IT Employment

How to troubleshoot a computer problem

What steps do you go through in order to diagnose a computer problem? There's more than one correct answer, of course, but here's one way to work through the troubleshooting thought-process.

What steps do you go through in order to diagnose a computer problem? There's more than one correct answer, of course, but here's one way to work through the troubleshooting thought process.

-------------------------------------------------------------------------------------------------------------------

We all have our methods to identify and correct computer problems. Most of us probably have certain things we go through, or steps we follow, and so on. I'm not suggesting that there's only one right way to diagnose and fix a problem, but they probably all have some commonalities. Moreover, different environments, applications, and countless other circumstances will influence any troubleshooting methods, so your particular circumstances might dictate something otherwise. Nonetheless, here's an outline of my own thought process when approaching a computer problem.

User support problems usually fall into one of the following categories: Hardware failure: This could actually be a failed component -- a bad motherboard or hard drive, for example - or it could be a matter of an unplugged cable, a router or modem that needs to be recycled, or some other such thing. Software issue: This would include the operating system, something affected by a Windows update, an application update, or an actual software bug. User-created problem: These are usually of the unintentional variety where a user might inadvertently disable something, change something, delete something, and so on. Training or documentation: Often times there is no real problem, per se, except for a lack of knowledge or understanding about something. In these cases, it might be that training is needed or documentation needs to be updated. Outside vendor issue: As much as I hate hearing the buck being passed around from one support group to another, it's sometimes real and justified. A custom software application or a supported piece of hardware might be experiencing some issues. High-end and high-output printing devices often fall into these categories. In my case, the support obligation usually ends at the printing or plotting device itself (except for low-end desktop printers), and short of restarting the unit, there isn't much I can do to troubleshoot and fix the problem. Such units are seldom sold alone, and they almost always come with a use-and-support agreement.

One thing I always do, however, when calling on such outside support help is to make myself available to assist in whatever way necessary. I would never say (or seldom say), for example, this is your problem that you need to fix. But rather I'd approach it as a joint problem that we, together, need to address until the issue is resolved.

I try to identify into which category the problem falls as quickly as I can. It is possible, I suppose, for a problem to actually have two (or more) underlying reasons, which can make finding the solution even more of a challenge; and in such cases, the fixes would probably have to be addressed in that order -- hardware first, then software (drivers, etc.), then user issues (configuration, etc.), followed by training or documentation. Nonetheless, I try to isolate the problem into one of those categories as quickly as possible.

Possible steps to seek a solution: Reboot: Try this preferably before support is called. I actually had a user who posted a message to himself that he taped along the edge of his monitor: Reboot first, call Joe second. Of course, on one hand, there's the situation where a reboot actually fixes a problem. On the other hand, however, requiring a reboot might really be a symptom of a deeper underlying problem. Rebooting once and being done with it is one thing, but if it becomes necessary to reboot too frequently to fix the same problem, then digging a little deeper to find the underlying problem would be required. Replicate the problem or issue: Make the problem happen again, see the error message, etc. Experience the problem yourself to get a better feel. Retrace user steps that led up to the problem: What did the user do immediately leading up to the problem? What was changed? What was installed (or uninstalled)? The answer is usually, “I did nothing different,” or something along those lines. But it's amazing how often we actually can identify that something really was changed by simply asking questions of the user. Device Manager: This is such a quick and easy way to get an overview of the hardware status. A quick look for the yellow or red warning tags might quickly identify the source of a problem. Error logs: Like the Device Manager, the error logs might reveal the source and frequency of a problem. Isolate the problem: Is it unique to this one computer, or are more people experiencing the same issue? For example, is it only one computer that can't print versus all of them that can't print? Seek obvious solutions: Unplugged cords and cables are often the cause of a problem or failure. Sometimes the most simple and obvious solution is the hardest to see. I recall having received a support call telling me that a computer hard drive had failed, but upon further investigation I discovered that a prankster had put a floppy disk into another user's floppy disk drive, and since the BIOS boot order was set as such, it was looking for a bootable floppy disk. Since floppy disks are used so seldom any more (in fact, many computers don't even have floppy disk drives), it was an easy thing for the user to overlook.

There's no one right way to troubleshoot all computer problems, and none of my outlined steps are necessarily in any correct order. But this is pretty much my own mind-set when I approach problems in my environments. Please share your own thoughts and comments.

63 comments
johnsondell
johnsondell

I was also facing same issue, I tried a lot to fix this however I was not able to do that. At last I just visited http://computervirusremoval.us.com/ and made a call on their toll free. I got instant help for my computer by a certified technician.

raywardme
raywardme

Whenever I have a problem with My PC specially in hardware, I always check the power supply first! and if I dont see any problem of it then I will go to Computer repair technician. Its so important to invest in a devices that can help you trace a problem with your PC.

TBBrick
TBBrick

For the most part, my users give me enough info to figure out if it's a network, hardware, or software issue. Even so, there's nothing like being able to put the ole eye-ball on the situation and fix it toot-sweet. Esp. for those at one of the eight sites. They love not having to wait on a tech to come out if I can solve it then and there.

bazza20
bazza20

I'd like to add an additional step - having fixed the original problem, check again for any malfunction. In my experience, multiple faults occur more often than you would expect. Often the faults seem completely unrelated. And yes, malware is a significant factor, especially when the problem is "My computer is sooo slow nowadays".

g3po2
g3po2

In my 33 plus years of experience troubleshooting computer problems, I have found that the Number One cause of hardware failures is the power supply. So invest in a $10 to $20 Pc power supply tester, and check it 1st. While you have the top popped, make sure that all the fans are working and that there isn't a large accumulation of dust that would prevent proper air circulation inside the unit. The 2nd most likely hardware item to fail is the hard drive(s). 3rd is memory chips, 4th the motherboard, and 5th--possibly some peripheral device. So detach all unnecessary items from the motherboard (CD-ROM drives, NIC cards, dial-up modems, etc.) If there is no online video on the motherboard, or it has been disabled, put in a cheapy PCI video card after removing any other video card and disabling the "on board" video in the CMOS. If all that doesn't show a problem, you can suspect malware is causing the problem. I've worked on computers that had as many as 3,000 unique malware programs!!! I have yet to find one single program that provides complete repair and/or future protection, so I use a combination of nearly 20 different products, mostly share-ware or free-ware, which seems to work very well if you can get your customer to do the updates. (Good luck on that last one!) Aloha from Hawaii, Glenn

Michael Jay
Michael Jay

the customer, if you know them well then you have a very large edge. If you do not know them it is time to get there. One of the things I love so much about fixing computers is there are few real repete problems, each issue is different. It could be a corrupt file, it could be a memory problem, it could be a user error. You must be on site to really tell just what is going on and no, there is no script. You look and analyze and find your starting point, from there you discover your first move. I just love this stuff. But I do not think there is any way to "troubleshoot a computer" each event has it's own troubleshooting plan which is laid out by what you find when you get there and see the problem first hand. Seeing the problem on a call ticket or talking on the phone just cannot give the info you will get by being there.

NickNielsen
NickNielsen

This should be the first check you make for most "won't boot" problems. I once spent 2 hours troubleshooting a server boot issue that was corrected when level 2 asked if there was a disk in the normally unused floppy drive.

Ravnor
Ravnor

Although it may not actually be part of the Hippocratic oath, I try to follow the maxim "First, do no harm". Unless it's a total meltdown, I back up anything that I may change so that I can at least return the system to the state where I started if my repair attempt fails. I also try to break the problem into separate components and test each component separately.

nospam.online
nospam.online

From my experience, in military and civilian work, Eliminate the user as being the source of the problem first by questions. We've all read the jokes about user error's with computers but know the truth is a lot of error's go back to user's having done something they were not to do or not paying attention to what they are doing and making changes that effected the problem if not outright caused it. Example, recently a teacher came to me complaining that some of the headset?s in the computer lab didn?t work, she had her aid remove all the ??failed ones?? (untested by me) and put them in a pile for me to replace. I tested them all on one computer and all worked. Had her aid reinstall them and list out the computers that didn?t work. Went in and personally tested each one and all of them had the same problem, when installing the aid turned the volume control (in the cable between headset and pc) down to prevent harming the kids hearing. Example 2: Recently got a report from a teacher that some of her kids couldn?t take a on computers to do a network test with, they complaint that 2 set?s of computers would not come on but students insisted they didn?t mess with them. On inspection it was found that the power strip had been unplugged from wall outlet and then plugged into itself. When it was plugged back into wall outlet they all worked fine, it was noted that while powering up they reported failed shut down. Example 3: Recently we had a 100% shut down of the server rack over a weekend all due to a guy decided to re-wax the main hallway of the school but in the process he popped the breaker for the main circuit the server UPS?s are connected to. But didn?t bother to tell anyone that the outlet or breaker quit working so the power could be restored and yet claims he didn?t hear the squalk of the 4 UPS?s telling that the power had gone out. After dropping to 5% power the UPS?s shut down all 5 servers. Took about 5 hours to charge back up and bring servers back online. Equipment fails, fact of life and that?s job security for some of us but there?s also the amount of work some if us have to do because user?s do things they should not. After that, it;s a combination of diagnostics and test's to find the failure and fix it. Normally for a PC I use a PCI testing card (kind of like the old Kickstart ones) and some software.

Joe_R
Joe_R

There's a big difference.

Joe_R
Joe_R

Multiple faults do indeed happen frequently. Thanks.....

Joe_R
Joe_R

Thanks - and Aloha back at ya'

pdr5407
pdr5407

What do you see most often that causes a power supply to go bad? I just replaced one today on a desktop that would not boot into Windows. After replacing the power supply and fixing the boot files it still had problems, and Windows had to be re-installed. But I isolated the problem to a flakey psu, and it is running better now.

Joe_R
Joe_R

You're right, there is no one way to troubleshoot a computer problem. That's why it's more of a midset. Thanks for sharing.

Joe_R
Joe_R

Thaty should be the first thing to check for most [i]won't boot[/i] problems. Thanks.....

santeewelding
santeewelding

That I cut you guys out of the loop by learning as much and as fast as I can about these things. Not there, yet; but, gaining.

knudson
knudson

First yes, Backup any user data. Then (unless a hardware issue, but even then too really) WHAT CHANGED !! And you may need to POLITELY pry it out of the user. They may have installed, changed something they shouldn't have, or done something without realizing it was a change or could have an effect. Start there, then normal trouble shooting. Had a almost impolite dealing with our cable companies help desk. They installed a new modem/router at the neighbors. I do tech support for him. No wireless devices would connect. The Help desk repeatedly told me to contact the laptopS (his 2 and mine) and desktops wireless NIC Mfg. I kept saying, it worked yesterday, you changed the router and it doesn't work. After too many iterations, including some email support. One tech, who did toe the company line, gave me a few pieces of information. End result, for some reason the router would not hand out IPs except after a power cycle ??? I sent the company a detailed report on what had happened, what they did, what I did, and the end result. I stressed Always review WHAT CHANGED !!! FYI they replaced his router a couple of days later, they even called him to setup the appointment. No issue's since.

matthew.balthrop
matthew.balthrop

always demonstrates a certain professionalism in ones work. Time constraints sometimes play a factor, but the ability to isolate a problem down to it's exact or near exact cause will earn the respect of your clients. I am applying this more towards a hardware perspective, and i know there have been many times when troubleshooting a failing computer that the ability to isolate a problem down to it's exact failing component/s has in the long run saved time and money along with bolstering a client's perspective of the technicians expertise.

bazza20
bazza20

Today I was called on site to replace a failed 5.25" FDD. Remember those? The user urgently needed to recover some data from an old disk. So I hunted around, found an old 5.25" drive, tested it on an old computer and fronted up ready for action, congratulating myself on my versatility. But first, I insisted on testing the faulty drive for myself. Worked perfectly. When asked to demonstrate the problem, the user inserted the disk sideways. "Oh," he said, "I thought I had put the disk in every possible way." I justified the hefty call out fee and hourly charge by installing a second hand hard disk, reinstalling DOS (yes, it was one of those), sorting out his configuration and menu files, and generally leaving the machine running nicely. Yes, this client has a very nice Toshiba laptop running Windows XP, but claims DOS can do some things better.

Betageek52
Betageek52

in certain situations I LOVE my PCI testing card....

Joe_R
Joe_R

Thanks for sharing.

Joe_R
Joe_R

.....on your own blog site, you should first do the following: 1. Ask for permission. 2. Give credit to the original writer. 3. Provide a trackback link to the original article. You did none of those things. Plagiarism is frowned upon by most people.

g3po2
g3po2

In my experience, I have found a number of reasons that I suspect cause power supplies to fail. 1.) The power supply can't produce enough power. Note that most "custom built" computers usually have at least a 450 watt power supply. I've seen some cheapy, over-the-shelf computers that have a 90 to 200 watt power supply. Problem is that if you start adding upgrades to the machine (a DVD-RW, for instance, you may not have enough power. 2.) Bad fans or dirty vents. If the fan isn't turning, even most power supply testers won't tell you. I've seen machines which have built-up a half-inch thick layer of dirt in a very short period of time (like three months or less.) No matter how good the fan, you still need that air circulation. 3.) Many years ago I read a report, I think it was a white paper put out by IBM, that the "average" number of severe power anomalies occur at the rate of 1,500 per day. Most surge protectors only "kick-in" when the power exceeds 330 volts A.C. Since A.C. is essentially a sine wave, the 120 VAC rating is actually R.M.S., which means, essentially, the mean average of electricity "under or above" the curve. 120 Volts A.C. will hit 177 positive and 177 negative during a normal sine-wave cycle. So if te power gliche occurs when the A.C. current is at 177 volts, you end up with a surge protector that won't "kick in" until 507 Volts is reached. Them's a lot of volts! Some high-end surge protector manufacturers off wave-tracking units, that will follow the wave and kick in anytime the voltage exceeds 330 volts, or whatever they set as their rated kick-in voltage. Power conditioners monitor power, and readjust it. So if you feed 90 V.A.C. into the unit, you still get back 120 V.A.C., with the sine-wave kept intact. Finally, there is the U.P.S., or uninterpretable power supply. Depending upon the make and model, they have various kick-in voltages. The most expensive of the bunch (made by "Best"), using multiple technologies to achieve this, the primary one being a large core transformer, which completely isolates connected equipment from the actual power source. This unit are very heavy to lift, because the transformers are huge. In addition, most U.P.S. models have circuitry that will monitor either the positive side of A.C. current, or in the better models, both negative and positive (sometimes referred to as a "double-conversion" unit. Hope that helps, Aloha from Hawaii, Glenn

Michael Jay
Michael Jay

so correct Santee, it is called CUSTOMER SERVICE, and called so for a reason.

Zenith545
Zenith545

We will also be pleased when "you guys" quit messing with our machines, trying to improve computer processes with your limited knowledge and overall just making things go FUBAR. Just sign the paychecks, please. LOL - in jest

Michael Jay
Michael Jay

finally unplugged that USB drive and the thing boots now. Cool.

Joe_R
Joe_R

To the [i]what changed[/i] question. It's amazing how often [i]nothing[/i] actually turns into [i]something[/i].

nospam.online
nospam.online

OK, I have to add this... I'm off campus looking at a place to buy and get three frantic calls that one of the east hallway printers quit printing. I drive back to campus (about 49 miles) I stop by my office, log into one of my office machines to check Cisco's and see if any error log...nothing. Pull up my printer folder and it's showing the printer as being "offline" so I walk over and sure enough, someone has clicked the "offline" button. I click it back online and it starts the warm up cycle, couple seconds later starts printing, prints off 7 or 8 jobs of teachers work...20 to 30 kids each... Sometimes I really wish I could get paid for doing this...

Joe_R
Joe_R

Thanks.......

santeewelding
santeewelding

You don't have to sugar-coat it. It is also requisite in computing the path of a molecule at STP. Maybe, though, those who crow about having everything about IT under tow will have problems understanding you.

NickNielsen
NickNielsen

Bloody, bleeding, swollen, fat, pooped, popped, and obese have all modified "capacitor" on the failure tags. :D

Joe_R
Joe_R

That's some usefull information.

g3po2
g3po2

I totally agree with Nick Nielsen. One of the reasons there are strongly worded warning labels on a PSU is because of the electrolytic capacitors, which can hold and maintain a large amount of D.C. current even though the PSU has been unplugged from power and completely removed from the system. One of the dirty tricks I heard of in electronics class was to put a full charge into a large(D-size battery size), then touch it to some unwitting student's wrist. The shock is brief, but quite grueling, especially if the cap is from an old-fashioned, vacuum-tube T.V. or radio, which ran at much higher voltages. If the read the find print on these devices, most are rated for a maximum voltage of 600 VDC. So a good size surge is enough to max one split open. In the old days, when PSU's were very expensive, and when I had a de-soldering station, I replaced many caps. Now, with the cost of labor versus the cost of a new PSU, it is difficult to justify the time to do almost any board-level repair, and that extends to motherboards. While over-heating is frequently the cause, I would guess that most PSU failures are the result of lousy A.C. power input. Sorry to say so little in too many words... Aloha from Hawaii, Glenn

Michael Jay
Michael Jay

commodify to treat as or make into a mere commodity to be bought and sold or to be used in selling something else What? I hug people is that ok?

boxfiddler
boxfiddler

Customer is only service when the customer is not a commodity. I think you don't commodify (?! :D ) people, Michael Jay. etu

santeewelding
santeewelding

With increasingly knowing and gaining eye. Fast-approaching is the moment when I ask you, as well as the entirety of your knowledge, how is it that it is and that you are? Falter just once -- just a little bit -- and I fire your ass.

santeewelding
santeewelding

You look to me, or to any other, is the moment you become a drone.

Michael Jay
Michael Jay

I was actually worried for a bit. My esteem is restored.

santeewelding
santeewelding

Mandriva, as I recall. $60 lesson. I also get to marvel, Michael Jay, at those who go about this unlike the so many drones I tire of hearing from. You are not a drone.

Joe_R
Joe_R

....that people overlook. Thanks.....

Zenith545
Zenith545

If you don't get paid for it, I don't understand why you drove the 49 miles back to campus. Ever heard of remote computing??

matthew.balthrop
matthew.balthrop

how many times we have had a user call about not being able to sinc their laptop to our server, while the ethernet cable is dangling next to them not plugged in. Or that they cannot log into their laptop because NumLock was on :(

Editor's Picks