PDA

View Full Version : Strange Start Up Windows 98



andrew93
10-01-2005, 12:24 PM
Hi everyone

I'm having a bit of trouble with a friends PC and whilst I would *like* to fix the problem with a hammer, unfortunately it's not an option.

It's a Pentium 2 PC no name UK retail brand (well PCP brand but it's meaningless to us Kiwis) running Windows 98. It was running like a dog when I got to it, so I flushed out a lot of the garbarge using Ad Aware as well as Spybot. I fixed a couple of minor problems with Stinger and subsequent scans with AVG 7 and Nortons show there are no longer any viruses. I also upgraded the RAM from an earth shattering 32Mb to 96Mb (i.e. 3 x 32Mb PC100 DIMM, SD-RAM). I did some minor tweaks with a couple of the non-essential settings in the BIOS and turned off all of the applications that had put themselves into the Windows -> Start menu. I also did a number of critical and recommended Windows updates.

Now, when I cold start the PC (as opposed to restart) I get a blank screen with faint grey lines in a grid pattern on the screen and the PC does not start. There is no BIOS beep, no whirring of the hard drive and Windows is nowhere near loading. It's as if the PC has gone into sleep mode and won't wake up & the keyboard does not respond to Ctrl-Alt-Delete (although the numlock light is on).

Being an older PC it has a 'reset' button on the front panel. If I press 'reset' while the PC is comatose then the PC boots into life as if nothing has happened and everything continues completely normally. I personally wouldn't be too upset by this but it's not my PC and it didn't do it before I started to fiddle with it. However, the problem goes away when I do a Windows restart - but if I shut it down and attempt to start the PC using the on button on the front panel then it's pot luck whether or not it will start (sometimes it will but often it won't).

I thought it might be the RAM so I downloaded MemTest 86 and there are no issues with the RAM. I also checked out the power management options and there was nothing there. I did a quick search through PF1 and whilst I found this thread :
http://pressf1.co.nz/showthread.php?t=52549
... I don't think it is the power supply but I will look at the mobo during daylight tomorrow for any crusty capacitors. I also did a couple of virus scans and the PC is clean.

What I am looking for are suggestions from other PF1 members as to what I could look at next to solve this non-start problem.

TIA, Andrew

theother1
10-01-2005, 12:48 PM
Andrew,
I returned from holiday on saturday to find the family comp doing stuff it shouldnt ("we didn't do anything to it, it just went like that").
Fortunately, before I went away I had made a set of AVG7 recovery disks and was able to restore the settings that saw it working perfectly well before I went away. Maybe you can do the same??
Good luck.
Rob

FoxyMX
10-01-2005, 12:54 PM
I thought it might be the RAM so I downloaded MemTest 86 and there are no issues with the RAM.
That does not always mean that the RAM is OK. Just to eliminate it from the suspect list I would remove the new sticks and see what happens.

andrew93
10-01-2005, 01:02 PM
Fortunately, before I went away I had made a set of AVG7 recovery disks and was able to restore the settings that saw it working perfectly well before I went away. Maybe you can do the same??

Hi Rob
I installed AVG after it started playing up because I thought Nortons was part of the problem, so recovery isn't an option. Thanks for the suggestion anyway - you've reminded me that I need to create some new recovery disks (although I'm not sure how useful they will be).
A

Prescott
10-01-2005, 01:28 PM
possible psu failure?
does this system have a graphics card?
try clearing the cmos and rearrange the ram sticks.

Billy T
10-01-2005, 01:51 PM
I'd check the health of the CMOS battery as well. My old P166/W98 box develops cold-start problems when the battery is flat. Just had to change it again last week.

Cheers

Billy 8-{)

Graham L
10-01-2005, 04:13 PM
That could be right, Billy.

The failure on cold start/good as gold on restart is typical. Once you have had the power on for a few moments, the BIOS registers have been refreshed after an automatic default configuration, and is running on power from the wall. I'm guessingh that this is an AT PSU, not an ATX, so it doesn't have standby power.

Usually the problem is that the BIOS complains that it hasn't got valid settings ... total failure is odd. I can't believe that a few sticks of RAM would overload the PSU.

Check that you didn't bump the battery while you had the box open. After all -- this didn't happen until you fixed it. :D It's always the way, disturbing something often provokes faults. (Telephone exchange people will remember "Provocative Maintentance").

andrew93
11-01-2005, 12:24 AM
Thanks for the suggestions everyone. I will try taking the RAM out and eliminating that as the culprit as well as swapping the sticks around. I didn't really want to go there given it runs like a dog on the 32Mb but I'm going to have to try that.

I don't think it is a PSU problem - it feels like a BIOS / CMOS problem given there is a partial startup (I can hear the PSU fan and the HDD light comes on briefly) but with no beeps so I will also try flashing the BIOS. It does have a graphics card (4Mb?) but being an old system I think it is the old PCI type.

Regarding the CMOS battery - I was careful not to 'knock' or touch anything when I inserted the RAM but yeah, sometimes things happen. ;) If the CMOS battery isn't seated correctly or is going flat wouldn't the system clock be wrong? I have seen flat batteries before and you can usually tell because the system clock starts to fall behind - this PC isn't doing that.

Thanks again everyone, will keep you posted.
Andrew

memphis
11-01-2005, 01:56 AM
Could the on/reset cables be in the correct plugs for each of them,ie on is on on, reset is plugged into reset on the motherboard?
Just check everything is connected and where it should be nothing loose etc.
Just making a guess sorry I dont know. :dogeye:

FoxyMX
11-01-2005, 10:30 AM
If the CMOS battery isn't seated correctly or is going flat wouldn't the system clock be wrong? I have seen flat batteries before and you can usually tell because the system clock starts to fall behind - this PC isn't doing that.
Yes, the clock normally falls behind (or sometimes rushes ahead) when the battery starts failing but not always.

My money is still on the RAM. ;)

Graham L
11-01-2005, 02:25 PM
Flashing the BIOS is something to do when there is a known bug in the BIOS code, which has been experienced on that computer.

It is very unwise to do it on a computer which has worked normally (even if slowly) until hardware is changed, and now doesn't boot reliably.

Look for a SIMPLER cause.

Is the power cord plugged in fully at both ends? (Silly, I know, but it happens.) IEC plugs can be "nearly" in and "sort of" work. A computer PSU pulls 80 amps or so (briefly) when it's turned on. Or tries to. A poor contact won't let it get that.

Are the power supply connectors to the motherboard properly in? Pull them out then push them in again.

I'm tending to suspect the power supply area. The PowerOK signal from the PSU holds the CPU reset until the DC voltages are correct and stable. When it releases the CPU, it starts the POST code in the BIOS chip.

You can't have a ram stick in backwards, otherwise it wouldn't work ever.

The date/time being correct probably eliminates the battery.

Before PCs and the wonders of Microsoft software, the first step in computer maintenance was to check every connector. Starting at the wall plug. They were always the least reliable elements.

andrew93
13-01-2005, 12:42 AM
Thanks everyone for your suggestions.

Status update :
no loose connectors or cables
all cables are plugged into the correct connectors (see below re glue)
no crusty or leaking capacitors
changed the BIOS battery but problem still occured
didn't need to flash the BIOS because changing the battery had the same effect to the point where I ended up with an IRQ conflict and the comm port was dropped by Windows such that I couldn't get on the 'net - since resolved

I didn't want to face up to reality but in the end I had to look at the RAM again (thanks for the persistence FoxyMX - I too suspected the RAM but didn't want to admit it, particularly to my friend). I removed RAM stick 2 and put stick 3 into the RAM slot 2, did not replace stick 2.

The cold start problem seems to have gone away but unfortunately I can't say it with 100% certainty because it is not a problem I can recreate at will.

Lastly, I'm not sure if NZ assemblers do this but inside the case, all cables and connectors had a blob of glue on them as well as the PCI cards. I think the assembler must have done this to stop unscrupulous customers swapping components and then claiming the PC doesn't work. Has anyone else come across this? Is this normal practice? I have never seen it in any of the PC's I have bought or handled over the years.

Thanks again to everyone.
Andrew

FoxyMX
13-01-2005, 11:19 AM
I didn't want to face up to reality but in the end I had to look at the RAM again (thanks for the persistence FoxyMX - I too suspected the RAM but didn't want to admit it, particularly to my friend). I removed RAM stick 2 and put stick 3 into the RAM slot 2, did not replace stick 2. The cold start problem seems to have gone away but unfortunately I can't say it with 100% certainty because it is not a problem I can recreate at will.

Just see how it goes for a week or so. If it is the RAM you may be able to exchange it if it is faulty or incompatible so don't think the computer has to make do with less.


Lastly, I'm not sure if NZ assemblers do this but inside the case, all cables and connectors had a blob of glue on them as well as the PCI cards. I think the assembler must have done this to stop unscrupulous customers swapping components and then claiming the PC doesn't work. Has anyone else come across this? Is this normal practice? I have never seen it in any of the PC's I have bought or handled over the years.

Uh oh, here we go again.... :eek:

<Waits for the PC Company fans to extoll the virtues of hot melt glue.> :D

Graham L
13-01-2005, 03:35 PM
Hot glue on connectors or plug in cards are a Good Idea. If the computer is to be transported for hundreds (or thousands) of miles, it's a way to ensure that it still works when it arrives.

I once unpacked a computer which had travelled 2000 miles in a Hercules. All the cards had fallen out. I was pleased that the NiCd cells on the memory cards hadn't shorted and caused a fire. I wasn't pleased with the guy who built the computer, and had ignored my suggestion that it was to be exposed to extreme vibration.