Boot problems

[Enjay]

Not GZDoom but more in the "Did your computer break?" category.

I have mentioned a few times that I think my computer might be failing. Here are a couple of relevant posts.

The thing is, it's intermittent and usual diagnostics don't really turn up anything useful. Sometimes the way it behaves gives the impression that it might be related to a failing SSD, but at other times that doesn't really seem to fit. I've tried the tips and advice that I've been given here and elsewhere but I still can't definitively say "the problem is X". If it was just the SSD, I could replace it. However, I don't want to go through the hassle of replacing it and setting everything up again only to find that I still have the problem. Not really sure what to try next TBH.

You will want to get that looked at by someone who has the resources to help you out. You won't have to set everything up again if you simply image the drive onto another. But first you have to rule out the SSD being the issue - and to do that you need another one (possibly a loan from someone else?) and copy your full system onto it and boot from that device, instead.

If the boot failures continue, then it is a software problem on Windows. If they do not, then you may want to purchase the drive.

I decided to buy a new SSD (they aren't as expensive as I expected). I cloned the drive (I know someone with a cloning docking station) plugged it in with new SATA cables (as suggested on here) and... still not working. I have also tried a fresh Windows install. Nope, still not working.

Basically the symptoms are:

When I switch on my computer, after a few seconds I get a "beep" and the Republic of Gamers logo screen (which says "Press DEL or F2 to enter your UEFI BIOS"). All that is normal.

If I then just leave the computer, usually what happens - if things are not working, instead of moving on to the startup screen (which should be the RoG logo + Windows spinning balls) the screen goes blank and I can see a text-mode cursor flickering at random locations on the screen. Eventually (maybe 20 seconds or so) I get the normal Windows 10 logo boot screen (instead of the RoG one - indicating it is booting in legacy mode, not UEFI) and the machine starts. From that point on, it seems to work OK, although I have had a couple of very rare Win10 versions of the BSOD (but the last one was a long time ago).

However, here is the interesting part (to me): If, when the computer is at the "Press DEL or F2 to enter..." I do in fact press del, I go into the BIOS, hit F8 to bring up the boot menu, the Windows boot manager on the SSD is listed*. If I click on it, the machine then boots normally in UEFI mode. i.e. it's a pretty reliable way to get the machine to boot correctly.

*occasionally, a very long time ago (more than a year) the SSD would sometimes (rarely) be missing from the list, but reappear on restart.

So, honestly, I'm stumped. The supplier of the computer (PC Specialist) initially identified the SSD as the likely problem. Clearly it isn't, because I now have a new one from a different manufacturer entirely. PC Specialist even took the machine back a little over a year ago, reconfigured it and sent it back. It worked for a few months, then the problem reappeared.

The problem has been rumbling on for a couple of years now, but the failed boots seem to be getting more frequent (basically all the time ATM - but that has happened before, and then rectified itself for a while). The machine is now out of warranty but I am keeping up a dialogue with PC Specialist. The conversations are very stilted though (emails are replied to within 3 working days etc).

I've flashed my BIOS, tried a new CMOS battery, run SFC /scannow, chkdsk, Open Hardware Monitor and other hardware checking programs (even some pretty low-level stuff provided by PC Specialist) but everything always comes up with a clean bill of health.

I'm constantly backing everything up but my fear is that one day I will switch on and nothing will happen.

My gut has been telling me that there is something wrong on the motherboard, but I really don't know.

Any ideas?

[Graf Zahl]

Maybe a problem with the cable connections? This sounds like it sometimes doesn't manage to get the needed info from your boot drive.

[Enjay]

That is what it sounds like right enough. I have already changed the cable to the SSD and I just tried plugging it into a totally different connector on the mother board. I still got the problem when I restarted.

[Rachael]

What happens when you put your SSD on another system? Will it boot?

[Enjay]

Yes, it does. It struggles a bit because my machine is set up to expect user files etc on a second hard drive, but it doesn't have anything that would seem like the symptoms that I get on my own machine.

[wildweasel]

What is the brand on your SSD? Some years ago I was running a Crucial MX100 series, and for whatever reason, that thing would not like staying connected. It'd work fine most of the time, but then briefly blip out - causing Windows to die (since it was installed on there!), making it disappear on reboot, what-have-you. It'd take powering the machine completely off and back on for the motherboard to see it was still there.

Never did figure out why, either; I ended up buying a different drive (a Samsung M.2 one) and haven't had problems with that one.

[Rachael]

Yes, it does. It struggles a bit because my machine is set up to expect user files etc on a second hard drive, but it doesn't have anything that would seem like the symptoms that I get on my own machine.

Your motherboard may be at issue then. I hope the second SSD was a loaner - if not at least you now have another storage device.

I'm really bad with motherboard repair - I usually just swap it out for a new one, myself.

Alternatively, consider that the PSU may be a problem too. If it's not delivering the kind of power that your SSD expects, then that will cause issues, too. And SSD's can be pretty power hungry, at least when they're not idle.

So those are two things I would swap out and check. Hopefully you can find some loaners so you don't have to drop the money until you figure out what the actual problem is.

If you're comfortable with a soldering iron and have a voltmeter, some flux, and a microscope, you might be able to nail down the problem to a motherboard component and get a replacement part that way. I wouldn't recommend that unless you know what you're doing or can follow a youtube video or something - but it's certainly a hell of a lot cheaper than replacing the entire board.

[Enjay]

What is the brand on your SSD?

The one that came with the computer is a Kingston HyperX. I replaced it with a Samsung. Both show the problem.

Your motherboard may be at issue then. I hope the second SSD was a loaner - if not at least you now have another storage device.

I'm really bad with motherboard repair - I usually just swap it out for a new one, myself.

Alternatively, consider that the PSU...

I bought the new SSD but that's OK because I have a case and a USB connector and I now have a nice fast Solid State USB drive that I can take with me.

Ever since this started, I've just had a gut feeling that it's something on the motherboard not doing what it should be. To me it was always about the SSD not being detected properly rather than the SSD itself being faulty (at least I've confirmed that now I guess).

I'm pretty sure that the PSU is OK. There is no evidence of anything else suffering from lack of/intermittent power and the unit is a good one (though, if it's faulty, it's faulty regardless of how good it is meant to be). I've considered replacing the motherboard. I've replaced motherboards before (a long time ago) and it wasn't that much hassle, I'd just rather not do it myself it at all possible. It is an option though.

I'm certainly not up to tracking down faulty components and re-soldering though. I can solder, but bigger stuff (mostly car electronics). Nothing with the level of precision required to work on a motherboard.

[Rachael]

This is why loaner components would be useful. If it boots up in another computer though, then it likely is the motherboard. Not a surefire bet, but it's more likely.

[R4L]

Is this a SATA based SSD? You could have a bad SATA controller. Not common, but I've seen it happen.

[Rachael]

That was kind of implied with the bad motherboard thing. It would specifically be the SATA controller on it, since everything else works.

Although - that does give me an idea that might save Enjay some money: They might still sell SATA addon cards cheaper than they sell motherboards - if they do then you might not have to replace the whole unit.

[Enjay]

I've been tending towards thinking that it is the SATA controller too. I wasn't aware that additional SATA cards were/are a thing. That might well be worth looking in to. As a side point though, I am not aware of any difficulty reading the HDD or bluray drives (both also on SATA connections). But, given that I don't usually boot from them, problems there would not be so immediately apparent.

Something that was just suggested by the PC supplier was to disable Compatibility Support Module boot support in the BIOS boot settings. I had not tried that, but I have now, and had two successful boots (more than I've had first time for the last few weeks). Once I get finished with the stuff that I want to do on my machine tonight, I'll try a few more reboots to see what happens.

However, they didn't really explain why this would help. From what I understand, disabling CSM means that the computer will not even try to boot into legacy mode, so it should be UEFI or nothing. Is that correct?

[Rachael]

However, they didn't really explain why this would help. From what I understand, disabling CSM means that the computer will not even try to boot into legacy mode, so it should be UEFI or nothing. Is that correct?

That would be correct.

In "compatibility" mode the UEFI spawns a fake BIOS that has all the services that a BIOS would provide to a DOS or early Windows system, and these include boot services that are compatible with MBR-formatted partition tables. UEFI's are compatible only with GPT-formatted partition tables, and that's why the compatibility feature is present.

In fact - funny enough - Apple even did this way back in the day with their Bootcamp app. Bootcamp would load a fake BIOS into memory in order to load Windows.

If you're wondering why some computers work with both - it is possible to hybrid-format a drive to be both GPT and MBR compatible, but it doesn't take much to mess these drives up.

[Enjay]

Thanks. So I guess I give it a few tries and it will either work or fail to boot entirely at some point. (I know that, if it does fail, it's not a permanent failure and I can easily get back into the bios.)

[Kappes Buur]

For what it's worth ....

I periodically run Memtest 86+ v4.00 (I don't have UEFI) over night, from a bootable CD, to get a few runs of the test in. That will check how well the data is conveyed from one point to another. Since then the program has grown up quite a bit.
https://www.memtest86.com/download.htm

Not so frequently I run CrystalDiskInfo 8.0.0 to check the health of my disk drives.
https://filehippo.com/download_crystaldiskinfo/
So far, all my drives report as Good.

And last but not least, to test my graphics section I run GPU Caps Viewer to find out what is supported and run the demos which are included to see how well it all fares.
https://www.geeks3d.com/20190322/gpu-ca ... #more-9874