Sign in with
Sign up | Sign in
Your question

My computer will not start, what to do?

Last response: in Systems
Share
Anonymous
a b B Homebuilt system
September 29, 2009 3:46:13 AM

I think I had a power surge. Computer will not start. Power comes up and fans turn but it will not attempt to start --- even with a ubcd. Hard drive is unplugged. Video Board and cd are only things plugged in. My board is ASUS k8n-e. I pulled all memory and replaced it with one 500m device.

More about : computer start

September 29, 2009 4:02:52 AM

Before the surge, when you started it did you hear a beep from your tower, and if so do you hear it when you start your comp now?
When you start your comp, do your keyboard lights flash within 1-2 min. of power on?

If the answer to the above questions are answered with a no, you either have a dead psu (probably faulty, since the fans still spin, but unusable nonetheless) or you have a dead mobo. If you have a spare psu to test with I would recommend doing that before buying anything else.
September 29, 2009 12:19:04 PM

You can have single rails fail on a PSU, it is not unusual, so pepperman is giving you great advice. If the PSU is still good, it may be the mainboard CPU staging MOSFETs (also known as the CPU power supply) could have taken a hit. Either way, you are looking at about $100-150 to replace either the PSU, mainboard, or both. Test each part if you have the means to do so. If you can use/borrow a known good PSU, you can at least see if you can get the mainboard to POST. If you can get it to POST, then reconnect your hard drive. If you can get your OS to boot, put your memory back in. At this point, your chances are very good that the PSU was the problem. If you can't get the mainboard to POST on a known good PSU, then you will need (at least) a new mainboard. Once you can get the new mainboard in place, test each component to POST, because chances are they could have taken a spike from the old PSU just like the mainboard. Check CPU, memory, video card and any hard drives by adding each in one component at a time. If it all checks out, you should be able to boot into your OS of choice.

You can also go down to your local friendly computer store and they will be more than happy to do all this diagnostic for you at a reasonable fee. With my own built systems, I tend to diagnose myself, but then, I was the one to build it. If the system is a prebuilt or integrated system (like Apple or Dell) I tend to take those in to the shop just on the off-chance that the repair is under warranty.
Related resources
September 29, 2009 4:04:25 PM

Quote:
I think I had a power surge. Computer will not start. Power comes up and fans turn but it will not attempt to start

Keep buying new parts until something works. A majority will recommend that due to insufficient technical knowledge. If your car mechanic did that, he would be quickly unemployed for being unknowledgeable.

Obtain assistance from the better informed by simply providing some basic numbers. First put everything back. Disconnect nothing. Using a 3.5 digit multimeter, measure voltages on six wires from the power supply. First measure the purple wire (where wire enters the nylon connector on the motherboard. That should read about 5 VDC. Report all numbers in three digits.

Second, measure the voltage on the green and gray wires both before and when power is turned on. Notice what happens when power switch is pressed.

And finally measure any one red, orange, and yellow wires when and after the power switch is pressed.

Those numbers make possible a definitive answer in the next post. No more 'it could be this or might be that'. Either have definitive answers or just keeping buying parts until something works. Those are your two options.

A meter is intended even for the simplest among us. Sold in Radio Shack, Lowes, Sears, Tru-Value Hardware, Home Depot, K-mart, or some grocery stores. Sells in Wal-Mart for less than $18. Has many other purposes beyond a computer repair. The meter is how to find a problem without disconnecting even one wire. Is it the power cord, switch connector, or power supply controller? Thirty seconds and numbers means the next post would be that definitive and useful. No more 'try this or do that' speculations.
September 29, 2009 5:10:37 PM

I think popping in a known-to-be-working power supply (if he has one) would be a lot easier than connecting a voltmeter to the inside of a running computer. My recommendation is: if you have a spare psu or can easily attain one, try that; if not, and/or you are experienced at using voltmeters, use the voltmeter. If you need a guide, try the link below.
http://articles.techrepublic.com.com/5100-10878_11-1056...
September 29, 2009 7:25:29 PM

pepperman said:
I think popping in a known-to-be-working power supply (if he has one) would be a lot easier than connecting a voltmeter to the inside of a running computer.

The minute even one wire is disconnected, you have complicated the problem AND put the computer at risk for even more damage. A mechanic that shotguns a car would become quickly unemployed.

How complicated is a meter? Anyone who cannot use a meter definitely finds the Ipod or a cell phone completely too complex. Why are meters sold in K-mart and in stores that also sell hammers? Meters are that hard.

How to triple the amount of time and work? Swap a power supply. Then still know little that is useful. Remain confused. A defective supply can boot a computer. An otherwise perfectly good computer can fail toith a perfectly good power supply. Those who don't know how computers work replace parts on wild speculation - also called shotgunning.

Shotgun: replace the power supply on nothing more than speculation. And still know nothing abou the other components in a power 'system. Shotgunner never even learn that another power supply 'system' component can be defective. In so little time - and disconnect nothing - numbers from those six wires report everything.

That techrepublic.com article is obviously rediculous. Don't disconnect anything. Why disconnect a power cord? Don't measure a power strip. Why do obviously foolish things. Listed were six wires to measure. Numbers from those six wires automatically test all that in thirty seconds. The safest solution - and to know immediately what is wrong without replacing perfectly good parts - is to probe with a meter.

Either spend 30 seconds to use the meter. Have a solution immediately. Otherwise swap perfectly good parts - put hardware at much greater risk - until half a computer has been replaced.
September 29, 2009 8:43:13 PM

All he will know is whether or not his power supply is interfacing/working correctly with his mobo. Computers are a lot different than cars, btw. Even so, if you don't have a sensor to hook to the car's computer, you replace the most likely parts.

Let's say you own an autoshop; you have all sorts of test equipment and manuals there to reference to. I'm guessing our friend doesn't own a computer repair shop, where he would have access to computer test equipment and manuals, so I presented from what information he told us what my best guess from my experience with computer repairs. That recommendation, provided he has a spare psu as many system builders do, but not necessarily a voltmeter (which unless you are an electrician or repair electrical equipment regularly, you probably wouldn't have one), was to swap psu's and evaluate whether or not the problem was fixed.

If you have any experience with repairing computers, you would know that if the computer will not start after a surge, the psu is most likely fried. Whether or not other components are dead as well could be determined by hooking up a working psu and checking.

As I said before, your method of the voltmeter will only determine whether or not the psu is working. This would be fine if he has a voltmeter, but if he doesn't, yet he has a spare psu, I don't recommend he buy a voltmeter when he has the means to test just as accurately without one.
September 30, 2009 3:50:22 PM

pepperman said:
All he will know is whether or not his power supply is interfacing/working correctly with his mobo. Computers are a lot different than cars, btw.

Good diagnostic thinking is same whether it is computers, cars, or toasters.

A good power supply can make a defective computer work (temporarily). A defective power supply can boot a computer. Swapping parts only provides 'maybe' answers. A definitive answer means an immediate solution - no more confusion so often found in shotgunning.

Let's see. Computer would not start up after a surge. Was the power supply damaged? No. Because we learned this stuff (and therefore save time by not shotgunning), a surge protector earthed that surge destructively through all networked computers. All power supplies were perfectly good. Why? Power supplies are where serious surge protection is located. We learned why the protector causes damage to a network of powered off computers - in part because we did not wildly replace parts by shotgunning.

Power supplies are most often damaged by manufacturing defects - not surges. You would know that if not shotgunning. We who know this stuff literally traced failures to the semiconductor. D autopsies periodically to further learn. What did we learn? Enough to know that a meter is how solutions are obtained in the very next post.

Only the most naive fear a meter. Only the most technically untrained shotgun. Then spin myths to justify what is really the only thing they understand. Swap parts until something works. Learn nothing from the experience.

OP can have a useful answer by measuring six wires with a tool even from K-mart. Then learn from that failure. A person who literally designed computers can provide zero assistance if those numbers are not provided. 30 seconds so that the next reply immediately identifies the suspect. Only alternative is to shotgun: wildly replace perfectly good parts and remain technically naive.
September 30, 2009 6:26:23 PM

Please learn to convey your thoughts in clear, concise English, first of all: "Let's see. Computer would not start up after a surge. Was the power supply damaged? No. Because we learned this stuff (and therefore save time by not shotgunning), a surge protector earthed that surge destructively through all networked computers. All power supplies were perfectly good. Why? Power supplies are where serious surge protection is located. We learned why the protector causes damage to a network of powered off computers - in part because we did not wildly replace parts by shotgunning. " I can't tell if you are referring to our questioner (in which case I have no idea where you got that information as he never posted a that he had a surge protector setup) or if its coming from your experience.

Second, I believe our friend wants to get his computer working again, not necessarily learn the inner workings of a PSU (he may correct me if I am wrong and I will humbly apologize). Based on this assumption, I gave him the most relevant recommendation to fix his problem, on an assumed tool assortment (in this case NOT including a voltmeter). If he has a voltmeter, I would recommend, along with you, that he should use one. However, I assumed he did not have one, and thus recommended the use of a spare PSU.

BTW, if a PSU is damaged by manufacturing defect, how come it comes AFTER the surge (for which they are rarely rated to handle a major electrical surge)? Knowing what failed in the PSU because of the surge doesn't tell you it was due to poor manufacturing, it tells you which part of the PSU can't handle the excess electricity.
October 1, 2009 6:27:39 AM

pepperman said:
Please learn to convey your thoughts in clear, concise English,


You shotgun due to no hardware knowledge. You would have the OP replace perfectly good computer parts using wild speculation. You don't even know a surge existed. And yet you know a mythical surge must have damaged a power supply. Curioius that surge did not damage everything else in the house. Or do additional fact get in the way of wild speculation.

Meanwhile, in 30 seconds, the OP can obtain numbers. Next post then identifies the suspects. Why? Because one of us knows this stuff. Those who actually know something do not replace perfectly good parts until something works.

"I think it was a surge. Therefore it was a surge." OP only defined symtoms typical of a manufacturing defect. But somehow you know it must be a surge. After all, if you feel it is so, then it must be true. Also explains your posted mayhem.

OP can have his computer fixed the first time, faster, and with less money by simply posting numbers.

October 1, 2009 1:04:09 PM

Westom, I'm glad you think everyone has enough experience using a multimeter to be able to diagnose a failed power supply by testing each rail. If the OP had that level of experience, then they probably would have already done it. I keep my multimeter handy for just such circumstances, but I also have a PSU tester on my bench that does a much better (and faster) job for me. I have this because I build systems and need it on a regular basis.

We are not proposing a shotgun approach at repairing his down computer. We are suggesting that he test the most likely components (ie, the PSU, and if that is good, the mainboard). How do you test a PSU if you (a) don't have a multimeter or the wherewithal to use one, and (b) don't have a bench tester? Why, you use a known-good PSU if you have one and try to boot the system. If that fails, then you know that there is something else that has been affected by the surge. Your first culprit, of course, is the mainboard. Unless you have some real elite soldering skills, there is just about nothing you can do to recover a blown mainboard. Most users have to replace. I did not see anywhere that pepperman actually said get out your credit card, go down to the local computer store and start buying parts. He (and I) would both suggest that he borrow a PSU either from another computer or from a friend and test whether or not it is JUST the PSU that died.
October 1, 2009 6:14:21 PM

westom said:
You shotgun due to no hardware knowledge. You would have the OP replace perfectly good computer parts using wild speculation. You don't even know a surge existed. And yet you know a mythical surge must have damaged a power supply. Curioius that surge did not damage everything else in the house. Or do additional fact get in the way of wild speculation.

Meanwhile, in 30 seconds, the OP can obtain numbers. Next post then identifies the suspects. Why? Because one of us knows this stuff. Those who actually know something do not replace perfectly good parts until something works.

"I think it was a surge. Therefore it was a surge." OP only defined symtoms typical of a manufacturing defect. But somehow you know it must be a surge. After all, if you feel it is so, then it must be true. Also explains your posted mayhem.

OP can have his computer fixed the first time, faster, and with less money by simply posting numbers.


As Houndsteeth said, we're not suggesting he replace parts, we suggest that he test parts. As far as I know, it doesn't cost any money to swap out a PSU from another computer. It does, however, cost money to buy a voltmeter that he may not have. I will restate my recommendation: if he has a voltmeter and knows how to use, that's all well and good and I recommend he use it to diagnose his PSU; if he does not, but he does have access to a working PSU, I recommend he use that to diagnose his computer.

When you say that the "OP only defined symtoms typical of a manufacturing defect." are you saying that when something doesn't work, it has to be a manufacturing defect? Haven't high temps, dust, possible misuse, or (it could not possibly be this according to you) a power surge crossed your mind. Even if he has a surge protector, it might still not have protected it, as most surge protectors are rated at around 330 to 350 volts. That's nearly triple the standard 120 V that comes out of your wall.

As for our "mythical surge" not damaging anything else in the house, how would you have any way of knowing? Has the OP contacted you with a personal message saying "Nothing else in the house was damaged"? He hasn't posted that info.

Regardless of whether the computer malfunction's cause, he still needs to diagnose it. Whether it was a surge, hardware defect, etc., he still needs to know what still works and what doesn't. My original recommendation still stands: if he has a voltmeter and knows how to use, he should use it to diagnose his PSU; if he does not, but he does have access to a working PSU, he should use that to diagnose his computer.

BTW, since when does helping the OP involve attacking my knowledge of fixing computers?
October 2, 2009 12:10:33 AM

Have you guys noticed the OP has not once answered any questions? BTW he also stated from the start that he "thinks" he had a surge.


Just an observation. I'll go back to my couch now.
October 2, 2009 4:28:14 PM

Supe_ said:
BTW he also stated from the start that he "thinks" he had a surge.

He posted symptoms of a manufacturing defect; not of surge damage. That was defined long ago by the few who actually know how computers work. Those who shotgun could not notice. Shotgunners immediately replace perfectly good parts until something works. Never learn to what should have been obvious.

If the OP uses thirty seconds and a meter, then the next post will provide definitive answers.
!