Requiem For A Motherboard

JimLynch writes "In my last DIY column, I discussed what it was like to build my first system. As time went by, unfortunately, my DIY system wasn't all wine and roses. This column tells the story of how I destroyed my motherboard through a series of ill-planned and stupid actions. It should stand as a shining example of What Not to Do for DIYers everywhere."

Comment removed

[account_deleted]

Comment removed based on user account deletion

Bad times

[keybsnbits]

I sympathize with this man's problems that have to do with computer temps. It's very hard to get the hot air to flow out, instead of having a mish-mosh of air currents within your case. If fans wouldn't work for him, you could just go the hardcore road of water cooling... none of that "hot air" is involved, although a conventional water cooling system is much more expensive then fans. Its the ever debated balance between $$ and quality.

Hit Print to read this story

[cacheMan]

This and many other sites like it offer a Print option that puts the whole story on one page. With the likelyhood that slashdot is going to take this site to task, it would be a good idea to get it all on one page before you start reading. That way, you won't get blue mouse trying to get to page 2.

Re:worst article post in a while

[nostriluu]

I have to agree. I wonder if they'll post a story about me eating a bag of potato chips. Oops, I dropped one! Better get a page or two in about that.

I've been building my own PCs for the past 20 years. Along with my own, I help friends and have easily build more than 100 systems, plus about as many upgrades. I used to ritually buy a bottle on the way home from the parts store and get smashed while assembling. I've put cards in backwards, splashed solder, forced all kinds of parts the wrong way, worked way past midnight, rarely think about grounding myself, and only ever ruined one thing, a CPU I was trying to rig for a dual setup.

PC assembly is meant for amateurs, so I don't know how this guy managed to do the damage he did, but maybe his next system should come pre-assembled.

Fried memory

[Punk Walrus]

I lost a computer once after upgrading memory. I installed another SIMM and upon boot, zzzzzt*BAP! The CPU actually sparked and smoked. Unplugged system. The CPU was coated with some white powder. How the hell could I fry the CPU after a memory upgrade?

Later, I found that when I had put the memory in, one of the plastic pegs that separated the mobo from the metal case fell off and the half the mobo was touching the metal case. I am not sure which short circuited first, but... game over, man. Lost everything but the hard drives, CD-ROM, and floppy drive.

Keep organized

[xray_dude]

I find that having everything you might need next to you before you start building/reparing a system is the best way to avoid danger. Then ESD is not a problem as long as you periodically keep touching the case, as long as there is no potential difference between you and the case you should be fine. The more you move around the more chance of you building up charge.

It also helps to read manuals and how-tos if you are new to this stuff.

xray

Re:no doubt..

[AKAImBatman]

This guy wasn't just forcing it, he was handling it in the most slapdash manner possible! And how could he get a heatsink that barely fits the case? Where's all the air going to go?!

Bah. Here's what I learned about building a machine:

1. Buy as much as possible from the same retailer. That way when something goes wrong, there's no back and forth on it.

2. Buy from a store. There are a lot of "cheap" internet sites that will happily sell you unreliable hardware, then become hard to contact afterwards. Swap meets are an especially bad place to purchase new hardware components. With a store, you can walk through the door and strangle the guy behind the counter.

3. Buy as big of a case as you possibly can. This will allow you a lot of room to work on the inside, as well as good airflow and extra mounts.

4. Find out what every cable is *before* you plug it in. Also, make sure which direction it goes. Sometimes they need forcing, but only force after you're SURE that it's supposed to fit that way.

5. Take your time and assemble the components as early as possible. Some things can only be inserted inside the case, but others (such as CPU, fan, and DIMMs) can be assembled outside the case. It also never hurts to leave things like the hard drive unplugged just to make sure your system turns on and functions. Remember, SLOWLY.

6. Buy quality components. It may look cheaper to buy that AZUZ motherboard instead of the ASUS one, but the difference is tremendous.

One last tip: don't buy the latest and greatest processor unless you absolutely have a reason to do so. The performance difference between that and the next model down is almost imperceptible due to wait states in the CPU. You're much better off investing your money into more RAM. Less heat, more speed. For gaming, go for the best vid card, though. Unless you like to upgrade, you'll be with it for a long time.

Why?

[WhatAmIDoingHere]

So this guy who doesn't know what he's doing decides that he can tell others what to do?

The clueless leading the clueless.

I mean, it would be expected from somebody at THG..

Actually, now that I think about it, THG will probably post these article soon enough.

Dumbest

[Apreche]

I've been building many a PC in my day. The first one I ever did is still humming away in the other room with my blog running on it. The BX chipset is by far the most reliable and stable motherboard chipset ever made. And the Abit BX6r2.0 is the best board with that chipset.

Anyway, I've never had a problem putting computers together. The reason for that is lots of research on the internet. Before I get ready to build a box I check everything. I never buy incompatible parts. I don't skimp on cost and risk getting a part by a no-name manufacturer. You don't know how many times I see people with broken hardware from no-names. Pay the extra 30 bucks and get the big name brand stuff. Abit, Asus, Leadtek, Gainward, Creative, Seagate, Corsair, Crucial, etc. If you get a video card from randomtaiwantech and it doesn't work, there's a reason.

However, me and my roomate did make a big mistake once. The reason was that there was no documentation concerning the issue on the net, and to this day, there still isn't. The first time I built a computer with a Duron isntead of a Pentium it wouldn't boot. I couldn't figure it out. The company I bought the computer from either didn't know. But what they should have noticed on my invoice was that I had a 300W power supply and that I needed more. Eventually after several RMAs I had a 300W power supply that worked somehow magically.

Later my roomate got a new PC and it had the same trouble. My computer died soon after and we realized something. The power supply is important and 300W isn't enough anymore. Motherboard manufacturers! In the documentation for a motherboard list how big a power supply is needed! You have no idea how long it took us to figure out what was wrong with several completel seperate machines not booting in the same fashion.

Let this be a lesson.
1) do research
2) do more research than is possible
3) don't be cheap
4) if you know what you're doing it wont go wrong.

Needless to say my current box has a huge expensive 450+ Watt Enermax PSU. I will never have THAT problem again.

Re:My guess

[vrt3]

On many sites including that one, it helps a great deal if you click 'Print this page' or something to that effect and read that instead. In this case: http://www.extremetech.com/print_article/0,1583,a= 130946,00.asp [extremetech.com].

Re:no doubt..

[AKAImBatman]

Wait states have not much to do with the CPU... good advice, but a bad explanation.

True, but it is the CPU waiting. The wait states have more to do with cache misses, bus speed, and memory latency. The end result is that your CPU rarely his 100% capacity, thus real world performance differences between CPUs is somewhat mythical.

Then again, I didn't really feel like giving a dissertation on why CPU speed doesn't matter. Getting a good MoBo is about all you can do to really improve the issue. Since "quality components" was already in my advice, I figured that anyone reading it would do just fine. :-)

Wow - BX chipset, you must have YEARS of exp.

[Anonymous Coward]

If the first computer you ever built had a BX chipset - then you couldn't have been "building many a PC" for too long.

Re:no doubt..

[mrchaotica]

You could always compile again and diff the binaries and find the bug that way, right?

Re:Let's see...

[Anonymous Coward]

I've built literally hundreds of computers, all working fine. Using a grounding strap is quite over-rated. As long as you aren't wearing nylon or wool clothing, simply touching a metal part of the case will be fine. In fact, I've never seen any component fried because of static when being removed/installed.

Just because a component still works after being exposed to static electricity doesn't mean no damage has been done. This page [circuitsassembly.com] has a picture of the damage that static electricity does to the surface of a silicon die.

Sometimes the static charge will only take a bite out part of a track instead of the whole track. While it will still work the life of the component has been shortened. The narrowing of the track will lead to heat build up at that point which can eventually lead to failure. As static damage is cumulative the component has a greater chance of failing the next time it receives a minor zap.

A static strap only costs a few dollars and only takes a few seconds to put on.

Re:no doubt..

[silicon not in the v]

Our company makes memory chips. 80C is considered standard operating temperature. We do high voltage operating stress on the parts to eliminate infant fails at 130C before they're re-tested and sold. I'm not sure how tolerant other types of IC chips are, but 50 or 60 doesn't seem too bad yet. However, if that is 50 or 60 air temp in your case, your components are getting hotter than that, and will probably have problems.

Baby steps, Bob.

[C10H14N2]

I started building out computers 20 years ago when components were so #$%ing huge and cast-iron that you'd have to be a total idiot to accidentally destroy anything. By the time a faulty heat-sink could destroy a system, I had built about a hundred of them, so I've yet to destroy a single machine by my own stupidity. Knock on wood, but there is something to be said for tinkering with something you don't mind losing before dropping $2k on a soon-to-be door-stop.

Definitely, though, the huge case advice should be heeded. Sooooo many problems can arise from that, not the least of which is, ta-dah, heat dissipation. To say nothing of proper (and improper) grounding. Finally, DON'T SKIMP! You're saving money on the labor, put it into quality parts -- especially cables (power/hdd/cat5 etc). There's no sense shorting out a $200 HD or $95 power supply, and potentially a $200 MB and $400 CPU, all over a 39-cent cable that you should have spent a buck on -- (or for that matter, the $20 power supply you should have spent $50 on)... and for the love of god, don't build this thing on the floor of your shag-carpeted living room walking around in new trainers, OKAY!??!

Re:Let's see...

[Halfbaked Plan]

Components aren't necessarily 'fried' by static. More often, mild 'latent static damage' occurs. A few pins on random chips within the circuit suffer partial damage. They become 'leaky' and inputs draw more current. The system slowly deteriorates from random mysterious problems.

Hot dogs who obsolete everything at six month intervals might not notice this sort of damage. The rest of us would.

Re:Boy

[silas_moeckel]

Ah the reason why programmers should not be allowed near hardware. Thats a general rule. Programmers with degrees in EE can touch hardware otherwise they dont have a clue.

The other reason is if programmers are allowed near hardware they will start to blame hardware for there issues.

Re:Dumbest thing I ever did to hardware...

[cr0sh]

Want to know what the buzz likely was? A dry bearing on the fan. Vibration sets up, "finds" the resonant frequency, and "buzzes". When you stop it, and let it start again - it is OK, for a little while. Sometimes, if you let it go on long enough, it will drop out of the frequency for a little while, but then sooner or later, it will be back.

The best way to fix it is to get a new fan, and install it into the power supply (don't just buy a new PS unless you need it for a good reason, a fan will be a lot cheaper).

Barring this, remove the fan, remove the label from the fan, and pop the retainer ring (don't loose it!) off the fan shaft. Remove the fan blade/magnet assembly, apply a drop of silicone oil (or a bit of petroleum jelly) to the shaft, reinsert it into the bearing (spin the fan to distribute the oil on the shaft), and reattach the retainer ring. Clean the area for the sticker (of any grease or oil), and put on a new one (a cheap paper "garage-sale price" label works good).

Such a simple repair takes only minutes, and will last for several months, if not longer (why spend the money if you don't have to?).

This guy's problem...

[cr0sh]

...was simple - a lack of patience. Indeed, throughout life one learns that most mistakes are caused by not taking time (to think things through, to actually do them, to looking at causes, etc).

This is as true today as it was 50 years ago. Furthermore, I think if you can't name "major" components on a motherboard, you shouldn't be messing with it - or at the very least, you should know this and *really* take your time.

I remember upgrading the memory, as a kid, in my TRS-80 Color Computer 3, from 128K to a whopping 512K. This was about 15 years ago. I remember the instructions (which I still have, along with the computer, and upgrade - and yes, it still runs great!) warning about handling the CMOS devices to avoid static electricity (when inserting each of the DRAM chips into their sockets). I ended up grounding myself using a length of steel wire tied to the kitchen faucet, then looped around my arm as I did the upgrade.

All in all, it took me about an hour to perform that first true "upgrade" on my Color Computer - being a 15 year old kid, impatient to get my upgrade going, but knowing that if I screwed up, my parents would be pissed (they paid for it, after all) - I took my time, grounded myself, and made sure I did everything right. So what do I have to show for it?

Well, patience, number one - but I also can still whip out my Color Computer 3, with floppy drive, monitor, and 512K of RAM - and boot it into OS-9 (8 bit multiuser/multitasking, baby!)...

Re:no doubt..

[WuphonsReach]

But I've seen everything from "60c isn't bad" to "60c is one step below the entire computer bursting into flames".

Different CPUs are capable of different temperatures. (case in point: Intell Prescott chips being derided as "Pres*hot*" chips)

Internal case temp should be in the 30-45C range (assuming ambient air temp of around 25C). My AMD cases are running 41-46C at the moment, but the A/C is off and the ambient temp in here is 31C. CPU temps are generally in the 50s, depending on the case temp and the particular chip. I only use AMD, but I get nervous when the chip hits the high 50s. At which point, I investigate larger heatsinks or higher cfm fans.

Only solutions for lowering internal case temps are either:

- Remove heat-generating components

- Get components that produce less heat (5400rpm drives instead of 7200rpm drives, older video cards instead of the dual-heater top-of-the-line beast, use an older and cooler CPU)

- Adjust/add fans to move more air through the case per minute (air flow). Make sure the exhaust fans are properly oriented so that air flows through the case as designed.

- Simply buy a larger case so that the heat producing components are farther away from each other (Antec Sonata / p160 or a full-tower case)

- If the video card has a fan on it, make sure there is at least one empty slot between it and the closest PCI card

- 7200+ rpm drives generally require active cooling (Antec p160/Sonata cases have drive bays with a dedicated 120mm fan slot). Putting a 7200rpm drive in an external USB/firewire enclosure that doesn't have a fan is a good way to kill the drive (been there, done that, now only use 5400rpm drives in those enclosures).

I tend to be conservative with my cooling advice because my office has poor climate control. Like I said, it's 31C in here at the moment, which is warm enough to be uncomfortable even in shorts and a t-shirt. However, all of my machines work just fine since that they're in good cases with good airflow.

Re:The article HAS to be a joke

[JimLynch]

Actually I'm the Community Manager over at ET. I manage the ET forum. Unlike Loyd, Dave or Jason, I'm not a hardware guru. I write software reviews and opinion pieces. So DIY is new to me but I'm learning fast and having a hell of a time along the way, as you can tell from the column. ;-) Please do drop by the forum and say hello sometime.

Re:imacs don't hold beer..

[bsartist]

I've had good luck with putting keyboards in the top rack of the dishwasher. Just make sure to let them dry *completely* before you try to use them again.