comp wont start unluess I unplug / replug it into the wall

[malagash]

running win 2k fully updated and scanned for everything on an ibm P3 866mhz, 256 meg.
computer won't start unluess I unplug / replug it into the wall. I get the "operting system not found" message.
when i unplug it and reinsert the power cord to the wall it starts normally...
TIA

 

[biglebowski]

Try pushing and holding the power button in for a minute or two with the power cable disconnected.

When I was born I was so suprised I didn't talk for 18 months

 

[malagash]

When I unplug it and replug it back in it starts normally. By discharging the mobo (which I understand happens when you hold in the start button) how would that correct such a problem?
Thanks Big

 

[paparazi]

Check for bulging/leaking capacitors on the motherboard.
See:

http://www.badcaps.com

For a guide.

Otherwise it could be a suspect power supply.

Martin

We like members to GIVE and not just TAKE.
Participate and help others.

 

[ceh4702]

Maybe the Power supply is shot or the power switch is not working correctly. It is not normally the switch because of how the newer power supplies work. You may see a bad motherboard or a bad power supply cause this. It could have bad components like a capaciter that leaks or something like that. Stuff just wares out sometimes.

If you do not like my post feel free to point out your opinion or my errors.

 

[DTracy]

I had a workstation with the same symptoms. I removed all peripheral devices and started with a known good keyboard, mouse, and monitor. Adding devices until the system failed again I found that a USB barcode scanner was causing the startup problem. Changed the scanner, solved the problem.

David.

 

[biglebowski]

have you got any usb devices plugged in, I had a similar issue with a usb hard disk. With it plugged in if I rebooted/started the pc it would hang at bios, if I then unplugged the power and put it back it would start ok. Plugging the usb hard disk into a usb hub resolved the issue.

When I was born I was so suprised I didn't talk for 18 months

 

[malagash]

There are no bulging caps and all looks fine on the mobo. I have seen units with cap problems and there is no instance of that.
The only USB device is the printer with which I tested and all was fine.
The only thing that remains is to swap the PSU the next time it happens and cross fingers.
Thanks everyone

 

[biglebowski]

If all else fails try this technical fine tuning tool

When I was born I was so suprised I didn't talk for 18 months

 

[malagash]

on the computer or the user?

 

[felixc]

Sorry for the vague answer, it can be the motherboard's standby circuit. It can be the power supply that does not behave properly when turned "off". It can be because of peripherals conencted to your PC.

When you power off from the front switch, there is still a standby voltage that the power supply feed to the motherboard. To turn that one really off you need to use the switch that is on the power supply itself. (some do not have one though...) Or to disconnect from the wall, like you've done.

Is there anything, connected to your computer like the monitor, that stays "on" when your computer is turned off? If these are "off" when you power the PC off-on, does it make a difference. Powered peripherals can reintroduce voltages into the PC and prevent a proper power-on on some parts of the PC.

I have had to modify I/O cards because of a printer port that reinjected current into a PC and prevented my card to power-on properly. So yes it can happen.

Tell us what happens after swapping the power supply.

 

[malagash]

I will have to get back to you as the unit has been returned to the owner. This is the 2nd time this has happened so, given the history I suspect that I shall be getting another call form him. When I do, I shall repost
Thanks

 

[spi200]

Hi

What you have described is called a "Crowbar" - this happen to a power supply when the limits are exceeded, often this is caused by a short - this is why its called a "Crowbar" because its like someone has thrown a metal object inot your computer and the powere supply is trying to protect it self. The "crowbar" condition is cleared by removing power to the device for a short period of time, usually 10 seconds will do it, but I have seen 3 to 5 seconds work as well. So you should not have to remove the plug from the wall or unit just switch is off at the wall and wait 10 seconds and switch it back on.
I this is what is happening then you have to find what is causing it - I usually found it was a fualty Psupply itself, but it can be a motherboard or other faulty device.
If your cover under warranty or similar take it back, if not start discontecting one device at a time. ie Floppy drive, Sound card, NIC card , DVD\CD Drive until you have nothing else but a motherboard, video and keyboard connected. If it still fails then my first choice would be a Power supply. You can of course check your power supply that it has a good 5 volts and 12 volts - without that your PC is not going far. Power supplys are cheap as chips, beyond that I would engage the local repair guy who can throw a mother baord or CPU at it to isolate the fault.
Hope this helps.

Good luck


Dave

 

[malagash]

Thanks Dave. I think I will go and retrieve the unit from the user and go back over it. I have another PSU if needed.

 

[edfair]

Sounds to me like the standby circuits are functioning as expected. As is the M/B except that it may be trying to access the hard drive before it has recalibrated.

My previous experiences with this go back to 8088 days when the solution was to hold the reset key for 2 seconds before allowing it to boot up. And the issue was resolved when I finally replaced the power supply. But to this day I still wonder if it wasn't the hard drive being slow.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[malagash]

except that it may be trying to access the hard drive before it has recalibrated.

Befroe the HD has recalibrated? Please describe the sequence.

Given that there is no reset button on the front and no switch on the back of the PSU the only option at the moment is to unplug it and hold the power button in for a few seconds which should reset everything however, that is not an option for the user (who happens to be a whipsnapping, cracking 90 year old gentleman).

I have all of his attachments except the monitor hooked up and thus far I have not been able to get this to fail (must be my magic touch). I'm thinking that my best bet is to replace the PSU and see what happens when the owner gets hold of it.

I'll let you know, thanks

 

[Grenage]

Yup, from your symptoms I would suspect and replace the PSU. It is the primary cause of this kind of problem/failure.

Carlsberg don't run I.T departments, but if they did they'd probably be more fun.

 

[edfair]

When a hard drive starts up from power down the motor cranks up. The controller starts counting the index pulses. When the count hits a predetermined mark the controller steps the heads away from the park position then recalibrates back to cylinder 0.

You might be able to check if this is an issue by escaping past the memory check or look in the CMOS for an option called fast boot. Anything that speeds up the boot should make the problem worse.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[CobolKid]

The “unplugme” is a classic sign that the psu's tripping one of it's overvolt/crowbar protection circuits. In the "old" days an overvolt/crowbar protector consisted of a silicon controlled rectifier (SCR) with it's cathode and anode wired across the particular direct current (D.C.) output's plus and minus, with a voltage sense circuit tied to the SCR's gate designed to turn the SCR on when an over voltage occurs. This protects against crowbars too, as a crowbar is simply a fast short duration overvoltage, in other words, a spike or series of spikes.

An SCR is an alternating current (A.C.) device, typically of a high current fast acting nature. It is turned on by the appropriate signal level voltage/current on its gate, and conducts across its cathode/anode until its gate is turned off AND its cathode voltage drops to (very close to) or crosses 0 volts. So, once the overvolt detector fires the SCR, the SCR turns on, shunting (dropping) the particular D.C. output by short circuiting it, and since the power supplies output is D.C, instead of A.C., the SCR stays on until the supply is turned off by unplugging, blowing a fuse, etc. etc. The typical SCRs fast response can clamp down an errant supply within a few hundredths of a volt, saving expensive hardware from catastrophic melt down.

Today’s protection circuits utilize integrated power management circuits that incorporate many additional features such as undervolt protection in place of just a simple overvolt/crowbar sense circuit, but the high current fast acting SCR is still an integral part, no other device will do it.

Although It may sound like the standby 5volt would be the culprit, it could actually be any one of the outputs, depending on the psu’s design, as the typical management chips are not limited to a one per one arrangement as were the circuits of the past, they tend to clamp down the entire supply when one critical output fails. This avoids damage that would occur if say the 3.3 volt supply section failed and all others remained on.

In reality, it may not be the supply, as several have mentioned, you could have a peripheral that has a failing device, for instance a power transistor (motor or solenoid driver etc.) that is slowly dropping impedance and is spiking the 5 volts with 12 volts, there are a myriad of possibilities.

Steve

 

[LawnBoy]

<pedant mode>

An SCR is an alternating current (A.C.) device

No. a Silicon Controlled Rectifier is a gated diode. It functions with direct current only.

You may be thinking of a TRIAC, which is basically two SCRs back to back. These are AC devices.

</pedant mode>