Windows 98 cold start problem
February 2, 2016 12:59 PM   Subscribe

I need help with a start-up issue with a Windows 98 Pentium III (yeah, yeah, I know).

I have an old Pentium III which I picked up a few months back for retro-computing fun that I recently installed Windows 98 second edition on. The first time I power it up on a given day, it loads windows, finishes playing the startup sound, then immediately reboots and goes through the whole scandisk routine. After that it runs and restarts with no issues for the rest of the day.

I have been pulling my hair out trying to figure out what is going on with the thing. I've gone through the registry, checked for hardware conflicts, checked drivers, installed a fresh CMOS battery, that sort of thing. Anyone have any clue as what is going on? Google has failed me, and my 1990's era windowz skilz are seriously rusty. I shut off scandisk to speed up the cold start routine, but I would much prefer fixing the problem. It is at the point where I am considering wiping the drive and re-installing everything. This is not my preferred solution as it took quite a while to get it set up, and finding and installing all of the correct drivers was a lot of work. I don't want to go through all that again only to have the same problem re-occur.

The computer is a Shuttle Spacewalker with a Twister FV25 motherboard, Pentium III, 500 mb ram.

Just to head off some helpful suggestions, I know about and use dosbox and virtual machines to run old games on newer hardware. Running it on old hardware is just more fun. When it works, of course.
posted by fimbulvetr to Computers & Internet (13 answers total)
 
Does it seem like Windows is rebooting due to a fault, or does it seem like the system is losing power for a moment causing the system to reboot? With hardware that old, it could very well be an issue with the PSU or motherboard.

If software, have you run MSCONFIG and taken a look at what's launching at start, then disabling items to see if anything helps?

(A final note: your system falls into the dreaded Capacitor Plague period and so I would limit any extreme measures to keep it running smoothly.)
posted by selfnoise at 1:12 PM on February 2, 2016 [1 favorite]


Response by poster: Yeah, capacitor plague was my first thought as I had to re-cap my old macs and it was a known issue with some of the early shuttle motherboards. But they are surprisingly fine. It doesn't seem to lose power, it just gets to the point where windows is loaded, finishes the start-up sound, and boop! it reboots. And then no problems at all for the rest of the day. I considered the PSU but it seems odd that it runs fine after the initial problem. I have even given it some pretty intensive multi-tasking to see if I can get it to crash. Nope.
posted by fimbulvetr at 1:22 PM on February 2, 2016


Is it a fresh install of Windows 98? What happens if you set the clock back to 1999?
posted by I-baLL at 1:39 PM on February 2, 2016


It's a long shot but I had this problem once with a system of about that vintage caused by a poor electrical ground. While in that case it was the building transformer an easy thing to try would be to plug the unit into a different circuit even via extension cord. Preferably something on its own circuit, possibly your refrigerator or a dedicated microwave outlet.
posted by Mitheral at 1:43 PM on February 2, 2016


This really rings a bell from my long-gone still-doing-Windows-desktop-support days, but I can't quite think why. The PSU seems like a long shot, but then again, power supply weirdness has been a larger percentage of my PC hardware problems than anything else I can think of...

Any chance something's up with a fan and it's overheating? I guess that doesn't make much (or any) sense if it runs fine after the reboot.
posted by brennen at 1:43 PM on February 2, 2016


Response by poster: It is year 2000 compliant!

I did a burn in with the system monitor running. It gets hot, but doesn't overheat. The fans are definitely all running, even the wee one in the tiny shuttle PSU. I forgot how noisy fans from that era were.

I'm betting on *something* in the start-up routine that only triggers once a day, but I can't figure out what the heck it is.
posted by fimbulvetr at 1:59 PM on February 2, 2016


Troubleshooting Windows 98 Startup Problems

I would start by checking the boot log. I had a similar problem in Win7, had to disable a problem task (Google Update making my system crash!) that was scheduled to run once a day. This made it really difficult to troubleshoot, because I could only recreate the conditions once every 24 hours...
posted by 0cm at 3:01 PM on February 2, 2016


I'd try getting it booted, run it for a (short) while, power off, leave for a couple or more hours, then boot it up again - this'll help narrow down if it's a "once a day" thing or a "cold boot" thing.

My inclination would be to blame hardware. Something like the PSU providing shoddy power for the first couple of seconds until something warms up, and that screws some other component trying to initialise, but when the system resets everything with a warm PSU it's all fine.
posted by russm at 11:12 PM on February 2, 2016


If you suspect a PSU warm-up issue, boot the thing into the BIOS menu and let it sit there for about twice as long as it normally takes Windows to boot, then give it the three-finger salute and let it start up again. If it's a bad PSU, this should make some kind of difference.
posted by flabdablet at 10:00 AM on February 3, 2016


Shuttle power supplies, and sff power supplies from that era are garbage. My old work and clients I've supported have had multiple issues with them. I'd take the exterior off the case and test with a normal ATX power supply before proceeding. I'd also, having messed with a glitchy ancient shuttle in the past year, give a very close mk.1 eyeball inspection of how every component is seated and whether stuff like the pci cards sits square in their slots.

After that, I know you said you checked drivers but I'd try and find different VERSIONS of the drivers. I'd start with the chipset driver.

On preview... Does it do this in safe mode? Have you used a tool like codestuff starter to see what's REALLY running at bootup? One of my first pot shots would be that something is running some kind of check that causes a reboot but fails to run after the reboot. Or runs, fails in some way, but is able to run after its failed once. This could be something really silly like a sound card tray app.

Another on preview is, some app could be expecting a file it can't read.

Have you verified the hard drive is healthy?
posted by emptythought at 12:07 PM on February 3, 2016


I'd also run a full memtest86! A warm reboot is not the same as a cold boot ram wise. There could be some slightly faulty ram in there. I've seen that cause a lockup or sudden reboot before that doesn't repeat.

Reflecting more this is the first thing I'd try after testing another power supply.
posted by emptythought at 12:08 PM on February 3, 2016


Response by poster: Lots of good things to look into here. It does start fine if I turn the machine off after it goes through through the reboot thing once, and come back later the same day, even a couple hours later, and turn it back on. The most vexing thing about chasing down this problem is I have to wait a day to see if whatever I tried was effective. I haven't run memtest on it yet, so I'll do that, and I have a spare full-sized PSU I'll have to dig up and try out.
posted by fimbulvetr at 6:54 AM on February 4, 2016


It does start fine if I turn the machine off after it goes through through the reboot thing once, and come back later the same day, even a couple hours later, and turn it back on.

In that case, the very next thing to do is find out how it behaves after being started in Safe Mode.
posted by flabdablet at 8:36 AM on February 4, 2016


« Older Name this blonde commercial actress with layered...   |   How do I pack for Los Angeles? Newer »
This thread is closed to new comments.