M.2 SSD boot problem

[fabmoraes]

Hi everyone.

I installed Windows 10 in a M.2 driver.

The installation went well, the system works normally but every time I restart for any reason, the M.2 SSD stops to be recognized by the bios.

The only way I found to go around of this was powering of the computer. When it is turned on again it goes back to normal.

This is not a big problem, but it is kinda annoying.

 

I am using the motherboard Asus X99 A-II and tryed different configurations in the CSM menu. Legacy only, UEFI drivers first.

I am also using the latest version of the Bios.

 

Does any one went throught this issue?

 

Thanks

[Master Disaster]

X99 used to have issue with booting from M.2 drives but I thought they'd been long fixed by now.

 

Unfortunately, due to the symptoms you described it sounds like your board doesn't like booting from M.2 devices.

 

Googling around throws up a few possible solutions, head into UEFI then set the following

 

Under Boot:> - SATA Support make sure this is “All Devices” or Windows will not restart correctly if you have any other drive connected.

 

And

 

Change SATA mode to IDE in PCH Storage Configuration under Advanced.

 

Try these out and see if they help.

[Naeaes]

It could be your PSU especially if it's an old unit. Older PSUs tend to powercycle slower but the UEFI needs the flash chips powered the instant it's looking for them, not a few hundred milliseconds later. The UEFI might have a setting to introduce a delay into drive detection. It's worth taking a look.

[fabmoraes]

7 minutes ago, Naeaes said:

It could be your PSU especially if it's an old unit. Older PSUs tend to powercycle slower but the UEFI needs the flash chips powered the instant it's looking for them, not a few hundred milliseconds later. The UEFI might have a setting to introduce a delay into drive detection. It's worth taking a look.

Oh, that's totally make sense, since recently I had to switch my PSU to an older one because the newer has to go to RMA. 

But if it is the case od the PSU causing this issue doens't mean it is a defect, right?

[Gonio]

Have you tried updating the bios? Maybe they fixed support that way.

[Naeaes]

22 minutes ago, fabmoraes said:

Oh, that's totally make sense, since recently I had to switch my PSU to an older one because the newer has to go to RMA. 

But if it is the case od the PSU causing this issue doens't mean it is a defect, right?

Not really, no. More like an incompatibility. I used to work as an RMA handler and came across these situations from time to time. It was a bigger issue when SSDs were new like five or so years ago. The PCI-E-based OCZ revodrives and whatnot were especially problematic. Our rule of thumb was that if the PSU and SSD were sold together, it would be replaced/rebated (since the seller should have known better) but they almost never were so we'd just explain the situation and offer a partial rebate at most.

 

If you're interested, this video demonstrates a tester that tells the power-OK delay (the PG value Linus fails to explain at 2:41) A premium modern PSU would have something like 200-240ms. That 340ms was very common at the time, usually fine for using SSDs. 400ms and up was pretty certainly not going to work perfectly. If you happen to get one of these somewhere, toggle the PSU on and off and see if the figure is consistent. It can go up and down randomly so a few dozen tests is required. This is all my experience and gut-feeling, no science.

 

For now, I'd keep messing with the BIOS but wait to get the new PSU back before throwing money at the problem. Maybe it'll fix this too.

 

(Edit: beside the point but in the video Linus doesn't have the PCI-E or EPS cables connected to the device, resulting in no detection on the +12V2 rail. The tester doesn't care about the number of rails, nor could it, just connected wires.)

[fabmoraes]

17 minutes ago, Gonio said:

Have you tried updating the bios? Maybe they fixed support that way.

Yes, thats the first thing I tryed. I also recently check in the website but there is no newer Bios version

[fabmoraes]

2 minutes ago, Naeaes said:

Not really, no. More like an incompatibility. I used to work as an RMA handler and came across these situations from time to time. It was a bigger issue when SSDs were new like five or so years ago. The PCI-E-based OCZ revodrives and whatnot were especially problematic. Our rule of thumb was that if the PSU and SSD were sold together, it would be replaced/rebated (since the seller should have known better) but they almost never were so we'd just explain the situation and offer a partial rebate at most.

 

If you're interested, this video demonstrates a tester that tells the power-OK delay (the PG value Linus fails to explain at 2:41) A premium modern PSU would have something like 200-240ms. That 340ms was very common at the time, usually fine for using SSDs. 400ms and up was pretty certainly not going to work perfectly. If you happen to get one of these somewhere, toggle the PSU on and off and see if the figure is consistent. It can go up and down randomly so a few dozen tests is required. This is all my experience and gut-feeling, no science.

 

For now, I'd keep messing with the BIOS but wait to get the new PSU back before throwing money at the problem. Maybe it'll fix this too.

I totally agree. I won't spend a penny more in this system for now. The system is running perfectly after changed to the older PSU. Only this small issue came with.

But regarding the Bios options, this delay you say is the "post delay"? Because it is the only option I found. I raised this delay and kept with the same issue.

The problem is not when I boot the computer for the first time, it is when I restart it. 

[Naeaes]

3 minutes ago, fabmoraes said:

I totally agree. I won't spend a penny more in this system for now. The system is running perfectly after changed to the older PSU. Only this small issue came with.

But regarding the Bios options, this delay you say is the "post delay"? Because it is the only option I found. I raised this delay and kept with the same issue.

The problem is not when I boot the computer for the first time, it is when I restart it. 

The POST delay is just how long it displays the POST information (or the foll-screen logo) on-screen. Set that to 0 for faster bootup times. The setting I was talking about was a dedicated drive detection delay but it's entirely possible it's not there. Also it might not affect M.2 at all. It's really for mechanical drives that take several seconds to spin up.

 

I just thought of a test you could do. Try to induce the issue by rebooting the computer but hit CTRL+ALT+DEL before you get back to Windows. That should do a software reboot hopefully without toggling the powersupply on and off. On the second go it might detect the drive just fine.

[fabmoraes]

1 hour ago, Naeaes said:

The POST delay is just how long it displays the POST information (or the foll-screen logo) on-screen. Set that to 0 for faster bootup times. The setting I was talking about was a dedicated drive detection delay but it's entirely possible it's not there. Also it might not affect M.2 at all. It's really for mechanical drives that take several seconds to spin up.

 

I just thought of a test you could do. Try to induce the issue by rebooting the computer but hit CTRL+ALT+DEL before you get back to Windows. That should do a software reboot hopefully without toggling the powersupply on and off. On the second go it might detect the drive just fine.

Hi, thats exactly the kind of reboot that induces this problem.

If windows restart, when the system goes back it doesn't recognize the M.2 drive. So I have to turn it off to work again.