I'm building my parents an Ubuntu box because I want to give them a simple system where they don't have to worry about viruses and such and where I can help them troubleshoot things remotely.

This is my second time building a computer and I think I just did something very bad. I assembled the thing and plugged it in (I set the switch to 230 volts, even though I'm on a 120 system, just to be on the safe side), and lights turned on and fans went a spinning and it seemed ok. I had my monitor plugged in but nothing appeared there and I tried to turn the thing off with the power switch on the front and that didn't work though it did turn off for a second, but then it would turn back on again. It wouldn't stay off.

So I figured that maybe I should turn the thing to 115 volt setting. So I did that and flipped the switch on and the same thing happened. I kept playing around to see if anything would appear on the screen (nothing ever did) and then suddenly the computer went dark, like a fuse broke. Now it simply won't turn on.

Now maybe I blew out the processor. That's a possablity I suppose. I never put any cooling compound on it, but that's because it looked like the cooling compound was already on the heatsink that came with the barebones system. In fact the manual said "Don't eat the thermal interface material". There was no tube of the compound material with the barebones system so I figure it must have been the gray stuff on the heat sink.

Anyway. Any idea of what I should do--how I should proceed to troubleshoot this thing? The barebones system I have, FYI, is:

http://www.newegg.com/product/produc...82E16856110061

Thanks in advance.

~daveZ

If you are on 110 power source why did you set it to the 230 setting to be safe--that is not the safe thing to do. That setting is for systems that are running on 220 powers sources. You probably blew the power supply, before it got to anything else. It is possible that you blew more then the power supply though. The first thing to do is replace the power supply. Power supplies that have the voltage option (many don't anymore because the are able to detect the input now) usually are set for the country they are sold in. You generally do not need to make any changes to this setting. But--if you are in the US or the few other countries that use 110 volts for power you want to make sure the setting is on the 115-- And if you are in the many countries that use 220 volts for (which you apparently are not--but if you ever happen to travel to one) you want to make sure the setting is on 230. Hope your problem did not go beyond the power supply if it did it is time to replace the hardware. This is generally not covered by hardware warranty as it is user error.

oh no...without looking at it I can't be 100% sure...but I am 98% sure.. you fired your cpu. A modern cpu can get to over 600 degrees in seconds. I think the system was overheading to the point it wouldn't function, but you kept turning it back on. One thing you could try, disconnect the power, (after touching the power supply to get rid of any static discharge) then pull the bios battery for at least 2 minutes. Then put it back in, plug in the power, with the correct settings as far as the 120 or 240 goes, then reboot with a live cd in and hope. That pad thing was your thermal paste, but you have to make sure it's seated correctly, and sometimes it comes with a bit of plastic over it, did you take the plastic off ?

David

As previously commented, there is no reason to think that this option is 'safe': In fact it may be the very opposite.
This could be a BIOS issue, Remember the on/off switches on the front panel aren't true 'disconnect the power, on/off switches'.

Did you actually check whether a fuse blew (you know, with a DVM)?

You won't like this (I guess that there is a more polite way of putting this, but I figure I might as well hit you with the unvarnished version)....but because the manufacturers generally try to make their systems idiot-proof (and not go into abrupt meltdown if the cpu fan fails), they should shut down if overheated....doesn't always work though. So, it should have survived. That plastic shim is a possibility, though.

Although, keep turning it on and off will probably do it in eventually.
i) does the power supply still work?
ii) have you cabled it correctly? Is there a short in any of the cables/connections?
iii) Do you (still?) have any friends with whom you can try swapping bits to see if they work in another system? Are you likely to if your processor breaks their system? Can you go back to the shop?

crap crap crap. I set it to 230 (which was the default) because the manual said that "switching to 115 V in a 230 V environment will seriously damage the system!" So I figured the converse was not true.

I'll check the power supply. But maybe first I'll need to get a real volt meter. Right now all I have is one of those two/light gadgets -- with a light that lights up on 120 volts and and another that lights up on 240 volts.

I did take the "plastic" off. It wasn't plastic, actually it was little rubber things that sat to the side of the thermal coolant presumably to keep the coolant from touching anything. I didn't see any plastic covering it. The manual said I should take the rubber off before I attached the heat sink.

OK. So as for now I'm going to try the following. Disconnect everything from the power. Plug the thing in and test for voltage. If I don't get a signal there I know at least that that's something that needs to be dealt with.

Question: If I don't see power coming out of the power box, does that mean that the rest of my system is likely still good? Or am I being too optimistic?

Presuming that the power does work, then I suppose it gets more complicated. Is there a way to test to see if the processor got fried without plugging it into another system?

~daveZ


EDIT: OK. I just checked the power with my cheapo 120/240 volt tester and I don't see any juice coming out. I'm going to play it on the safe side (assuming that this really is the safe side this time) and get a real volt meter to test it. I've been wanting to get one anyway. Thanks a lot for everyone's comments by the way. LQ forums rock.

How did you test it? This is how you generally test an ATX power supply:

1. Take the power supply out of the computer.
2. Turn the rocker switch off, switch the power supply to the 115V setting, plug in the power supply (to mains), turn the rocker switch on.
3. On the large 20-pin connector, short pins 14 and 15 (e.g., with a paper clip). The power supply should now be on (if it has any internal fans, they should turn on).
4. Use a voltmeter to check the DC voltage with respect to ground (any black wire) of the 3.3V rails (orange), the 5V rails (red), the 12V rails (yellow).
5. You should do this with the various four-pin power connectors as well (the ones you use to power drives and such). They will not have orange wires.

You should probably look at a pinout to be sure which pins you need to short. The pinout will also show you the colors (and voltages) of the other pins, so you won’t have to strain too hard. On the four-pin molex connectors, the wire colors are easy to see.

This test of course only makes sure the voltage readings are correct, and in particular, it does not test the performance under load. The only way to do that is to use a good BIOS which has a PC-Health Status section (which you should try after you verify that the voltages are sane). Obviously, if you cannot even get the motherboard to POST, you will have fried something on the motherboard.

OK. I bought a real voltmeter. Testing the power box again I can confirm that there's no juice coming out of it. I guess this makes sense since the fan doesn't run. (The power box fan should run if the power box is working, right?) I opened up the power box and didn't see any obvious fuses.

So do you think I should just replace the power box and hope for the best?

~daveZ

Osor, I just saw your post. I hadn't seen it before I wrote my last post. Hold on I'll short the 14 and 15 pins to see what happens. ~daveZ

not sure if i saw it, but is there power going to the outlet right now? you could have tipped the breaker.
you can most likely test your system with pretty much any ATX power supply you can pull out of another computer.
good power supplies have internal protection circuits, much like breakers but they reset themselves when the power is unplugged. such power supplies protect your system from many abmormalities in the power grid.
give another power supply a try, that could be an easy solution.

Actually, short pin 14 to ground momentarily. Pin 15 is a ground so that is OK, but the emphasis is on momentarily - don't hold the short any more than you hold the power button on your computer in (which is what you are simulating).

As others have said, setting the mains to 230 on a 115 system isn't a good idea. I have no idea what happened to you, but blowing the fuse in the PS is certainly a possibility. I have my serious doubts about any downstream problems from doing this; my first thought is that you would just never get a power good signal in the PS and it would therefore crowbar itself off.

I really can't say, though, because I have never tried it and I have never studied a PS schematic with an eye to doing that to see what would happen.

Ok. Well I did short 14 and 15 momentarily following the instructions here:

http://www.duxcw.com/faq/ps/ps4.htm

And the fan spun. So I guess that means I blew up something on the motherboard. Hopefully that doesn't include the processor.

So I guess I should order a new motherboard, then?

Question: If I get this new motherboard and I also have a broken processor that I install in the new motherboard, will I hurt the new motherboard?

I'm a hardware dunce (as y'all can probably tell) so any advice on the next step would be handy. Thanks,
~daveZ

typically the cpu fan will come on as well if the board is still good, even though the cpu is fried.
so if no running cpu fan i would assume that the board went bad.

Not sure what you mean about the 'rubber' stuff on the heatsink but many cpus come with a heatsink compound already in place. It's usually a think lump with a plastic cover. The usual thing is to buy some good heatsink compound, clean up the cpu and heatsink and apply your own.

If you took off the compound that was supplied but never put any proper compound on, you may have fried the cpu, or it may just be powering off at a certain temperature.

First thing though is to check the PSU is working and make sure there is power going to the board. Then check and recheck that everything is in place and connected. Make sure you put some good heatsink compound on the cpu - don't overdo it, it only takes a thin layer.

Another thing to look out for are the posts that hold the motherboard to the case. Remove _any_ that don't have a screw through them. Only leave the ones that are holding the mobo.

To test the system only connect up what is really necesary to start up and get into the bios - 1 ram module, pc speaker, on/off switch and graphic card. Then there is less chance that a faulty component will freeze the system.

If you manage to get into the bios with no beeps start adding drives, sound card etc and retest.

Ok. So I will replace the motherboard. What a freaking pain. Now if the motherboard went out that probably means the CPU did not, right? I'm just worried about using a broken processor in a new motherboard and frying the new motherboard. ~DaveZ

Well the compound on there wasn't anything like a "lump", but it was a thin grey layer.


I never took any compound off.

There is power going to the board. I connected it again and a green LED on the motherboard turns on when I flip on the power. However, the optical drive, when connected, won't run anymore and neither does the harddrive seem to move anymore. (I tried all this stuff before I read your post).

Do you recommend that I put on my own coolant compount despite the compound material that was supplied with the barebones system?

Great. Thanks for the advice. I'll check this thing out before I finally order a new motherboard.