Strewth! 45 posts here!

Your brick psu is separate, right? How did you get that? Your box is a laptop without a battery? Is there a little 'DIY' going on here.

Is the noise coming from the power supply brick, or the box? All the big inductors, and all the coils and capacitors should be in the brick. If you're getting it from the box, it could only be fans or disks.

OK, next report. I found an AntiX live CD and tried to reboot into that. I got stuck at Freeze Point no.1, which is what usually happens when I reboot out of X. So I power-cycled, which so far has been a universal fix for the problem.

AntiX booted (painfully slowly!) and I then rebooted from the AntiX desktop menu. The default desktop is either fluxbox or icewm, I can't remember which, but the menu system is the same for both. I removed the CD when prompted and pressed Enter, and the system rebooted smoothly into Slackware.

So, obobskivich, your instincts seem to be sound. I think there's something about my Slackware kernel that sets the video into a mode which later causes problems during program handovers. An electric pulse seems to reset it.

It's a riddle with no obvious solution. A lot of people like riddles.
No. That's how I bought it. There's no battery except for the coin battery that runs the rtc. It is in effect a laptop that thinks it's a tower! Look at the link in Beachboy's last post for more info.
It comes immediately from the case, which must be resonating to something inside it. I also think it's the disk. But increasingly I think that the noise and the reboot problem are not related after all.

Well, at least we can isolate to that (and hopefully that means your hardware is undamaged!) - unfortunately you're probably better qualified to diagnose it from there than I am. My super-duper lazy solution would just be to re-install the whole thing clean, but depending on how much customization/extra software/etc is involved that could take a really long time and there is no gurantee it 'fixes' the problem because we (it seems) don't know what actually is wrong, other than it is somewhere in the Slackware install.

On the coil whine/buzzing/noise: it may be the result of the CPU going in/out of power saving modes (or some other voltage/current change) - I know on high end graphics cards (where this can be 'louder') usually the noise is heard when they switch in/out of their higher performance modes, because Vcore changes. I agree it is probably a red herring here - just a quirk with that machine's physical build, and probably has gone unnoticed until you start scrutinizing the machine due to the above error(s). If you want to better isolate to the noise, based on the pics in the linked review of this tower, it has a 'normal' motherboard, you could try running with the side panel open and get 'closer' to the board to hear where on the board the noise is coming from (assuming it isn't just the disk). If it is the coils near the CPU there isn't a lot you can do 'on the board' but I know the SPCR-types were putting adhesive foam on the insides of their towers to combat noise like this, and I believe Silverstone (the hardware company) sells pre-cut sheets for this purpose (and I'm sure automotive dynamat has also been used), if the noise really bothers you. It shouldn't do much to cooling if you're just covering 'flat metal' and not blocking any fans or putting the foam directly on the hardware (in other words, you're just applying it to the inside of the case side panels, floor, etc to reduce echoes and increase dampening).

Bah something else I thought of: the noise *may* be a fan failing, either a bad bearing or a cracked rotor, and it is audible at certain RPMs (and I assume the fan on this system, like many modern computers, changes RPM with load). That would be worth checking out to rule out a secondary problem - if the fan is failing, you want to replace it before it stops working.

1 members found this post helpful.

Yep I'm one of them for sure.
Is there any settings in the bios that have to do with graphic cards?

1 members found this post helpful.

Or it could be a loose nut or screw that needs to be tightened?
Just a thought.

1 members found this post helpful.

Oh lord - reminds me of a tower I had some years ago that had a persistent 'rattle' that I could never sort (and after a while just gave up on it - I never had 'problems' with the machine itself). Finally, after 8-9 years it was time to retire it, and while gutting parts out of it before sending it off, I discovered the source of the 'rattle' - a screw (one of those ones with the captive washers) had fallen down 'inside' between the motherboard panel and the case, and was just down there, making noise whenever the machine was moved, for years...

1 members found this post helpful.

Bios keeping time. Maybe a cmos battery problem.

Just to give you a idea hazel. I pulled out a old netbook. Batteries deader than a door nail. Plugged in power brick and hit the power button.

Try to get into bios. Netbook acts slow and sluggish. So I wait a bit longer on charger.

Fire it up again.

Hardrive is not seen. Made me take the netbook apart to check. It's there.

Charge it up some more.

Now hard drive is seen.

Here I thought the power brick would be enough to get things working. I found incorrect cmos settings < dead cmos battery > and just being being in hurry did not let the hardware catch up.

Good Luck.

I thought for a while that there was a consistent pattern of failure and was studying it by rebooting each time in a slightly different way. But that doesn't seem to be the case after all. There is however an explanation that is consistent with everything that has happened so far.

A couple of days ago, I tried leaving X running rather than shutting it down, but doing the actual reboot from a virtual console. If I do it from X, it always seems to stop on what I call Freeze Point 1, just before the bootloader takes over. If I close down X first, it stops at Freeze Point 2, when the bootloader hands over to the kernel. Well, that time it didn't freeze at all; it went straight into a full reboot. So I thought "Good, that way seems to work better, but let's check if it's reproducible." So I carefully did the same thing again last night and guess what? It stopped immediately after closing down the kernel without even loading the uefi, but the power light stayed on and, when I put my ear to the case, I could hear that the fan was still going. So there's a Freeze Point 0 as well!

And what normally happens at Freeze Point 0? A handover of the cpu and the video from one program to another, in this case from the kernel to the uefi. So that is the common feature. At Freeze Point 0, no disk access of any kind takes place because, as I understand the thing, the uefi program is read out of a rom chip. And at Freeze Point 2, the reading of kernel and initrd off the disk has already been reported as successful, so there can't be any further disk access there either. At Freeze point 1, there could in theory be a bad reading of the elilo bootloader off the disk but more likely it's the actual handover of control that fails, just like in the other two cases. Which of the three handovers actually fails in a particular reboot seems to vary and sometimes none of them does. In every case so far where there was a problem, power-cycling has fixed it.

I don't think there is anything more that can be done. It's not a severe problem in practice because I very seldom do a reboot (except now when I am doing it regularly as an experiment!) and I have never yet had a failed boot when starting from cold.