Strange start up problem

[twreeve]

I have a Dell XPS 8100 running Win7 64. Day to day everything works fine, except that is until I turn it off (as opposed to sleep / hibernate). If it is ever powered down I need to carry out a convoluted start up process I discovered by accident. My procedure it to 1) unplug the mains cable, 2) hold down the power button until the flea / motherboard orange light goes out, 3) disconnect the 24 pin connector, 4) plug back the mains cable into the chassis, and finally 5) simultaneously hold down the power on button, spin the cpu fan, and reconnect the motherboard connection. (and yes I know that playing around inside a mains connected pc is hazardous...). Without this process (especially the manual spin of the fan) it won't start, bleep, spin or anything. The fan incidentally spins freely, and the power button works flawlessly in all other situations..

So clearly I have a problem. So my question is aimed at the more cerebral tech wizards; what might it be the issue - what is it that I am doing (draining the circuit, spinning the fan, etc that allows the system to function, and what component should I examine.

Many thanks in advance.

 

[Ralston18]

Have you ever replaced the CMOS battery?

And do immediately stop "playing around inside a mains connected pc". It is hazardous and will not end well at some point.

If the battery replacement does not help, then I would suspect a failing PSU.

 

[twreeve]

Thank you Ralston18. I have indeed replaced the CMOS (I should have mentioned that). And yes I agree with you re the hazards (though you should see what I get up to for leisure ;-) ).

More seriously, I agree the PSU probably needs replacing, but for both intellectual curiosity, and also frugality should it not be the PSU at fault, I was wondering if anyone could cast any light on why this phenomenon is occurring: is the spinning of the fan creating some kind of voltage that triggers a circuit? Is the power button connectivity not working due to some residual charge that needs to be dissipated first? I fear this is a deeper level of knowledge than most normal IT specialists (to use a medical analogy more of a consultant than a GP/MD), but perhaps someone who understands the intricacies of the PSU circuitry might be able to shed some light....

 

[Ralston18]

I am not a PSU expert per se and will happily accept other suggestions and comments.

Most things are made not to be repaired and as the complexities involved increase it becomes difficult (at least for me anyway) to be able to explain exactly what is happening and the why. One cannot simply get into the "guts" of things any more to examine, test, swap smaller components, etc..

PSU problems can manifest in many ways and vary accordingly. They can be damaged by incoming power related problems and can pass that damage along into the host PC's motherboard and components.

You can with some knowledge and skill (plus lots of caution) test the outputs of a PSU to determine if there is a voltage problem on any given output connection. And perhaps eliminate the PSU as a problem all together by such testing.

Here is a good starter link about how PSUs work. There are many such "tutorials" to be found online with as much detail as you desire.

https://www.lifewire.com/power-supply-unit-2618158

No need or reason to go into the internal circuitry to troubleshoot.

Do Not Open the PSU under any circumstances. Just test the outputs if you wish to go that far.

Agree that the problems could be elsewhere.

Key is to methodically and safely troubleshoot and identify the cause either directly or by elimination.

 

[twreeve]

That's an extremely good pointer. I shall do some research and have a think.

Thank you.

 

[smorizio]

if the system is from 2010 it may have bad cap. a lot of older dells had cheap cap that used to fail causing no post or post issues.

 

[twreeve]

Thanks Smorizio. Good point. It is indeed a Dell from May 2010 (its an XPS8100). I didn't think of that. I do have a solder iron, and while I like to tinker, replacing the capacitors might be stretching my skills. I will have a look for bulges next time I have to open the box.

 

[smorizio]

with any power supply you should be able to jumper it to come on and then with a volt meter check the output on the pins.

 

[twreeve]

Thank you. I shall. Fingers crossed.