PSU died, new PSU, bios wont recognize SATA devices, now wont turn on

[nstiver]

Synopsis: PSU died, replaced PSU with different model, system boots to bios successfully but does not recognize any devices in SATA ports, nor do those devices appear to be getting power. Now can't even get to BIOS, entire PC flicks on and off repeatedly as soon as I turn the PSU on regardless of whether I hit the power button on the front of the computer.

System Specifications

Original PSU that died: SeaSonic M12II 750 SS-750AM 750w PSU
CPU: Intel Core i5-3570K Ivy Bridge 3.4GHz
GPUs: ASUS 670 non-top version, SLI (2x)
MOBO: ASRock Z77 extreme 6
RAM: Mushkin Enhanced Blackline 8GB (2 x 4GB) SDRAM DDR3 2133
Optical drive: Sony AD-7280S-0B 24x SATA Internal DVD+/-RW Drive
SSD: SanDisk Extreme 240GB SATA 6.0 Gb-s 2.5-Inch (OS is installed here)
Storage drive: Western Digital Caviar Black 1 TB SATA III (This was not connected when power supply died)

Detailed explanation:About a week ago my PSU (Seasonic 750w) died randomly while the computer was left running in sleep . I RMA'd the dead seasonic. In the mean time I happened to come accross an extremely good deal on a new PSU, so I bought it (Thermaltake smart m850w PSU) to have a backup and to get my machine working again.

I hooked up the new PSU to GPUs, CPU, fans, ATX power connector, and to all of the SATA devices, (the SSD, the HDD, and the Optical drive). I made sure to connect the data cables to the SATA devices in the same configuration they were in when working properly. The computer boots successfully to the ASROCK BIOS, all fans spinning. However, I get the error code "A6" on "dr debug" (the led display on the mobo). The manual says this means "SCSI detect". I have no idea what that means.

I was able to load the UEFI, and bios recognized all components except for any of the SATA devices--(SSD, the HDD, and optical drive not recognized). The RAM, GPUs, processor, and fans were detected. At this point I was stuck in BIOS because I couldn't boot off anything.

I did not have the HDD connected when the PSU failed, but now the BIOS won't read that either. I don't think any of the SATA devices are actually receiving power. Again, the UEFI showed that there were no SATA devices connected to any port, when in reality they clearly were connected both by power and data cable.

Then things get really strange. I unhooked all of the SATA devices except the HDD (the one that wasn't hooked up when the original PSU died). I plug the AC into the PSU, then turn on PSU (not the power switch for the computer, the switch on the PSU). The computer starts flicking on and off in this weird rhythmic pulsating pattern "click...click...click...click..." with about a 1/2 second between each click. Each time the power flicks on, the LEDs come on and the fans spin briefly. The strangest part is that I don't need to hit the power button on the front of the case for this to happen, it just starts as soon as the PSU is turned on. In fact, I can't get the computer to turn on at all--the switch on the front of the computer appears to be dead.


Measures I have taken in an attempt to fix the problem:

1. Unplug the power cables from everything and re-seated them.
2. Reseated the data cables to the optical drive and the SSD on the drive end, not on the MOBO side as they are covered by my GPUs and I would have to take them out in order to do that.
3. Tried hooking up a SATA power cable from my old modular PSU and using that in the new PSU in an attempt to power the optical drive (didn't work, I know, bad idea--I learned you aren't supposed to do this shortly after I tried it)
4. Setting all preferences to default in the UEFI.
5. Disconnecting the power cables from one or both of the GPUs.

I think that this is either a problem with (1) the motherboard, (2) the new PSU, (3) all 3 SATA devices (unlikely). The fact that I can't even get into Bios anymore tells me that this may be a MOBO or PSU issue.


My main question is how to procede from here? I suppose I can wait for my Seasonic to come back from RMA--perhaps that will clear up the issue. I just can't shake the feeling that this is a MOBO issue though. I am worried that if I keep screwing around with it I may end up frying my CPU and GPUs. I can eat the cost of a MOBO or PSU if necessary, but I can't afford to buy a new processor and two new 670s.

Thanks for the help, and if you need more info or detail I will hapily provide it!

 

[OvenMaster]

I had a similar problem when I had an Antec PSU die on me. After I replaced it, my Biostar mobo's BIOS would no longer retain settings, and both my Pioneer DVD burners were completely dead. Not recognized in the BIOS, they would not acknowledge any discs inserted, nothing worked. I also started getting BSODs pointing to a corrupt BIOS. Replacing the burners didn't help. Putting the Pioneer burners in a new machine didn't work.

It turned out that the dying Antec PSU had wrecked the Biostar's BIOS chip and both Pioneer DVD burners. Replacing the BIOS chip was more expensive than buying a new motherboard, and reflashing the chip wasn't possible because the flash utility told me I already had the latest BIOS version. I ended up trashing the board and buying an ASRock board and Samsung burners.

If you get a new motherboard (or try another known-good working PC), and you connect your present SATA devices and they don't work, then the dying Seasonic may have burnt them out, too.

 

[nstiver]

Well, I am strangely able to get into BIOS again. I tried unplugging all SATA devices and connected a HDD that I had as a spare. BIOS was able to read the spare HDD! Perhaps that means that the MOBO is not in fact faulty, rather the dying PSU killed the optical drive and the PSU -- all the SATA devices that were connected at the time of the original fail.

I'm going to test the SSD at a friends house tomorrow. If it is in fact dead I will try to RMA it--it's still under warranty. I just hope dandisk doesn't reject it--that was a $200 drive.

My concern is that the system will continue to eat hardware. I simply cannot afford to keep throwing money at the machine only to have it kill the hardware. The GPUs and CPU in particular are very expensive and I would not be able to replace them for quite a while.

What should I do from here?

 

[essenbe]

I would try all of your hard drives in a friend's machine. If you find them to be dead, I would contact Seasonic technical support and insist that they replace the bad hard drives. Now, it probably won't happen, but it has happened before. It does not cost anything to try.

 

[bobafetthotmail]

Another AsRock mobo surviving a PSU meltdown without damage? Seems to look like that, and wouldn't be the first time. Most of their stuff that eventually got in my hands survived lots of abuse.

Check the capacitors on the mobo, if you see some that are swollen, leak or look damaged (some mobos have capacitors with glue or weird resins over them to stabilize them in place, so even if they look messy it could be normal if it's glue/resin), then something bad has happened and the mobo is unreliable (i.e. change it).

If no capacitor is damaged and you get to see the BIOS, it is safe to assume the mobo, CPU, and RAM work fine. If GPU are damaged you'll see it when trying to load an OS.

See if you can boot a linux live CD like say ubuntu, placing it on a USB stick if your DVD burner is dead, or for the sake of having something handy while also testing the hardware, make a Win 7 installation pendrive, which will be handy in case you have to reinstall everything.
If that works, then the mobo/CPU/RAM/GPU are fine.

As others said, check the other devices on another computer, although I'd bet that they are dead.

My concern is that the system will continue to eat hardware.

This is usually caused by a crappy PSU overloading things, Thermaltake is a good brand afaik, i never heard of SeaSonic.

I had very very very few processors and no RAM/GPU that died when the mobo died. Usually the mobo death happens before the other components sustain damage and if the mobo is fine the components are fine as well. ASrock is a good brand and never failed on me without very good reasons.

The computer starts flicking on and off in this weird rhythmic pulsating pattern "click...click...click...click..." with about a 1/2 second between each click.

If this sound comes from HDDs... well it's called click of death for a reason.

 

[nstiver]

Thanks for the reply bobafetthotmail,

Seasonic is one of the most highly regarded and quality PSU makers on the market, so it was certianly not a crappy PSU--maybe it was just a lemon. I didn't skimp on any componants, everything in the system was nice. Maybe 750w wasn't enough to power 2 670s, a i5-3570, peripherals, and 2 SATA devices?

Also, the "clicking" sound was the sound of little jolts of power going to all the case fans and mobo -- as soon as I would turn the PSU on, the system would get little "clicks" of power, the fans would spin like a quarter rotation each time, hence the "click...click....click" sound. The HDD wasn't hooked up when the clicking happened, so that wasn't it. In fact, the HDD is the only undamaged sata device I have, and not supprisingly the only one that wasn't hooked up when the PSU blew. The only SATA devices that were hooked up when the clicking happened were the damaged SSD and optical drive.

I HOPE that the SSD and optical drive are the only things damaged by the PSU failure. I really, really, really hope that the 670s are not damaged...I couldn't stand to lose my babies. They are also to costly for me to replace right now.

I will test the SSD and optical drive today. You would recomend making a windows 7 pendrive and trying to boot from that? I think the USB ports are just fine.

 

[bobafetthotmail]

Frankly, I cannot tell offhand, but 750w for a dual GPU setup doesn't sound that well.
There are calculators that help you figure out your build's power requirements. This one is what I use myself and most here recommend to use before purchasing a PSU. You fill your specs there and it tells the wattage you need to run it (even taking into account PSU aging). If you pay a little cash it gives much more electrical-oriented data but it's usually unnecessary.

Still, if you were overloading the PSU it would refuse to boot or do GPU-related issues (as the GPUs are in your case by far the most hungry component, thus the first to get shafted if there isn't enough power), not go kamikaze on your hardware. Definetly a lemon.

as soon as I would turn the PSU on, the system would get little "clicks" of power, the fans would spin like a quarter rotation each time, hence the "click...click....click" sound.

This usually happens when the PSU isn't enough to run everything (also when it's malfunctioning and it's producing less power than what it should). 850w is quite a bit but I don't know your power consumption, see the calculator above.

But you say you managed to run the rig afterwards, so... I'm a bit puzzled. If it was the behavior with the old PSU then it's ok as it was faulty. But if it happened with the new one, that would raise some eyebrows if the calculator says 850w are enough for your rig. Maybe something on the mobo said "screw that, I'm gonna leave a GPU offline", some mobos are smart enough to shut down non-critical stuff when power isn't enough.

You would recomend making a windows 7 pendrive and trying to boot from that? I think the USB ports are just fine.

Yes, but I had bigger schemes.
I was thinking to use that pendrive to install Win 7 on the spare HDD you have just to test everything, and a USB pendrive is by far the fastest way to do so (as even average USB speed is ridiculously higher than anything read from a DVD). If you manage to run a OS (any os, either windows 7 or the ubuntu I linked above, YUMI helps you placing it on a USB stick if you want to boot from that) on that machine without issues, then you can assume all the hardware the OS recognizes and can work with is fine. (NVIDIA does have good linux support so ubuntu should have/find the right drivers off the net easily, in case you don't want to use the spare HDD)

 

[Zepher]

It's possible that that new psu doesn't work properly with the Asus board.
Over the years I have experienced 2 situations where a perfectly fine psu would not work with a motherboard.
One was an Antec and an intel board and just the other day with a 600watt Thermaltake and an Asus board with a 3570k
Computer would power up for less than a second and then turn off.
It's rare since I have built over 100 machines in the past 16 years and this problem only happened on 2 machines.

 

[nstiver]

@ Zypher-- the mobo is ASROCK, not ASUS. The fact that I can get into bios now makes me think that compatibility with PSU and mobo isnt the issue, but who knows!

@ Bobafet -- the PSU calculator thingie says I need only 517w. I had a feeling 750W was enough because those 670s actually draw much less power than previous generations of NVIDIA cards. I also had several people who I trust tell me that Seasonic 750w would be more than enough.

Also, there is the fact that the "pulsing" thing would start as soon as I turned on the PSU, not when I tried to pwr on the computer itself from the power switch at the front of the machine. In fact, when the clicking was goign on, I couldn't actually get the machine to turn on--it was stuck in a perpetual clicking loop. Wouldn't the pusling start when I tried to turn on the computer if it was a low pwoer issue instead of when the PSU was connected and switched on?

Anyway, I THINK that the reason the clicking stopped and I was able to get back into BIOS was that I disconnected both the "dead" sata devices--the SSD and the DVD/RW--and reconnected the spare, functional HDD. Could the fact that the two dead devices were connected to the PSU and mobo cause the clicking/power issue? Of course, I also reseated some cables and moved cords around in there, so it could have been something else.

 

[essenbe]

Yes, dead hard drives can stop the whole machine, just like it is shorted out. I know from experience.

 

[bobafetthotmail]

the mobo is ASROCK, not ASUS

Asrock is a subsidiary of ASUS.(technically it's more complex than that as there is another company in the middle and they do keep the two companies separate, but in the end it's ASUS the owner)

ASUS boards and Asrock's do share some similarities and issues, as far as I remember.

Yes, dead hard drives can stop the whole machine, just like it is shorted out. I know from experience.

This.

Another thing could be that you screwed up the front panel connections, so the machine was receiving a constant "reboot" order like you were keeping pressed the hard reboot button.

 

[nstiver]

@ bobafet -- That is interesting, I didn't know ASROCK was a subsidiary of Asus. I was under the impression that Asrock had a higher reputation with MOBOs than asus.

Anyway, the plan now is to borrow an optical drive from my friend and try to install windows onto my HDD from there. The HDD has some random files and other junk on it, but I suppose I should be able to install windows.

Then I can see if my GPUs are acting up--if I can actually get into windows I can test them with games/benchmarks.

In the back of my mind though I am still wondering if the MOBO didn't escape the catastrophe 100% unscathed.

Assuming I am able to install windows, how do you all think I should procede with testing my hardware to make sure all is good?

 

[nstiver]

Update -- Well, I borrowed my friend's optical drive and have successfully installed windows on my spare HDD onto the computer. In fact, I am typing from the computer in question as we speak.

There seem to be no problems with any components, however, I have not tested my GPU's under gaming conditions yet.

I do, however, have a new hypothesis, one that reflects more poorly on me.
After the initial PSU died, I bought the replacement PSU. I, being stupid, overeager, and ignorant, decided to simply use the cables from my Seasonic PSU to power my hardware and just plug them into the modular Thermaltake--save some time, right? Of course, nothing powered up, and I flicked the PSU on and off a few times, giving anything that might be connected a nice, big, jolt of electricity.

What if this is how my SSD and my optical drive got killed? It makes sense--everyone I talk to says that a nice PSU is unlikely to wreck other hardware when it dies. Not impossible, but certianly unlikely. Maybe everything would have been fine had I just used the new cables that came with the PSU. However, my GPUs seem to be fine--maybe they were somehow not damaged by using the wrong cables.

Do you guys think this is a plausible explanation for what happened? I am not knowledgable about electricity and power in general. Do you think that this may be the case?

 

[bobafetthotmail]

I was under the impression that Asrock had a higher reputation with MOBOs than asus.

Yes, it can be described as "ASUS with less press coverage and more quality". It started as ASUS's cheapo mobo subsidiary, and when even its cheapo stuff was selling because it was significantly more reliable than other cheapo stuff like foxconn or random rebranded boards, they moved to do fully-fledged boards like yours as well.

Do you think that this may be the case?

Most computer components are designed in a very very dumb-proof way, the only way for that to happen is if a cable/connector is damaged. I really doubt it is the case, but look for burned cables or bent pins. But really, it's something very obvious that you should have noticed if your eyes were functioning correctly.
If connector pins are bent enough to make it fail you would need a hammer to make it fit, and this should clearly indicate that there is something wrong. (actually, they won't fit well before they actually short-circuit, had to fix quite a few pins)
The cables are usually designed in a way that even if they fail and you manage to use them anyway the current does not flow in the failed parts, not that you shovel 500 watts through the wrong pin. So at most you have stuff that remains unpowered, not that blow up.

Really, something went boom in the PSU. It happens. Even with good ones. Actually the only things that do fail pretty often (from a tech support point of view) are spinning HDDs and PSUs.

Testing a mobo is a very straightforward (and time consuming) affair. Run all things (USB, etherent, integrated graphics all audio ports) and do stress tests with "sacrifical" hardware (that is, components that you don't care a lot of in case the whole thing melts down). But more often than not, any damage is very apparent. All mobos that survived a PSU failure that I had in my hands didn't have any issue (like say damaged USB controllers or whatever, everything ran fine). Those that die refuse to boot or do very weird things right from the start (like what the guy above said, and also simply shutting down randomly is another all-time favorite).

The current goes from PSU to mobo and then to GPU/CPU/RAM, so if the mobo is fine, nothing on the mobo received damage.

If you notice, stuff that failed was all connected directly to PSU. If your cards weren't connected directly to PSU (some do have such additional power connectors), they should be fine.

 

[nstiver]

Hmm, both of my cards WERE connected to the PSU by 8 pin connectors. The thing is that I played an hour of Starcraft 2 with SLI enabled and they honestly seem fine. I'm still waiting for BF3 to download so I can put them under some real stress, but temps, usage, voltage, fan speed, all seem great on my graphics cards--just as they were before the accident.

Maybe the voltage spike was enough for the MOBO and GPUs to handle but too much for the SATA drives.

There is no obvious damage to my mobo. No blown capacitors or burn marks on ANY of my hardware, including the devices that got fried or the cables that fried them. I may have gotten off with just a couple of dead SATA devices

Well, now begins the process of RMA for the SSD. Should I just say "it stopped working"? It's not technically a lie, right?

 

[bobafetthotmail]

Good to hear.

stay evasive on what happened to the SSD and it should end well.
They rarely take the time to actually figure out what is wrong with stuff you RMA as it's cheaper to send a replacement (hours of a technician to determine who's the fault and lots of times it's hard if impossible to tell for sure). They won't miss 200$, as the SSD failure rate is ridicolously low.

If you feel funny you can try to RMA the DVD drive as well.

Btw, if you want to stress test the CPU, we have a tutorial (read the disclaimer therein).

In case you want to really stress test the GPUs, FurMark can help you do so (forcing them to render tons of 3D fur and thus use 100% of their power, something I doubt BF3 will do). Be warned though, if your GPU has any defects, it will fail and force a computer shudown, and possibly suffer permanent damage (they included pics of an overclocked card that FurMark killed). The flip side is that if it does not fail then you are 100% sure it's fine, and that the mobo is fine as well.

I tend to go full-nazi on used hardware I want to resell, and it must endure a full night of CPU and GPU stress test, but you may not need to be so harsh.

 

[nstiver]

I've been gaming on the GPUs and I'm convinced they are fine.

Another thing I noticed just now -- I used to have horrible coil whine coming from my case but I assumed it was just the GPU. Now that the old PSU is gone however, no coil whine at all--I guess it was coming from the PSU all along. I wonder if that was a sign that it was dying, it was getting progressively worse and worse.

Yeah I'm going to RMA that SanDisk and the dvd. In retrospect I'm kind of bummed I didn't get the Samsung 830 series 256 gig SSD instead of my SanDisk extreme 240 gig. The Samsung is cheaper and bigger, and I think just as fast.

I am hoping that they won't take the time to test the SSD. If they screw me with the RMA though I'm dropping SanDisk immediately and going over to Samsung.

Boy was I spoiled by those fast load times from the SSD. It seems to take forever for anything to load using this old fashioned HDD now.

 

[Layback Bear]

I have been watching this thread for a few days. I think you found the problem in the power supply harness. To be sure of things I would double check the routing of the new harness and make sure they are not in a bind or rubbing on anything. Sometimes when a harness is in a bind it will put undo stress on a connection.

 

[bobafetthotmail]

hm, usually coil whine happens when a PSU rail in a multi-rail is at near max capacity, if it was single-rail PSU (what does single-rail mean?), the whole PSU was at near-max capacity. Which is a clear indication that it had issues as the nominal power is pretty above your needs. What Ladyback Bear says is also possible, cables doing very tight turns could increase a lot the resistence on that connection, requiring more power than usual.

Keep fingers crossed, because if your SSD is no more in stock (if it's older than 2 years it can be the case) they will offer a newer model (usually better) for replacement.

Pestering SeaSonic and ask for a refund or something because they killed some of your hardware is another thing that should have been tried.