PDA

View Full Version : "Disk boot failure" and FDISK /MBR Command



Randolf
15-07-2005, 10:16 AM
My Windows 98SE system sometimes doesn't boot up properly and hangs on the screen message "Disk boot failure..." I am able to make it boot properly by inserting my emergency start up floppy and on the A: \> prompt typing FDISK /MBR. then all is OK for a few days then the problem comes back.

Would it help to also type SYS C: at the prompt?

One tip I read said to type "C:" first, then the above two commands. What effect would this have?

Why does typing FDISK /MBR on the A: \> prompt with the emergency S/U disk in fix the proplem, albeit temporarily?

What else should I be doing? Would prefer nothing too drastic if possible, as it seems to me the problem can't be too bad since it can correect and reboot properly for a while.

Any help would be most welcome!

Murray P
15-07-2005, 10:39 AM
Firstly, I'b be backing up all my important data to somewhere other than the drive in question. You may have a failing hard drive.

Then check that the drive ribbon cables are all securely in place. Next run the drives diagnostic utility. If they have a utitlity that can differentiate between the IDE controller and the logic board on the drive and test the communication between them would be ideal (rather than just a surface scan of the disc).

Follow the instructions for repairing/replacing the MBR for the system directory C:, not the A: floppy disc you have inserted.

pctek
15-07-2005, 02:10 PM
Yes sounds like imminent HDD death is approaching.

Graham L
15-07-2005, 03:13 PM
...
Follow the instructions for repairing/replacing the MBR for the system directory C:, not the A: floppy disc you have inserted.

It has been fixing the hard disk's MBR. FDISK doesn't know about floppy drives. ;)

SYS C: just replaces the DOS boot (hidden) files and the COMMAND.COM (which is the DOS OS). If the "fixed" MBR boots OK they are all right.

It seems a bit odd. It's unusual for just the first few sectors (or any) to become "temporarily" bad. I'm wondering whether the drive is just a bit too slow in getting up to speed. (Cold weather, drive getting a bit old and tired ...)

Instead of booting a floppy when this happens, try just a Ctrl/Alt/Del reset. If I'm right about the drive being a slow starter, it will boot with no more nonsense. It will be up to speed by then.

Of course it could be symptoms of a forthcoming bad failure. :D

Randolf
15-07-2005, 03:51 PM
Yes, it does seem to happen more when the computer has been off overnight or for a day or so. I read somewhere that older hard drives can develop a sort of film on them, which slows them down just long enough to miss the boot command initially?? I will try what you suggest meantime. A bit puzzling at present. :confused:

SurferJoe46
15-07-2005, 03:57 PM
check psu voltages at start too..sometime a component in it fails ever so slowly and lets the boot voltage drop too low with all the demand at start up.

just a thought.......................

Randolf
16-07-2005, 03:43 PM
[QUOTE=SurferJoe46]check psu voltages at start too
QUOTE]

What's the easiest way to do that please?

Randolf
18-07-2005, 09:43 PM
check psu voltages at start too..sometime a component in it fails ever so slowly and lets the boot voltage drop too low with all the demand at start up.

just a thought.......................


Thanks. This could be valid comment, as PSU is only 250 watts. But how do I check the voltages?

Murray P
19-07-2005, 03:46 AM
Apart from a multi meter, I guess, most modern(ish) BIOSs will show your running voltages for various things. Sisoft Sandra may do the trick, but it does assume some things.

SurferJoe46
19-07-2005, 06:08 AM
The problem with using the bios monitor(s) that you may have is that you are gonna see the values once you are up and running...too late! It will not record or show the starting voltages.

I suggest gently probing the 4-pin connex at the fans or the HDD wires and find the 12 VDC line while the unit is up/running, and then do a reboot and watch the voltages change.

Try this also for the 5 VDC circuits...but I recommend NOT probing the 3.3 VDC areas at all, they will be a waste of time and they aren't that much affected by the other side of the PSU anyway..not that they MIGHT NOT be bad, but they are really for the MOBO anyway, and may not suffer as much.

DO NOT use an analog VOM! The load that it creates is possibly gonna throw everything into the toilet anyway. Use of a Digital DVOM/DVM will not load the circuit you are testing and make a lot better guesstimate for what is happening.

SurferJoe46
19-07-2005, 06:18 AM
PS:

Here's a neat site with some voltages and values for a lot of ...well...everything that you might need for what you are looking: PINOUTS AND PORTS (http://www.pccables.com/cabtech.htm#power)

You should be able to find just about anything you need here, and they are nice enuff to segregate the various systems for you too!

One thought: just probe from a colored (red, yellow, blue, green etc) TO the Black wires, as Black is Ground, (Earth to NZ's I think!), and you'll get some really weird values if you probe across...say...red to yellow!

<sorry as there's no YELLOW on this site...hadda use ORANGE>

Graham L
19-07-2005, 02:18 PM
Unfortunately, using a digital meter will not help. At all. DVMs don't show changing values very well at all. They use samples. Slow samples. So even in the unlikely event that this is the problem, you won't see anything. A DVM will show the first stable reading it can get. After everything has happened. The only really useful instrument for looking at rapidly varying voltages is an oscilloscope. Preferably a digital one to see one sweep.

What sort of analogue meter have you got, Joe? Old moving iron switchboard meters built by Edison? Most of mine pull 50 microamps at full scale They are the industry standard 20000 ohms/volt" meters. I've even got one with 10 Meg input resistance, just like a DVM. It's got a FET amplifier. ;)

Even a very cheap 5000 ohm/volt (200 microanp full scale) meter won't bother a 4 or 6 amp 12 volt supply.