TR Forums

I installed a new PSU (Evga Supernova 1200P2 Platinum) on my old - LGA1366 - desktop. The first PSU always bothered me tho there was nothing wrong with it (Corsair 850x).

Just prior to installing the new PSU I tried to install Corel Painter 2019 - the demo version. The installer hung till I finally reset but the computer wouldn't boot. I couldn't even get to Safe mode even with my installation disk. I thought maybe the PSU had finally crapped out so I put in the new PSU.

Now out of 2 SSDs and 4 HDDs the BIOS only shows 1 HDD, which is not my boot drive. The optical drive shows up too. But none of the other drives.

I checked the connections many times, I swapped power cables - I tried using the "good" drive's power and data cable on my boot SSD - no luck. I used different SATA ports on the PSU ... I tried Fail-Safe and Default in the BIOS. I also tried changing to IDE mode but that didn't work either. Last I removed all RAM except for 1 stick, and power and data from all drives except the boot SSD - nothing...

It feels like it's got to be something with the new PSU and I'm guessing the Painter hang is a separate problem and incidental - but I'm not sure. I'm at a loss how to proceed.

Well thanks for reading - if you have any ideas let me know - thanks!

Try connecting the drives to a different system and see if they are recognized.

If you don't already have one of these (or something similar) in your toolbox you may want to order one. Useful for testing and pulling data off drives.

thanks for the link - i'll have it on Tuesday - nice tool, easier than the enclosures I've got stuffed in my closet...

hard to believe 4 drives suddenly went bad, though I suppose I must eliminate that possibility by testing them.

canoli wrote:

thanks for the link - i'll have it on Tuesday - nice tool, easier than the enclosures I've got stuffed in my closet...

hard to believe 4 drives suddenly went bad, though I suppose I must eliminate that possibility by testing them.

If either PSU was providing grossly out-of-spec voltages, that could've damaged multiple drives.

I would expect system instability and a myriad of other issues at the motherboard-level if your PSU voltages were off by enough to fry multiple drives.

I did notice your new PSU is modular. Are these drives all on the same cable, perhaps? Have you tried multiple cable ports on the PSU? I suppose we'll find out more when you get the external dock.

Both of those PSUs are modular. You didn't happen to use one of the Corsair strings on the EVGA, did you? That could instantly fry every drive attached to that string.

bfg-9000 wrote:

Both of those PSUs are modular. You didn't happen to use one of the Corsair strings on the EVGA, did you? That could instantly fry every drive attached to that string.

oh no... Yes at first I did. I never imagined a different pin out between cables (tho it's obvious now). I only thought if they're both capacitor-free they'd be okay .

well I guess I'll find out tomorrow once I try them in another system....

Btw if I did indeed fry the electronics on these disks I assume the data is still recoverable provided I can find a new circuit board for them. I guess not for the SSDs but the magnetic drives....yes? (Please give a glimmer of hope lol)

canoli wrote:

Btw if I did indeed fry the electronics on these disks I assume the data is still recoverable provided I can find a new circuit board for them. I guess not for the SSDs but the magnetic drives....yes? (Please give a glimmer of hope lol)

Not nowadays, sorry.

If you have valuable data on there which you haven't been backing up, you should go to a properly professional data recovery firm e.g Kroll Ontrack not Bob's Local IT Shack. It will be expensive, though.

canoli wrote:

bfg-9000 wrote:

Both of those PSUs are modular. You didn't happen to use one of the Corsair strings on the EVGA, did you? That could instantly fry every drive attached to that string.

oh no... Yes at first I did. I never imagined a different pin out between cables (tho it's obvious now). I only thought if they're both capacitor-free they'd be okay .

well I guess I'll find out tomorrow once I try them in another system....

Btw if I did indeed fry the electronics on these disks I assume the data is still recoverable provided I can find a new circuit board for them. I guess not for the SSDs but the magnetic drives....yes? (Please give a glimmer of hope lol)

You used to be able to buy a drive with the same controller board and swap them... haven't needed to try this in quite a few years.

I don't know why I shouldn't be able to do that now...although finding a working Samsung 1 TB 103SJ isn't going to be easy. I saw 1 at Amazon for $175!

canoli wrote:

I don't know why I shouldn't be able to do that now...although finding a working Samsung 1 TB 103SJ isn't going to be easy. I saw 1 at Amazon for $175!

The reason that this isn't straightforward anymore is the firmware on the controller board. The current "practical" solution is to identify the chip(s) on the board that contains firmware, and swap them between the good and the bad board. Success is not guaranteed by any means, either.

Swapping the boards without doing this, depending on the application, can permanently destroy data.

For decades when flash was expensive, the standard practice was to put the firmware (the drive's OS) as well as the locations of any relocated sectors onto a hidden service track on the spinning disk media itself. The controller board would only contain the BIOS/CFE needed to boot the firmware.

This could be a bad thing--there's a performance penalty from using up some of the advertised cache just to store the firmware after boot (which was done because it's too slow to run it directly from disk), and of course if the media in the service track itself got even slightly corrupted then you have a shiny new brick. But it did let you swap controller boards because all the important stuff was on the spinning part.

Nowadays, flash is cheap enough to dispense with this.


I'm not surprised that PSU manufacturers choose to repurpose cheap commodity PCIe power connectors for their modular cables, or that there is no standard way for wiring the PSU end, but the current fashion of coloring all the wires black so you can't even see if the color-coding is correct is annoying.

Wow it gets deep (technical) pretty quick, but I think I understand - thanks you guys for the explanation. Damn, well no biggie, everything is backed up, and my SSDs were the OS on one and a cache drive was the other. And who knows maybe they'll come to life....I'll find out later today or tomorrow.
Thanks again!

roncat wrote:

canoli wrote:

Btw if I did indeed fry the electronics on these disks I assume the data is still recoverable provided I can find a new circuit board for them. I guess not for the SSDs but the magnetic drives....yes? (Please give a glimmer of hope lol)

You used to be able to buy a drive with the same controller board and swap them... haven't needed to try this in quite a few years.

Yes, it used to be possible but this was maybe 15-20 years ago. With any reasonable modern drive this is no longer possible. The tolerances are so small that each drive needs individual calibration data to work and this is stored on the circuit board. Using the mechanics with a calibration from a different drive is very unlikely to work.

Depending on where the wrong voltages went, it is also possible that the head actuator has been physically damaged.

roncat wrote:

You used to be able to buy a drive with the same controller board and swap them... haven't needed to try this in quite a few years.

The only way that's worked for me is when the drives are identical - down to revision, lot & date codes. Even same drives with different revs has failed.

Well it looks like I killed 2 magnetic drives and 2 SSDs. Luckily for me the one drive I hadn't backed up was spared, a WD 5TB drive loaded with pictures.

Why the %#@* the manufacturers can't settle on standard pinouts, at least for SATA power is beyond me.

one question: With my old Corsair HX850 I added a Y-splitter to the EPS 8-pin cable and an extension as well (my board has 2 EPS connections). EVGA supplied 2 EPS cables but they're too short to run them the way I want. I don't dare use that old extension/Y-splitter before I know for sure whether it's safe. Anyway I'd prefer to buy new extensions as I only have 1 now. I see them for sale all over the web...is there any danger they'll fry my CPU?

Thanks you guys - happy Fourth of July to you!

canoli wrote:

Well it looks like I killed 2 magnetic drives and 2 SSDs. Luckily for me the one drive I hadn't backed up was spared, a WD 5TB drive loaded with pictures.

My condolences. That really sucks.

canoli wrote:

Why the %#@* the manufacturers can't settle on standard pinouts, at least for SATA power is beyond me.

Heck, back in the day, Dell even used a non-standard ATX motherboard power connector pinout, with potentially destructive consequences!

An old article from those days which describes the issue: http://www.informit.com/articles/article.aspx?p=339053

Obligatory XKCD link: https://xkcd.com/927/

canoli wrote:

one question: With my old Corsair HX850 I added a Y-splitter to the EPS 8-pin cable and an extension as well (my board has 2 EPS connections). EVGA supplied 2 EPS cables but they're too short to run them the way I want. I don't dare use that old extension/Y-splitter before I know for sure whether it's safe. Anyway I'd prefer to buy new extensions as I only have 1 now. I see them for sale all over the web...is there any danger they'll fry my CPU?

Extensions and splitters (* see below) should generally be pretty safe, since they're intended to pass the existing pinout through unchanged. Mis-wired ones do happen occasionally though, especially if you're ordering stuff from random third-party Amazon sellers or eBay. It is best to at least do a visual inspection to verify that things appear to be wired "straight through", or (if you're more paranoid) invest in a cheap DVM to allow you to ohm out all of the connections to check for correct wiring.

* There's another potential issue with splitters though: If the power draw from the device(s) the splitter is feeding exceeds the rating of the contacts in the upstream end of the splitter, the connector can overheat. Overheating in turn damages the contact surfaces in the connector, resulting in further resistive heating (i.e. a feedback loop), which can eventually result in the destruction of the connector itself, or even (if you're extremely unlucky...) a fire. Splitters should only be used if you know the current carrying capacity of the upstream (pre-split) connector won't be exceeded. For lower-power devices like HDDs and SSDs, I generally don't worry about using power splitters. But for higher power devices like a motherboard or GPU, splitters can be a problem.

Thanks JBI - interesting article - a little sad it was written almost 20 years ago and the situation is hardly better - what he describes still applies. I thought only Apple practiced such proprietary nonsense - they certainly take it to the extreme - but I never guessed it's the standard business model across the PC parts industry.

re: the EPS cables - when I put this system together (2010) I don't remember considering any potential problems extending the Corsair EPS cable - I think I just bought the cheapest name-brand I found on newegg. I do remember being cautious about splitting the power to the board for the reason you noted but can't recall how I satisfied myself it was safe to use.

When you said "ohm out all of the connections to check for correct wiring" - if I make sure Terminal A (for example) on the PSU side has continuity through the cable and its extension to Terminal A on the board side...will that do the trick?

Thanks again.

canoli wrote:

Thanks JBI - interesting article - a little sad it was written almost 20 years ago and the situation is hardly better - what he describes still applies. I thought only Apple practiced such proprietary nonsense - they certainly take it to the extreme - but I never guessed it's the standard business model across the PC parts industry.

At least Apple generally designs their stuff so the connectors aren't physically compatible, so you need to get a little creative to blow stuff up by connecting an Apple thing to a non-Apple thing. With the Dell PSU issue, it was the SAME damn connector, just wired completely differently.

canoli wrote:

When you said "ohm out all of the connections to check for correct wiring" - if I make sure Terminal A (for example) on the PSU side has continuity through the cable and its extension to Terminal A on the board side...will that do the trick?

Yeah, that's basically what I was getting at. If you're extra paranoid, you can also check for shorts between the individual pins, though that would typically be pretty obvious from even a cursory visual inspection.

To be fair, the ATX spec does not mention anything about modular cables at all, and in fact using them makes it harder to meet the required electrical and ventilation specifications. Modular cables were something invented by the PSU manufacturers themselves in order to cater to the tidy-case crowd, and have become a fashionable way to distinguish their higher-end models.

The standards bodies are private entities (like Underwriters Laboratories) which can do nothing but license their name and logo for a price. I can guarantee those nonstandard Dell PSUs and motherboards were not labelled "ATX" anywhere on them. Not meeting the standard does not necessarily mean they aren't interoperable or safe though--perhaps the manufacturer simply did not want to pay for an "80-plus" efficiency rating or to use the "WiFi" logo. It's not just PC parts either--oil companies must pay if they want to sell something branded as "Top Tier" gasoline, which can be arguably worse because the specification for that requires alcohol in it! You are perfectly free to start your own certified PSU-end modular cable standard, but good luck trying to get people to pay you to use it.

What really gives me the willies are those adapters that repurpose common, logo-ed cables for other purposes--like those remote PCIe x1 extenders that use USB cables, or video extenders that use ethernet cables. Imagine someone plugging a PCIe port directly into USB power.

I think I'm going to try swapping boards (and BIOS chips) on these Sammy drives. I found the boards online, about 40 bucks each and there's a shop close to work that can swap the BIOS for me. This site seems to think it's at least possible to get the drive spinning again, long enough to grab some files I don't have backed up.
https://www.hddzone.com/bf41_00329a_samsung_pcb_repair.html

If it fails it's okay I have most of the data on BD. I'm not worried about killing the drive further since it's already toast. I'm curious to see if it'll work. I'll let you know...

canoli wrote:

I think I'm going to try swapping boards (and BIOS chips) on these Sammy drives. I found the boards online, about 40 bucks each and there's a shop close to work that can swap the BIOS for me. This site seems to think it's at least possible to get the drive spinning again, long enough to grab some files I don't have backed up.
https://www.hddzone.com/bf41_00329a_samsung_pcb_repair.html

If it fails it's okay I have most of the data on BD. I'm not worried about killing the drive further since it's already toast. I'm curious to see if it'll work. I'll let you know...

Yeah, you've basically got nothing to lose at this point beyond the cost of the boards and BIOS work. Before you pull the trigger, consider that there's a pretty good chance the firmware chips on the existing boards are fried. So don't get your hopes up.

Hi...I trust you didn't disengage the SATA information links from the hard drives or then again on the off chance that you disengaged them you re associated the information links to the right drives.
In the event that you have an extra PC you could eliminate the drives each in turn to the extra PC and run circle diagnostics
for example, the Western Digital Data LifeGuard Diagnostics - WinDlg (DLGDIAG for Windows)

Dont attempt to boot up the issue PC until you know whether the issue is a broken hard plate or legitimate debasement in the RAID format