| Date: | July 20, 2007 / year-entry #264 |
| Tags: | other |
| Orig Link: | https://blogs.msdn.microsoft.com/oldnewthing/20070720-00/?p=25923 |
| Comments: | 43 |
| Summary: | I had removed the cover from one of the computers in my office in order to upgrade one of its hard drives (from 20GB to 200GB, woo-hoo). The hard drives are kept in a removable cage, so first I had to unscrew the cage, then unscrew the drive from the cage, then swap in the... |
I had removed the cover from one of the computers in my office in order to upgrade one of its hard drives (from 20GB to 200GB, woo-hoo). The hard drives are kept in a removable cage, so first I had to unscrew the cage, then unscrew the drive from the cage, then swap in the new drive. Of course, you don't screw everything back in until you've tested it out, so I had the computer running with its innards strewn about my floor until it ran to my satisfaction. Okay, time to put everything back together. I screwed the drives into the cage, screwed the cage into the case, but before I screwed the cover back onto the case, I turned on the computer just to make sure everything was still okay. A colleague of mine happened to stop by as I was doing all this to discuss a technical matter, and we chatted about the problem while I sat on the floor with a screwdriver. I said to my colleague right before I turned the computer on, "Look, I bet it won't work." Lo and behold, the computer didn't work. It just made a horrible beeping sound. Screwing the computer parts back together is the most dangerous step in computer assembly because once you do that, there's a pretty good chance that something will stop working. I spent the next fifteen minutes re-disassembling the computer, removing and re-attaching every cable that might have wiggled loose, all to no avail. Eventually, I found the loose connection: While mashing the cables around, one of them accidentally pushed against one of the RAM release levers on the motherboard. As a result, one of the RAM sticks was not fully-seated. Push the RAM back into its socket, power up the machine, everything works again. |
Comments (43)
Comments are closed. |
Sometimes this is literally the problem.
I’ve seen one member of my family nearly destroy a hard drive by using a screw that was too long. It pressed against the circuit board and cracked it, breaking a trace.
One family member and one friend have shorted out their motherboards by using screws with too large a head. Traces were grounded, one board was destroyed the other fine once the problem was discovered.
I’ve had three computers of mine crash and lock up intermittently, problems which were eventually traced to screws. Two to grounded traces caused by tightening the motherboard down before all screws were in. Poor placement tolerances on the mounting posts left one too near traces and heating or vibration would cause intermittent contact. The final one was due to a lose screw that fell and lodged beneath the motherboard. It didn’t rattle, it didn’t fall out, but it did occassionally short with vibration or heat. Six months I spent cussing Windows 95 before I found the screw and random errors went away.
Finally, I have seen more CD-ROM problems than I can count, and a few floppy drive ones, caused by mounting screws being tightened too far in. Doors not opening, scratching noises, all sorts of stuff that disappeared with different screws or a slight loosening.
Mr. Shiny & New: That sounds like it may be a grounding problem, or something like that. It may be that inside the case, one of the pins into the motherboard (or another card, but it’s usually the motherboard that this kind of thing happens to) is touching the metal in the case, which is hooked up to ground through the motherboard screws. Outside the case, of course there’s no grounded metal.
(On the related law: Yeah, I’ve seen that too. It’s even annoying for the expert, because now they don’t know what exactly was wrong.)
15 minutes? You realise that the beeps were trying to tell you roughly where the fault was, don’t you?
Could be worse. Once, after installing a new hard disk, I put the case back on a PC. Only I managed to put one of the case screws through one of the 12v connectors from the PSU. There wasn’t very much in the case that worked after that.
Try inserting the RAM while it’s powered on, it might produce a nice smell :)
So do all of the programmers service their own hardware in your neck of the woods, Raymond? Or are you a special exception?
Maybe, maybe not… if you don’t have a manual handy, how can you interpret them anyway?
I’ve had this happen to me WAY too many times, soooo frustrating.
I recall a wonderful poster of Murphy’s Laws, which included the following gems (quoting from memory):
The part that has failed will be behind a panel held in by sixteen screws, and you will strip the head of at least one screw when removing the panel.
After you have replaced a panel held in by sixteen screws and tightened all the screws, you will find that you have forgotten a small but significant part.
I recall a wonderful poster of Murphy’s Laws, which included the following gems (quoting from memory):
The part that has failed will be behind a panel held in by sixteen screws, and you will strip the head of at least one screw when removing the panel.
After you have replaced a panel held in by sixteen screws and tightened all the screws, you will find that you have forgotten a small but significant part.
The only thing more dangerous than a software guy messing with hardware is a hardware guy messing with software.
Another war story was a buddy of mine ended up doing something very similar–only he somehow accomplished the feat of bending the tension pin connectors in the memory socket out of place at the same time…
Suffice it to say he said it took him nearly an hour to finally figure out what was wrong, another half hour to get over the "oh my god, what have I done–did I just destroy my machine what will I do now oh god its 11:30 at night, and I need somebody to help me figure out what to do" moment, and then about an hour and a half after that patiently trying to maneuver the pins back in to place with a needle. He said he was finally about to make the repair and the machine booted and worked just fine afterwords…
Neal:
Your point is exactly why I advise people to also purchase a stand alone packet of mounting screws whenever they buy something that involves opening the case. (e.g.the Fry’s "pre packaged" baggies that contain the plastic motherboard mounts, metal mobo mounts, and a bunch of the properly threaded screws, generally designed for folk who are building a PC out of components)….
Yes more often than not this is a bit of overkill, especially for an upgrade, but it avoids the alternative "feats of creative engineering" when you don’t have the proper parts…
How many programmers does it take to screw in a light bulb?
.
.
.
.
Can’t be done: it’s a hardware problem.
Neal:
Yep, I did exactly that same thing while reassembling a friend’s laptop that had failed while visiting me on vacation. Fixed the LCD display ribbon cable that came loose. Was feeling all proud and geeky and promptly drove a too-long screw into the case and through the hard drive logic board.
Oops. Hubris leads to downfall.
[What does "loud continuous beep" mean? (Note: No manual.) -Raymond]
Who needs a manual when you’ve got the Internet (and another computer that can go online when the first computer is dead, obviously)?
Searching online for POST beep codes I’m guessing you have either an IBM or a Compaq. The long continuous beep on the Compaq BIOS indicates bad RAM, in any case.
I’m sure there must be plenty of computers in a Microsoft office building. Though there might not be as many people willing to give up their computer for a minute to someone who just broke his own. ;)
Ben,
Similarly I once pulled a Voodoo 2 video card from a turned-on computer and replaced it with a Voodoo 3. The computer was locked hard, but surprisingly all the hardware worked once I had the mess sorted out, even the expensive Voodoo 3 I plugged in ‘live.’
Man, I miss 3DFX. Those carts were hot stuff back then.
From the "words to live by" collection:
"Beware of programmers bearing screwdrivers."
also,
"There’s nothing more dangerous than a bored engineer."
15 minutes to find the loose connection isn’t too bad. My next-to-most-embarrasing one is a full day to find a bad power cord. After swapping literally everything else I sat down and asked myself, what haven’t I swapped? Answer – the power cord.
"It’s always in the last place you look."
The next time a power cord failed on one of my computers it only took me half a day to zero-in on it :-)
Basic computer beeps to never forget:
Single beep – Boot OK!
Continuous Beep at steady interval and tone- RAM problem
One long beep followed by 3 short beeps of same tone – Video Card
High beep followed by low beep – Bad BIOS
No Beep – Vague Mobo problem or no / bad CPU
I’d say 99% of the beeps that I’ve heard as a service tech fall into those categories.
Blood helps. If, at some point during the process, you cut yourself badly enough to bleed, that greatly improves the odds that your newly-reconfigured hardware will work when you put it all back together.
No beep can also mean you’re using a Shuttle mini-PC. I think I tested nearly the complete power set of parts before finding the one bad RAM stick.
I can totally relate – a new piece of hardware is likely to fail if properly screwed into place when first installed. If the computer case is also close and screwed in then failure is almost guaranteed.
I think the first time I learned that rule was back in the days of modems with jumpers :)
My system doesn’t beep at me. It talks to me, in a monotone female voice via the PC speakers.
It drives me up the wall to hear "System completed power-on self test. Computer now booting from operating system." on every reboot. Still, it’s infinitely more understandable than a series of beeps.
Having worked in computer service for several years I can totally relate to this problem. It’s really almost a Law of Nature: When you put the case back on things stop working*. This is guaranteed if you didn’t test it before putting the case on. Heck, multiple times I saw computers that worked properly if they were totally disassembled, but when you put them into the case they stopped working. Take them out, they work fine, put them in, no go.
* A related Law is the Law of things not working until an expert arrives, who then does exactly what the novice did but gets different results. Drives my (non-expert) wife crazy.
I call that the Law of Tech Support, and formulated it as such:
"The likelihood of something going wrong is inversely proportional to the availability of the person who can fix it."
In other words, as an extension to Murphy’s Law:
"Things go wrong when tech support is on vacation."
There is another realization here: some OEM and ODMs do not make it easy to perform simple upgrades, such as adding memory.
In our new Windows Vista PC, we had to remove the case cover (three screws… no, four), remove the CD/DVD (one screw, detach several cable connections), detach a face plate, remove an internal cover (one screw) and (last) unplug two cables that cross overtop of the memory DIMMs.
To make matters worse, none of the information was included with the new PC – it was found after 10-15 minutes of coaxing it from the OEM’s web site.
Once apon a time, a long time ago, I was rejigging some parts in a machine I had that was diskless and fanless.
I took out the RAM and put in some new RAM. As soon as I had a matched pair of SIMMs (yes, this was *that* long ago) I heard the telltale beep of the BIOS POST and I saw the OS booting on my monitor.
Turned out that I’d actually hit the reset switch instead of the power button, ripped the RAM out and then put the new RAM in all with the power on. I’m amazed not only that the machine worked after all this, but that it actually managed to start booting as soon as enough RAM was seated without any ill-effect aside from the fact that half of my RAM wasn’t in the system.
Old hardware was much more resilient. I’ve done much less ridiculous things to more modern hardware and had it catch fire, explode or just melt.
Yet another way to mess up (low light version):
Power off. Open case. Put new board in bus slot.
Boot. New board unfound.
Remove new board. I had guillotined the CD audio player line (the 4-pin connector on a CD drive)
I don’t play CDs so it was no loss to me–reseated, everything worked.
[quote]
So do all of the programmers service their own hardware in your neck of the woods, Raymond? Or are you a special exception?
[/quote]
Because I happened to be also an unofficial technician(My official title is programmer) of my company, when I add hardware, I have to self-serve. I have touch to parts of every machines I have access to except the servers, which people have real title "technician" will take care.
"So do all of the programmers service their own hardware in your neck of the woods…?"
I always do – I don’t trust anyone else. :=)
My lessons learned:
1. Always use "standard" screws. There are basically only three different screws in normal PC.
2. Always double-check cables and chips are seated just before turning on the power. A hard failure isn’t so bad because it is *obvious* something is wrong. An intermittent error due to a loose cable is much worse.
3. Always watch where the cables are routed. They might snag a DIMM clip, or get scratched on a sharp metal edge, or get caught in the CPU fan. (Thank goodness for thermal shutdown, or I’d be $250 poorer)
These days it’s harder to confuse this as you have to turn it off at the back of the machine.
At my first job there was this new computer that was always on. Since it was not a server I thought I’d take it off for a night or something – To find it didn’t come back on …
Also past years there’s been these faulty capacitors causing issues. I had a nice guy do a cheap replacement job of changing two that I had visually identified as being faulty on a quite expensive motherboard which sometimes didn’t power up and took like 10 minutes of warming up before it powered with some other minor issues. It was pretty exciting (I anticipated smoke) to turn it off after the servicing and to see that it worked like new again. And has since.
Thank god they took a hint of that and use new type of capacitors that probably last a little bit longer in the motherboards today. (the $150+ ones that is)
That’s why I never put the cover back – my PC is open 24/7 =)
Back in the highschool, like 10 years ago, we installed a new shiny HDD on a school PC and accidentally dropped a screw on the drive while testing the configuration…then we had to lie to the CS teacher that the drive was defective..
Once it took me an hour to realize that the beep doesn’t come from the PC, but from the overloaded UPS next to it.
Can’t Microsoft afford to buy you a ThinkPad? I just swapped both hard drives from one of my ThinkPads to another. It must have taken all of sixty seconds. One screw for the main drive, no screws for the second drive. No pieces laying on the floor, and not much to go wrong. I am *so* glad I gave up desktop computers nine years ago when I got my first ThinkPad.
Dave wrote:
Steve Nuchia wrote:
In Russian computer folklore, we say, “The three most dangerous people are: a programmer with a screwdriver, an engineer with an editor, and the boss with an idea.”
Beeping means either ram or video card isn’t in place or damaged. Next time start from there.
I get the same feeling but worse when I replace a car stereo. Step one of the task is to carefully figure out how to take apart the dash without damaging it. The last step of course is to put it back the way it was. In some cars this takes some bending, forcing and lining up of invisible clips, not something you want to turn around and undo 10 minutes later.
Ray: [What does “loud continuous beep” mean? (Note: No manual.) -Raymond]
Ever heard of Google? – I can’t believe I had to type that.
[How do you google for the manual for a no-name motherboard? -Raymond]
My God, this is rudimentary. You look for the silk screened model number on the PCB then Google for that. – I can’t believe I had to type that.
Or if you can’t find a number, maybe use this query “POST beep code” + whatever BIOS you have.
But I guess the real question is… who buys this no-name stuff anyway?
Another good thing to remember is to always find that last screw (the one you are sure you don’t really need anyway and is probably under the desk somewhere). It’s horrible to find it a couple of months later (when you moved the computer a couple of inches!) inside the case having just shorted something expensive.
"I have no RAM, Help!"
This is one of the things I like about the G5 MACs (and the newer ones, too) – everything is easily accessible. Also, nice cable management.
Beep codes haven’t changed that much in 15 years.
I’m late to the party, but it sounds by his comments like Raymond is trying to wriggle out of not knowing that motherboards/BIOS make error codes with their beeps!
Top Google hit for “asus bios beep codes”: http://bioscentral.com/beepcodes/awardbeep.htm
2nd entry:
Award BIOS Beep Codes:
Beeps: Repeating (endless loop)
Error message: Memory error
Description: Check for improperly seated or missing memory.
Memory generally does a longish beep (5 secs?) repeatedly. Regardless, memory and video are the first two things to check, as that’s where the computer puts its brains.
Of course, I built computers for a while in college, so I remember these things.
For fun, try diagnosing crap power supplies.
[Beats me. It was a hand-me-down. All it says on the motherboard is “ASUS”; no part number that I can find. -Raymond]
Actually, ASUS homepage provides you a feature that allows you to use “CPU sockets” and “onboard chipsets” to narrow down the choice.
Btw, I thought all ASUS motherboard be produced later than 2003 have model number be printed on PCB, and an additional sticker be sticked on the serial/parallel port module.