Has my Corsair AX760 gone to meet God--err, Mr. Gerow? Or something else have a meal of dust, or high-BMI female human serenaded my PC some other way?

[PianoPlayer88Key]

Okay .... Sooo.... I came to my desktop PC yesterday evening...to find it powered off.  Hit the power button on the case, nothing happened.  Hit the power button on the B550 Taichi motherboard, nothing.

RGB lights on the motherboard were still on, doing their default pattern.

Powered the PSU off and back on (which turned off the RGB lights).  Tried power buttons again.  Nothing.  (This time the RGB lights didn't light up.)

 

Unplugged everything from the Corsair AX760 PSU except the A/C power and 24-pin, and jumped it (pins 4 & 5) per the instructions on Corsair's site.

Flipped the PSU power switch to on, didn't see the fan move, flipped it to off after a couple seconds, and right that moment the fan started to spin a bit, then after several seconds eventually stopped.

Tried several times again, nothing.

 

 I've had that Corsair AX760 since January 2015, so I'm suspecting it may have died but I'd be a bit surprised since it wasn't run within an inch of its life and the warranty only just expired last month.

 

 Also that was the only component I could reasonably test while isolating others, I don't have a spare AM4 motherboard or CPU laying around.  (I do still have an LGA 1150 board & CPU and DDR3 RAM, although I wouldn't be able to do much with it cause the Hyper 212 Evo's fan is kinda seizing up.

 

Anything else I should try, or is it likely the PSU kicked the bucket and I need to buy a new one?  (I have the $ as long as they're not being scalped, just wasn't planning on getting a new one yet.)

 

 

I recorded a video a little bit ago of what the fan does when I try the paper clip test:  https://photos.app.goo.gl/3wmE3rsdHLThD6xMA

 

The fan control was set to normal, not hybrid.

 

 

I hope I put the paper clip in the right place, I'm pretty sure that's where https://help.corsair.com/hc/en-us/articles/360025085372-How-to-Test-a-power-supply-unit said to do it:

 

 

 

PCPartpicker list (to see what parts I have and what PSUs might be compatible):  https://pcpartpicker.com/list/9s6gLs

 

 

I also asked in a couple discords ...
Someone in one discord suggested a Corsair VS 600W, but I don't think that's appropriate with my build.
(He also suggested an EVGA T6 1000W paired with an RTX 3080 Ti but I'm not planning a GPU upgrade yet, especially not one like that.)

 

 

I'll also need to figure out what other upgrades I might eventually do.  (As for the GPU, i have a psychological cap in that I don't want something with a TDP higher than about 200-250 watts, when I eventually upgrade my 1060 3GB.)

 

I was thinking somewhere around the $100-150 range or so for a PSU, I'm guessing, that may be subject to change pending a bit more research.  (At least I have my laptop to use in the meantime.)
 

I'm generally looking at least at 80+ Gold with 12 or more SATA cables, but might consider 80+ Bronze with at least 8 SATA.

If I could get something that's cable compatible with my AX760 as much as possible, that would be nice.  I especially don't want to have to re-route the 8-pin EPS/CPU cable.

 

 

 

A few Corsair PSUs I'm considering include the:

 

HX750 (Platinum) (or maybe 850 but that's pushing the budget)

RM (2021, Gold) 750W or 850W, maybe 650W, or RMx (2021, Gold) 850W
 

 

Looking at the cable compatibility guide at https://www.corsair.com/us/en/psu-cable-compatibility, it looks like I could maybe re-use the EPS/CPU, PCIe and SATA/PATA cables from the AX760, I'd just have to use the 24-pin ATX connector from the new PSU.


I'd also consider some other PSU models/brands, preferably Tier A although I might consider Tier B.  (With a 5950X though I should probably limit to A or better, you think?)  I'd most likely have to replace all the cables though.  I believe the AX760 was based on a SeaSonic platform, but I doubt there's a current non-Corsair PSU that's at least compatible with the 8-pin cable.

 

 

 

OR .... is it possible something else might have died?  I hope it wasn't the motherboard or CPU.  The fact that nothing powers on when I hit the power switch makes me suspect the PSU, although it surprises me that it would have died this soon....  (Also the motherboard's RGB lights had still been on when I first came to the PC, and the fan does spin a little bit when doing the paper clip test if I wait a couple hours or so between tests...)

 

 

[Somerandomtechyboi]

I doubt the psu is dead when the lights still power on, problably just mobo refusing to turn on cause it thinks theres a short if it isnt the psu

 

Put the psu in another system to really test if its the psu, and heat soak takes awhile to overheat a cpu with a heatsink esp on idle so you are fine with running it as long as you arent stress testing the cpu or something

[Guest]

5 hours ago, Somerandomtechyboi said:

I doubt the psu is dead when the lights still power on, problably just mobo refusing to turn on cause it thinks theres a short if it isnt the psu

The motherboard RGB runs on +5VSB.  So as long as that rail is working, those will light up.

 

5 hours ago, PianoPlayer88Key said:

If I could get something that's cable compatible with my AX760 as much as possible, that would be nice.  I especially don't want to have to re-route the 8-pin EPS/CPU cable.

Good news.  Only the 24-pin cable is an odd ball in that cable set.  The rest are type 3/type 4.

 

[PianoPlayer88Key]

8 hours ago, Somerandomtechyboi said:

I doubt the psu is dead when the lights still power on, problably just mobo refusing to turn on cause it thinks theres a short if it isnt the psu

 

Put the psu in another system to really test if its the psu, and heat soak takes awhile to overheat a cpu with a heatsink esp on idle so you are fine with running it as long as you arent stress testing the cpu or something

 

2 hours ago, jonnyGURU said:

The motherboard RGB runs on +5VSB.  So as long as that rail is working, those will light up.

 

Good news.  Only the 24-pin cable is an odd ball in that cable set.  The rest are type 3/type 4.

 

 

Okay now something's interesting...

Last night as I said, the RGB lights were still on when I came to the PC, but it wouldn't turn on.
Switching the PSU off then back on, the RGB lights wouldn't come up.

 

Several minutes ago, though ...
I plugged the CPU and PCIe cables back in, hit the power, nothing.
Held the Reset CMOS button for a few seconds, then realized I forgot to plug the 24-pin ATX back into the motherboard.
Plugged it in, which lit up the RGB a second later (forgot to switch the PSU off again).
Hit the power, and the system booted up to Windows.  (I didn't have SATA plugged back in yet, but Windows is on an NVMe SSD on the motherboard.  Also I noticed the PSU fan behaved the same way as in the video I posted earlier, so apparently that's normal operation?)

 

Temporarily powered it off so I could put the case back where it normally sits, now going to see if it powers back on and go into BIOS.  (I'm gonna guess it didn't actually reset CMOS cause I did that while the ATX connector was unplugged.)

 

Powered up again, but hitting Del to get to BIOS settings turns up a blank screen.
Rebooting, and letting it go normally, gets me to Windows.

Shift+restart .... -> restart to UEFI .. THAT gets me to UEFI settings.

Settings look like they're default ... BUT, now I'm not so sure they weren't reset before.  My RAM was running at 2400 before the system went down (it's 3600 but for some reason doesn't like running there so I back off  bit, like 3200 to 3500 or something when I have XMP enabled), and I usually enable SATA Hot Plug and turn off the motherboard RGB, but SHP was disabled and the RGB was lit up while the system was running previously.

Hmm....

So I may not need to get a new PSU yet.... but what could have caused the system to behave like that, any ideas?

[Guest]

1 hour ago, PianoPlayer88Key said:

I didn't have SATA plugged back in yet

But it's plugged in now and it's working?

 

1 hour ago, PianoPlayer88Key said:

 Also I noticed the PSU fan behaved the same way as in the video I posted earlier, so apparently that's normal operation?)

No.  If you have Hybrid mode off, the fan should always spin.

 

[PianoPlayer88Key]

50 minutes ago, jonnyGURU said:

But it's plugged in now and it's working?

Yeah it seems to be working now, idk what happened earlier.

 

50 minutes ago, jonnyGURU said:

No.  If you have Hybrid mode off, the fan should always spin.

The fan was spinning with hybrid mode off, just really slowly, like about 10-15 rpm... but then I was thinking, maybe the LED light I was using had a bit of a strobe effect.  (I shone the same light on the fans for my AIO and the blades looked like they were spinning slow, but the sticker on the hub was a blur.)
One thing I just thought of (but idk if/when I'll try it) was to stick something soft and non-conductive in there and "listen" to the fan blades grazing it.  That would probably tell me the actual fan speed if my light + camera isn't exactly truthful.

[PianoPlayer88Key]

Well ... back to square one.

 

Was away for several hours, came back, PC was off, RGB lights still on.
No response from the power button.
Unplugged / powered off the PSU (which extinguished the lights a second or two later) and held a Clear CMOS button for a few seconds.
Plugged back in ... power button (both on the case and mobo) .. nothing.

 

I haven't turned the case on its side this time (yet) to see if the PSU fan spins (for the paper clip test).

This time, I'm gonna try something else that's similar to what happened previously ...
Leave it powered off and unplugged for maybe a couple hours or so (or longer, idk, last time it was overnight and partly into the next day),
Then see what happens, see if it powers back on.

I'm still not sure if it's the PSU or the motherboard or something else.   It's quite weird that it'll start working, then shut off while I'm away.

Would be nice if I had a way to diagnose it (like if I'm able to get it back up and running again), to see what might be causing it to shut down and not power back up.
 

A thought briefly came to mind about possible borked sleep settings in Windows, but I would have thought you could still power on the system after that.  Maybe messed up sleep settings in BIOS?  (I don't remember changing anything like that though...)

I really hope it's not something like a motherboard short (although if it was I would have expected the issue to crop up right away, not a couple months after building the system).  I would absolutely hate having to disassemble the entire system (like removing the motherboard) -- I'd rather replace a dead-of-old-age (death during warranty = infant mortality) top tier 10 or 12 year warranty fully-modular / cable-compatible PSU a few times before I went through the steps to unplug and remove a motherboard.  (And I don't have all that much plugged in either, just 8 hard drives, 2 M.2 SSDs, an AIO, a GPU, a few fans, front panel connectors, etc.)

I really don't want to have to replace the motherboard already, as what I would really like (DDR5, PCIe 5.0 or 6.0, supports a full range of CPUs from entry consumer to flagship server, long enough CPU/RAM/etc upgrade support to last through 3 or 4 PSU replacements, 7+ PCIe x16 slots, 16 DIMM slots supporting RDIMMs & LRDIMMs, dual CPU sockets, >12+ SATA/SAS ports, several NVMe slots, 10G LAN, enough PCIe lanes so every port/slot can be used simultaneously without bandwidth sharing, etc) doesn't even exist yet, and if it did exist now, would cost several times my current entire-system budget JUST for the motherboard.  (I don't like to replace a motherboard unless I'm doing a wholesale major upgrade on everything.)

 

 

 

 

 

Any ideas on what to try next?
If I'm able to get it running again, any diagnostics I should run, or maybe have some way of monitoring and periodically recording certain stats onto an external SSD (I have a 500GB Samsung T5 sitting nearby), then if/when it goes down again plug that SSD into my laptop and see what the most recent status was?
If I'm not able to get it back up.... test the PSU again?  Maybe there's a way to temporarily power on my ASRock Z97 Extreme6 mobo without the CPU connector (enough to see LED indicator lights or spin a fan; also see a previous post for an idea of how hard it would be to unplug the EPS12V connector from the B550 Taichi)?

 

 

 

EDIT (at 3:06am, this reply originally posted at 12:45am)...

Plugged the PSU back in & turned it on, the RGB LEDs lit up.
Hit the power button and the system powered on.



I still don't know why it keeps going down.  Gonna check the Windows sleep settings... (I thought I had it disabled except for the monitor....)

 

 

[Guest]

So the PC powers up once it "cools down".  Is it safe to say that?

 

[PianoPlayer88Key]

2 hours ago, jonnyGURU said:

So the PC powers up once it "cools down".  Is it safe to say that?

 

Well, I'm not so sure about the "once it 'cools down'" part, but it does power up once it's sat and rested for a while.

 

 

Came back to it this morning, and it was still running.  (The case was pulled out from the desk though.)

Joined my church's Zoom meeting, that was going alright so far.

Opened HWInfo64 to check temps and other stats.
Did a quick Cinebench R23 run, got a score of about 26.3K or so I think.  (Not sure if that's typical or low or high for a 5950X though.)

Started a custom 3-minute Cinebench loop just to see what happens, while monitoring a few things in HWInfo64.
CPU temps didn't exceed 60°C (I think 59 point something was max),

but core voltage looked a bit high, I saw 1.48V, is that normal for a 5950X?  (It would have been way too high for my 4790K and 6700K.)
(The 5950X is on an ASRock B550 Taichi, under an Arctic Liquid Freezer II 360, paired with 128GB DDR4-3600 RAM that was running at 2400 cause XMP was off.)

That confuses me... high voltage combined with low temps ... is something not being reported correctly, or is the LFII360 that good at keeping temperatures in check?  (With the 212 Evo (even when the fan was working) on the 4790K + Z97 Extreme6, it would hit 100°C at like 1.3V under heavy load (like Prime95) at stock, I think, and would still be in the 80s, maybe low 90s in Cinebench R23.

 

 

Then, about 2 minutes into the Cinebench loop (with about a minute left), the PC just spontaneously clicked off.

 

 

So this time I actually caught it in the act ... I don't think it's overheating cause it didn't exceed 60°C, but I am a bit concerned about the high CPU voltage.  (I haven't reloaded my custom UEFI settings, just running at stock, and even the custom ones didn't have any OC applied except for RAM, I just changed a few other things.)

This time I'm going to see if it'll let me power it on after a few hours leaving it plugged in.  Yesterday when I got it back up and running walking, I'd had it unplugged, and I also reset the CMOS (this time I'll try not doing that and see what happens.)


I wish there was a way to see if a protection on the PSU is tripping, or a rail is going out of spec or something, but I guess there isn't cause I have the AX not the AXi....

 

[Guest]

@PianoPlayer88Key

 

And when it shuts down like this, you can't immediately turn it back on?

 

If not, does cycling the PSU switch also not allow it to turn back on?

[PianoPlayer88Key]

56 minutes ago, jonnyGURU said:

@PianoPlayer88Key

 

And when it shuts down like this, you can't immediately turn it back on?

 

If not, does cycling the PSU switch also not allow it to turn back on?

No, couldn't immediately turn it back on,

And I cycled the PSU switch (although this time it took about 5 seconds before the RGB lights went off), and it doesn't come back up (at least not yet).

(It's only been probably less than an hour or so since it went down again...)

Gonna switch the PSU off again and leave it off for a while (but still plugged in) and try again a bit later.  If that doesn't work, I might try the unplug trick again (and wait a couple hours or so).

 

(I wish I'd thought to take a screenshot or 2 of hwinfo64 before it went down, or maybe even a short video clip with my cell phone camera.)

 

 

1:27PM (PST) edit:  switched PSU back on, RGB lights came on.  Hit power switch, system seems to be coming back up...

Still want to figure out what's taking it down though...

[Guest]

@PianoPlayer88Key

 

I'm willing to bet something in the PSU is overheating.

 

[PianoPlayer88Key]

2 hours ago, jonnyGURU said:

@PianoPlayer88Key

 

I'm willing to bet something in the PSU is overheating.

 

I wonder, maybe so ... but any idea how I might find out what?

 

Thinking about things, I think it could be more likely the PSU than the mobo or CPU, cause .... (LOL there it went right now (2:52PM PST) as I was typing this on my laptop, just had Chrome open, no stress test.) ... if it was the mobo (or related) I would think it would power right back up, or at least try then shut down after a second or two.  But, it won't come back up at all until I unplug the PSU and let it sit for a while ...

hey there's another thing I could check too...  (Just need to find my dad's outlet tester, idk where he keeps it) ... this house was built in 1962, and while the outlet I had the PC plugged in previously was properly grounded, some outlets in the house have an "open ground".  I recently moved the PC across the room, and suspect it may be now plugged into one of those outlets...

(While testing, unplugged the PC from the strip to plug the tester into that spot, then decided to move the tester to a different spot where it's easier to see, and when I plugged the PC back in at 3:06PM PST the RGB lights came back on.  Haven't hit the power switch yet though...until 3:19PM and it's coming back up.)

 

Yeah ... here's where it was before...

And this is where it's plugged in now.


I wonder if that could be part of it...  idk where else I'd plug it in though, moving the PC back where it was isn't an option now, and to get to any other outlet I'd have to use an extension cord, and it could be a trip or cat chew / litter box spill hazard.

 

I don't have it in my bedroom, cause there's no physical place to put it right now, and the outlets in there are all:

I'm curious about something now .... I might just leave it on the Windows login screen (not logged in) for a while and see if it goes down again sometime today.... (or maybe it needs to be doing something to trip up whatever's causing it to shut down...)

Well I didn't have to wait long ... 3:22PM ... *click* (sound of fans spinning down)


Okay now that's interesting... I went away for a while, came back (at 4:21PM), and the system is on, at the Windows login screen.  I don't remember touching any power buttons...
Windows is not set to automatically restart after a BSOD, and it's not set to lock the screen after a period of inactivity.  (I can't remember if there's a UEFI setting that enables automatic power on (for example when power is restored after an A/C power loss but that's not what happened here), although I also don't think I have wake on LAN / USB enabled either...

I'm remembering now, there were a couple other times in the past couple months that I'd been away from the PC for a while, and came back to find it on the Windows login screen instead of the desktop with my apps/etc open.  (And there was no BSOD / stop code error logs then either.)

I almost wonder if it could be related... hmm... maybe it's not the outlet after all?  (Cause it may have shut down & come back up while I was away earlier, while plugged into the properly grounded outlet.)  I'm still not sure if it would be the PSU, or maybe the motherboard / CPU or something....
I really hope it's not something like a short between the mobo & case, although I'm pretty sure I installed all the standoffs correctly and all that.  (I would have thought that would cause a lot more obvious problems though, not be intermittent like this is.)


I had actually bought another PSU (Seasonic Core-GM 500) intending to use it to power up my Z97 / LGA1150 platform, to possibly turn it into a NAS or something, but I was unable to find a compact case that would support a full ATX motherboard, Hyper 212 Evo, ATX PSU and at least 10 3.5" HDDs (with them mounted like the picture in the spoiler),.

Spoiler

and the holiday return policy for the PSU was nearing its limit in January, so I sent it back al huevo nuevo, y no tuve problemas con ellos.

 

So while I briefly had a 2nd PSU, I don't anymore.  I had another one a while earlier - an OCZ StealthXStream 500W bought in 2008 for my AM2 build then, but that system passed away in 2012 - I thought it was the motherboard then (it would power up, but no video out, I was using onboard graphics and didn't have a dGPU and I think someone told me the southbridge or something may have died so even with a dGPU I still wouldn't have gotten it running) but more recently I've wondered if it could have been that PSU.  I donated the OCZ PSU (and some other stuff I no longer had use for, including 3x each fully working 4TB and 5TB HGST Deskstar NAS HDDs, and didn't want to go through the hassle of trying to sell them) to a local non-profit recycler a couple years ago.

 

Fry's is gone, I'm not so sure if Best Buy would be the "best" option, and pretty much all the local computer stores have closed down, so if I was going to take the PC somewhere and try swapping parts, idk where I'd take it.  Micro Center in Tustin, CA, is an hour and a half drive each way, not counting traffic.  (At least I'm south of L.A., actually east of San Diego, but there's still some traffic usually.) 

 

Also if I don't have to spend any $, other than replacing parts that are out of warranty (like the PSU - bought the AX760 in January 2015 and its 7-year warranty expired January 2022 I think), that would be preferred.

 

 

So are we still maybe suspecting the PSU?  Or could it be something else, and maybe we should consider having a moderator move this topic to, maybe, the troubleshooting section?

[PianoPlayer88Key]

Okay that's interesting...this time it was my LAPTOP that spontaneously shut down (while I was away for a few minutes) and isn't powering back up for the moment.  (Desktop is still up from the earlier restart for now...)

 

It's plugged into the same strip with the desktop now, and has been plugged into the same spot for a few years but I think that's the first time in a long time (or ever, I have seen it BSOD then click power off while counting up the percentage but I wasn't there watching it this time) it's done that.

 

I'm wondering about the ungrounded outlet the strip's plugged into, maybe that's also contributing to the issues?  (I doubt it's too much load, I'd think everything on that outlet / break would trip off simultaneously if that was the case.)

 

 

Maybe this is why the laptop shut down...

It's still fairly warm to the touch underneath several minutes later.  (I JUST cleaned it though maybe a couple months ago or so...)

 

[Guest]

@PianoPlayer88Key

 

Well... being ungrounded isn't going to hurt unless there's a surge that has nowhere to go (typically: ground), but it could be something else to do with the mains power.

 

Here in CA we have horrible power.  Resistive and inductive loads are not impacted, but capacitive loads are.  We don't tend to get brown outs, and weather rarely effects me because my lines are underground, but for some reason the voltage will spike from 115V to a few hundred voltages for about a millisecond.  We actually had our heater's control switch die because of this phenomenon.  And LED light bulbs last me much less time than classic incandescents, which is ironic since incandescent light bulbs are pretty much "banned" in CA.

 

Of course, still could be OTP on the PSU since the PSU has to "cool down".  Just like any CPU or GPU, your SR MOSFETs use TIM and it can dry up and if they overheat, they're going to tell the PSU to shut down.

[PianoPlayer88Key]

Ahh @jonnyGURU ... yeah I still would be a bit more comfortable if things were properly grounded though.

I'm in southern CA (east of San Diego), main issue I've had with power is it being so expensive (like, i'm guessing it would cost less for people in cheaper areas like TX, WA, etc to mine actual bitcoin on, say, a GeForce 6800 XT, than it would cost me to mine etherium or dogecoin or other current ones on an RX 6800 XT ), and parts of the electrical system in our house not being up to code (and probably still has the original wiring from 1962 for the most part, idk how much longer that will last before it becomes a hazard).
Our lines are above ground strung between poles out on the street.  We don't get a ton of rain or storms that directly hit the lines I don't think, but once in a while a tree might come down and cut power that way.

OTP ... I wonder if there's a way I could monitor the temperature while it's running, and see what happens when it shuts down.  I believe the AX series is rated at 50°C ambient, although with it being 7+ years old I wouldn't be surprised if it eventually starts derating.  (And it's nowhere near 50°C in the room right now, maybe 26-28°C or so.) But maybe something else is getting a bit toasty...  (I'm not qualified / trained to open it up, of course, so I won't be attempting that.)

Anyway I cleaned out the laptop, then powered it on (had to plug it in as it didn't respond otherwise at first)... and apparently the power connector had gotten dislodged, making it run on battery.  What I don't know though was why didn't it give me any low battery warning at all, it just clicked off...
When I got it back up and running, I checked the battery status in HWInfo64, and it said the charge level was about 5.4%, and wear level was 10%.

Jump ahead to now, and HWInfo64 says the battery is fully charged, but with a wear level of about 30%.  (I've had the issue before with a previous battery, that once the wear level started climbing, the system would just shut off without warning.  Eventually it got to where the battery might last 2 or 3 minutes unplugged, then the laptop would die if I dislodged the plug a little, that's when I replaced the battery.)

 

 

 

Back to the desktop ... it had been back up and running since Sunday afternoon / evening, but just maybe 25 or so minutes ago while I was listening to some audio files in VLC to try to find something (filenames were pretty generic, from some things I had recorded in the past; also had other stuff open in the background, but nothing like stress tests), and it clicked off again.

If it's the PSU (which would be probably easiest to replace if I get one that's at least compatible with the EPS12V connectors) that's shutting down the system, then why is the 5VSB still up (lighting the motherboard's RGB)?  I guess that's a low enough draw, or different rail or something, that it's "immune" to the other issue, or requires something more severe to kill it?

Plugged it back in just now, it lit up and I powered it up, we'll see how long it lasts.  (One thing I'm tempted to do is underclock / undervolt the CPU a bit, but I'm not sure how best to tackle that with a 5950X and B550 Taichi, it's definitely different than my previous desktop build's 4790K & Z97 Extreme6, or my laptop's i7-6700K and P750DM-G...

[Guest]

@PianoPlayer88Key

 

@50°C ambient is assuming the condition of the components within the PSU are running at spec.  @50°C ambient, you have components running well over 100°C.  If some of those have dried up TIM, it doesn't matter if it's 50°C or 30°C... it's going to overheat. 

 

 

[PianoPlayer88Key]

10 hours ago, jonnyGURU said:

@PianoPlayer88Key

 

@50°C ambient is assuming the condition of the components within the PSU are running at spec.  @50°C ambient, you have components running well over 100°C.  If some of those have dried up TIM, it doesn't matter if it's 50°C or 30°C... it's going to overheat. 

 

 

Ah, and since it's out of warranty it's probably more likely.   (BTW does the 50°C ambient rating mean the temperature in the overall room, or the temperature inside the case or the air temperature in the PSU before cooling?  And just curious, how do people properly choose PSUs for operating in environments that can exceed 50°C, like some places in the southwest USA desert can get to in summer, and even like 37-40°C at night?)

Also I came back to it just now (6:41PM) and it was still on.  Then, all I did was hit the alt key (to alt-tab to another window, or see which virtual desktop I was on cause upon restart and restoration of my Chrome windows there's a bunch scattered across several desktops), and it clicked off again.

Now that's interesting... 7:01PM and it powered itself on again, I didn't hit the power button (and hadn't unplugged it).  It's possible I might have reached for (or bumped) the wrong keyboard or mouse (I have 2 keyboards & mice here, one for the laptop, one for the desktop, I'm looking up replacement PSUs on the laptop right now).
7:50PM - tried to log into Windows (after letting it sit a while, was partly through typing my password I think) ... clicked off.

 

Maybe I should go ahead and order something, I just don't like doing that unless I KNOW that's the issue.  (I don't like the idea of, say, replacing a CPU when it turns out the RAM is kaput, or replacing a motherboard when it's the PSU that used its hoof to send a bucket flying.)

 

I'm thinking, if I get a PSU that uses partially different cables, I should probably somehow write which PSU the cables are for, at least on the AX760 so I don't mix up the ones that shouldn't be interchanged.  (A while ago I put some tape on some SATA cables and wrote in sharpie which port on the motherboard they were plugged into, but the tape eventually loosened and came off, so maybe there's a way to write directly on the cable, although black on black might not show up very well.)  Or maybe there's a better way to keep track...

 

 

I'm looking at the 2021 RM750 (CP-9020234-NA) and HX750 Platinum (CP-9020137-NA).  The RM850 is $139 at Newegg with a promo code (yes I know the recent issues but I personally haven't had problems with them), and the RM850 is $98 at Amazon.

 

 

From what I can tell on the pinout pages on your site, and https://www.corsair.com/us/en/psu-cable-compatibility , the AX platinum uses its own 24-pin connector, but the CPU/PCIe and SATA cables are standard Type 3/4, which are also used by the RM850 and HX750 (I hope I got it right.)  So it looks like I'd just need to use whichever new PSU's 24-pin ATX connector, but could reuse the others (SATA, CPU, PCIe).  (And as a "bonus" I'd have a few extra SATA cables so if I ever needed to hook up a lot of HDDs...but then if I run out of places to plug them in....)

Spoiler

 

 

 

 

Okay I don't have all those drives now (got rid of most of the PATA drives, most of the SATA HDDs ≤5TB and some older SSDs), but I still have:

• 2x 8.4GB IBM DTTA-350840 (one was dead when I got it but I might want to scavenge some parts (like the magnet) and see if I can figure out how it died)
• 80GB WD800BB
• 750GB WD7500AAKS
• 3x 8TB HGST HDN728080ALE604
• 8TB Toshiba N300 HDWG180UZSVA
• 2x 10TB HGST HDN721010ALE604
• 12TB Toshiba MG07ACA12TE
• 2x 14TB Toshiba MG07ACA14TE
• 240GB Crucial CT240BX300SSD1
• 256GB Crucial CT256M550SSD1
• 2x Team Vulcan 500GB (2.5" SATA)
• 2x 1TB WD Blue 3D (2.5" SATA)
• 1.05TB Crucial CT1050MX300SSD1
• 2TB Seagate Barracuda 120 (3.5" SATA)

 

 

 

 

As for non-Corsair PSUs, the Phanteks Revolt Pro 850W (PH-P850GC, $140), BitFenix Whisper M 850W (BP-WG850UMAG-0/6/7/8/9FM, $100) and ADATA/XPG CORE Reactor 850W (COREREACTOR850G-BKCUS, $110) are on my current shortlist.  (There were about 53 other models, including 22 from Corsair like CXF, HX850+ Platinum, other RM & RMx, originally on my list.)

http://jongerow.com/Corsair_pinouts/Corsair_pinout_AX_Platinum.htm says the Phanteks Revolt uses the same 24-pin pinout as the AX Platinum, and http://jongerow.com/Corsair_pinouts/Corsair_pinout_Type_4.htm says the Bitfenix Whisper M and XPG Core Reactor use Type 4 as well (which AX uses for CPU/PCIe and SATA/PATA), but for the Phanteks, does it ONLY use the same 24-pin (but different PCIe/CPU & PATA/SATA)?

 

Most of the PSUs are 80+ Gold, except the HX750 which is Platinum.  I wonder how much difference it would make with our electricity rates -- I think we're on either the TOU-DR1 or TOU-DR-P plan - I know we have peak, off peak and super off peak, but don't know if we have the "reduce your use" event / $1.16/kWh though.  We usually blow way past our baseline every month.

System (with AX760) is still off right now (4:29AM), PSU's been switched off for a few hours but not unplugged.  I'm gonna get some zzz then come back and see if it'll start up again, or if I actually need to unplug the PSU.

[Guest]

3 hours ago, PianoPlayer88Key said:

Ah, and since it's out of warranty it's probably more likely.   (BTW does the 50°C ambient rating mean the temperature in the overall room, or the temperature inside the case or the air temperature in the PSU before cooling?  And just curious, how do people properly choose PSUs for operating in environments that can exceed 50°C, like some places in the southwest USA desert can get to in summer, and even like 37-40°C at night?)

That's 50°C inside the PC.  If it's hotter than 122°F inside their chassis, the PSU  functioning properly is the least of their worries.

 

3 hours ago, PianoPlayer88Key said:

From what I can tell on the pinout pages on your site, and https://www.corsair.com/us/en/psu-cable-compatibility , the AX platinum uses its own 24-pin connector, but the CPU/PCIe and SATA cables are standard Type 3/4, which are also used by the RM850 and HX750 (I hope I got it right.) 

Correct.  The only odd ball cable is the 24-pin.

 

3 hours ago, PianoPlayer88Key said:

but for the Phanteks, does it ONLY use the same 24-pin (but different PCIe/CPU & PATA/SATA)?

ONLY the 24-pin cable is the same on the Phanteks.  All of the others are different pinout.

 

[PianoPlayer88Key]

Ahh okay @jonnyGURU looks like that scratches the Phanteks off the list.

 

 

(Powered up the system yesterday afternoon around 5 or 6pm or so (I forget exactly now), it's still up right now as of 7:48am.)

 

 

Hopefully this will help things a little...

 

(Yes, I'm aware of the recent issues with Newegg, however, I've pretty much always had good experiences with them, and they estimate I should get it tomorrow (Friday the 4th).  Amazon would have been the same price, but estimated arrival would have been Tuesday, 8th.)

(And i hope it won't end up being a  , I don't expect it to. )

 

If I got it right, I think I just need to use the 24-pin ATX cable from the new PSU, but can reuse the rest of the cables (CPU, PCIe, SATA)?

(Also I guess we'll see how hard it is to tell them apart; don't want to, when I need an extra cable in the future, accidentally grab the wrong one.)

 

Maybe there's a way to mitigate this as well... (look closely at the connectors on the ends of the cables near the camera)

(I believe it's courtesy of...)

 

50°C inside the PC ... so I'm guessing if it's 40°C in the room (my cousin's place a few miles from me has reached that at times in summer, her place has no A/C and is poorly insulated, it's an upstairs apartment in an almost 100 year old house) the PC might be a wee bit warmer than 50C.   (If someone needs a high-end system in a hot environment, then I guess they'd either get a higher wattage PSU (so it's not running right at its limit) or figure on replacing it sooner.)

[PianoPlayer88Key]

Well that's interesting.  The PC has been behaving itself so far the last day or two, and I haven't installed the new PSU yet.  (It arrived this afternoon.)

I ran a stress test this morning on the system (Cinebench R23, 30-minute loop) and it passed.  (Previously, it had shut down about a couple minutes into the test.)

One observation, it was a bit warmer earlier in the week and last week, up in the 80s or so (may have flirted with 90 idk), but today it was a lot cooler, high of about 65 or so.  But, we now have A/C in the house, although we don't necessarily keep it at a consistent temperature.  (Sometimes it's closer to 70, sometimes closer to 80.)

I still don't know what's going on.  I still plan to swap the new PSU in.  Intermittent issues are always un dolor en el glúteo mayor to troubleshoot as I'm sure some of you well know.

[PianoPlayer88Key]

Okay so quick update (it's been a bit over a week).

I haven't yet swapped the new PSU in, BUT, the PC has NOT powered itself off since the day I ordered the PSU (March 3).

I still have the Newegg box sitting on a chair next to the tower.  (I've cut the tape but haven't actually opened it yet.)

I'm now debating whether I need to put it in at all, if it hasn't been giving me much problems.  That said, I haven't been actually USING it that much, although it has been on 24/7 with Chrome and HWInfo64 open.  Also I've done a few 30-minute CInebench stress tests which all passed, unlike a test I did a couple weeks ago when the PC shut down about 3 minutes into the test.  I haven't yet pushed the tower back into the spot in the computer desk, it's sitting on the floor right now.  (well, on a plastic piece that my chair is also on, there's carpet under that but the PC isn't directly on it.)  First time it shut off it was in that cubby hole in the desk, but it had also shut off a few times when it was sitting out where it is now.

 

I don't know how much of a factor it is but I'll mention it cause JonnyGuru seemed to think it was something overheating ... we were having a bit warmer weather when it was acting up, with temperatures up in the 80s °F, possibly pushing 90°F.  The last week or so it's been a fair bit cooler, highs in the 60s or so, lows in the 30s and 40s.  Also the house itself has A/C, although we don't usually have it consistently locked at a particular temperature all the time, we let it vary quite a bit.  (In summer it's not uncommon to be in the 90s or 100s here at least outside.)


Am I the only person this kind of thing happens to?  Your computer is having a problem, you narrow it down to the part you think you need to replace, and the moment you go to order it (and for the next week or two after and counting) the PC works fine even without replacing the part?

(I just hope that if I send it back, the PC won't start acting up then.)

(Hey what's the name of the "Law" that essentially says when the technician is around the device works perfectly?  (Maybe this is a variation on that - when it feels threatened that it'll have a part replaced it decides to act normal?)