750 has a real hard time with my 3070Ti and 5900X with PBO. It trips so hard that I have to flip the switch. My cheapo EVGA can do it though.. mildly disappointed. But it is running my 3070Ti and 58X3D just fine.
 
Status
Not open for further replies.
Back
Top