- Part I: The Intermittent
- Part II: Hardware Purchase Strategies
- Part III: Themes, Letters, Articles, Trademarks and URLs
Editor's Desk
By Steve Litt
Mid July 2008. As the harsh white sun baked
Orlando lifeless, and Hurricane Bertha threatened from thousands of
miles away, I was blissfully unaware that Bertha was just the
beginning. Tropical cyclones Fay, Gustav, Hanna and Ike waited
patiently, in the future, some to flood me and some just to make me nervous. Also waiting in my
future was a tiny bug in my daily-use desktop. A bug that would stop my
work dead in its tracks. A bug that would drag me through the worlds
of intermittence, corrective maintenance and assumption testing.
An insipid little intermittent whose only
power was its anonymity.
Oblivious, I worked on...
Dang!
By Steve Litt
And then my work stopped dead as the computer instantly powered down.
Dang! I lost the last 10 minutes of my work. Why didn't my UPS keep the
computer running through the almost instantaneous power outage?
My
mind reviewed the last 10 seconds. No lights had flickered. There was
no power outage. Just to make sure, I disconnected power to the UPS.
The computer kept running. No power outage could have caused what just
happened. It was an event -- an intermittent that had happened just
once. A real pain in the posterior. Dang!
Having been
around the block a few times, I knew better than waste a lot of time on
a sparse intermittent. The time-efficient strategy was to wait until
the symptom occurred with regularity. And maybe it wouldn't happen
again.
But the a week or so later it happened again. Just once. Dang!
Ignoring It
By Steve Litt
My courseware and books list ignoring intermittents as a valid
intermittent busting tactic, especially for sparse intermittents that
don't cause too much trouble. The sparser an intermittent, the harder
it is to reproduce and troubleshoot, and the easier it is to live with.
I lived with this intermittent reboot for the next two or three weeks.
Early
August saw me on a business trip, with my laptop computer handling
Troubleshooters.Com business. But a few days after my return, the hung with increasing regularity. Shutdowns went from every
other day to every day
to twice a day to several times a day. I'd need to back up every 5
minutes, and still lose data. It was clear this problem needed to be
handled. I rolled up my sleeves and went on in...
Troubleshooting the Intermittent
By Steve Litt
First I tested by plugging the computer into a
regular surge protector instead of the UPS. A few hours later the
computer shut down. OK, the UPS is ruled out.
Time
for corrective maintenance. Can of electronics lubricant in hand, I
removed, sprayed and replaced every internal cable, daughtercard, and
the RAM sticks. No change -- a couple hours later it shut down again.
Shut Down?
Shut
down is perhaps a simplistic word for what it did. The first few times
it appeared to simply lose power, but on occasions a closer look saw
the power supply fan still spinning. It was, at least on some
occasions, a hang so complete as to make a click in the speakers,
instantly remove all video, and remove all keyboard responsiveness, and
prevent pinging from other boxes on the network. |
In
hindsight, the complete loss of everything with the fan still spinning
should have been a hint. But although hindsight is 20/20, at-the-moment
vision can be blurry indeed.
By this time the symptom was occurring every few minutes.
I replaced the video card, powered up and waited awhile. It once again hung.
I
removed 2 of the computer's 3 RAM sticks and let it run for awhile.
Symptom unchanged. Used a different RAM stick. Symptom unchanged.
Switched the RAM stick's position to each of the three slots. No change.
I disconnected the power supply, and without installing it in the box, connected another one. Symptom unchanged.
Somewhere
during this troubleshoot I booted Knoppix in order to replace the whole
operating system. When the symptom kept occurring, it pretty much ruled
out software.
Meanwhile,
the constant hanging and rebooting had taken their toll. The filesystem
on the system disk was corrupted, and the computer kernel panicked on
startup. I'd ruled out the UPS, power supply, video card, RAM,
and internal connections (ruled out by corrective maintenance). What
could it be except the motherboard or the CPU?Steve Litt teaches courses on troubleshooting. You can email Steve here.
Pragmatism
By Steve Litt
Troubleshooters.Com's website remained accessible, but as a business,
Troubleshooters.Com was now shut down. No email, no Paypal, no order fulfillment. Some of you who ordered books in
mid August know that only too well. It was obvious I had to transfer
operations to a new computer as quickly as possible, rather than
continuing to find the root cause of this frustrating intermittent.
My decision was supported by troubleshooting theory. One aspect of Step 2: Make a Damage Control Plan
is preventing business outage, and getting business running as quickly
as possible, even if that means a delay in finding the root cause.
Once
upon a time transferring operations to another computer would have been
easy. In 2003-2005 I bought about 5 similar computers from LEK
Technology Consultants -- all Athlon/Asus. Putting my hard disks and
RAM into one of those other computers wouldn't have been hard.
But
two of the computers had died of intermittency -- actually with
symptoms that might have resembled what I had now. One of the still-working computers
had been pressed into service as an IPCop box when I switched to
broadband, and I hesitated to remove it from that function and put back
my mid-1990's Dell that had been doing that duty back in my dialup
days. Another of the working computers is being used on a daily basis by one of
my daughters, and also serves as the family's backup server. The
fifth was my daily driver desktop that was now too intermittent to use.
So the deep computer redundancy from a couple years ago was gone. I
needed to be practical and buy a brand new computer to replace my
current one.
It's Harder Than It Looks
By Steve Litt
Coincidentally, for the past few months I'd been
researching a new computer to replace my 3 year old daily driver
computer. Because most of my Athlon/Asus fleet had succumbed to or at
least been affected by intermittent hangs, shutdowns and reboots, I'd
decided to go Intel/Intel this time. I'd even gotten prices from a
local computer shop. But I procrastinated.
Meanwhile, in late
July, it was announced on the Ubuntu forums that Foxconn manufactured
boards malfunctioned with Linux (announcement URL in URLs section). For the past four years I'd assumed
that motherboards all worked with Linux, and only periperals like
sound, video and network could fail to support Linux. Now all of a
sudden I was in danger of buying a doorstop motherboard. I had to be
much more careful.
After a couple days of considering piecing
something together by buying parts at CompUSA, the need to get
Troubleshooters.Com operational RIGHT NOW became evident. I transferred
all operations to my laptop, using a desktop monitor, keyboard and
mouse on it. This widescreen notebook is ugly on a regular geometry CRT
monitor, and nothing I could do to xorg.conf could change that, but at least I could work, albeit at a slower pace.
In
the end I bought from LEK (URL in URLs section), for the simple reason that LEK will let me
load Linux before I buy, meaning whatever I buy from them will be Linux
compatable. I ended up getting an Intel/Asus combo. Considering the
problems I had with my 5 previous Asus mobos, I didn't like going with an Asus mobo, but I
had to get back to work.
Here's
a little peek into the future -- I now think Asus had absolutely
nothing to do with my problems, on my main computer or on the others
that had gone intermittent.
One
problem with LEK's Asus mobo -- it incorrectly registered the CPU
temperature as 48 Farenheit -- a refrigerator :-). They upgraded the
board's bios to cure the problem.
Troubleshooters.Com's Desktop
By Steve Litt
Once LEK's hardware worked, I loaded Mandriva
2008 Linux on it, transferred the old computer's data disk to the new
one with a disk to disk transfer, and then did a differential rsync
from the notebook computer so as to have the current data set. A few
tweaks to software, and Troubleshooters.Com's work was transferred from
the notebook to the desktop, and it was business as usual.
Now:
what to do with the carcass of the old desktop? How to give it parts
necessary to make it a working computer? I went to CompUSA to figure
the cheapest way to repair the old computer. This search for the best
way to repair that computer would bring to light several complications.
Salvaging My Old Computer
By Steve Litt
As
mentioned, of the five original Athlon/Asus computers bought in
2003-2005, three were broken, one was a firewall, and one, which had
become increasingly intermittent, was being used by my daughter. And my
son's computer is a frustratingly slow nine year old dinosaur. Bottom
line, I needed more computers, not merely a replacement for my daily
driver.
So I figured I'd buy a motherboard, CPU, memory, and
SATA hard drive from CompUSA, slam it in my old case, and have a second
computer for not too much money. My existing power supply had the
connectors to accommodate a modern motherboard. I bought an Athlon dual
core and an Asus mobo, and put it together, tested it, and it worked
perfectly. Happily, I turned off the fixed computer moved on to other
things.
Oh Man, What Now?!
By Steve Litt
A few days after turning off the repaired
computer, I turned it back on. It ran for 5 minutes and rebooted
itself. A few minutes later it rebooted itself again. Then it sort of
hung, requiring a powerdown and powerup. Repeatedly, it would run for a
minute and hang. What the heck happened to this thing in the few days I
didn't use it? How was I going to get the money back from CompUSA?
Disgusted, I turned it off.
A couple nights later I couldn't
sleep. Why oh why were modern computers so difficult? Tossing and
turning and going through the situatation, the following made itself
apparent:
- New mobo
- New CPU
- New RAM
- New hard disk
- New operating system (Mandriva 2008 instead of Mandriva 2007)
- Power supply proven to be not the cause good in previous diagnostic tests
It
was a totally new computer except the case. Hmm, the case. HEY!!! That
was the same case that housed the original broken computer. And this
set of components was exhibiting approximately the same set of symptoms
as the first set of components.
What could be wrong with a case?
Well, maybe it bends daughter cards, or maybe it has an exotic
grounding problem, but by far the most likely failure in a case are the
switches, LED, speakers.
I turned on the computer and watched it
reboot several times. Then I yanked from the mobo the leads for the
reset switch, power switch, all LEDs and speaker. It stopped rebooting.
I kept it in a specific bios config screen and went to bed. Waking up
late the next morning, it was still in that same bios screen. I put
back the leads and it started rebooting.
GOTCHA SUCKER!
Spraying
electronic lubricant into both switches didn't cure the problem, but
simply leaving the reset switch disconnected did. The reset switch had
been intermittently closed and stayed closed, creating a symptom which
casually looked like an unscheduled powerdown and on closer examination
looked like a system hang with loss of video. Or sometimes it would
close just for an instant and reboot the computer.
So I put it together, left it cooking for a day -- no further problems. It had been the reset switch!Yet Another Computer
By Steve Litt
This meant that my original motherboard, CPU, RAM
and disks were good! They hadn't yet been thrown out. I put the
old components into a case purchased a $45.00 case from LEK, loaded an
operating system, and it worked perfectly. Left it on for hours, no
problem.Steve Litt is the author of the Samba Unleashed. You can email Steve here.
Maybe More Computers?
By Steve Litt
My daughter's
Athlon/Asus has been intermittently hanging. It was one of a fleet of computers
bought from LEK in 2003-2005. All with the same type of case.
It won't be long now before I replace that computer's switches. Of
course, different cases use different switches, so I can't purchase an
exact replacement to fit the computer's button set.
But what I
can do is wire a cheapo doorbell button to the wires for the switch.
The button would need to be surrounded by a raised "wall" to prevent
accidental turnoffs and reboots, but that isn't hard.
EDITORS NOTE, 6/26/2010
I disconnected my daughter's reboot switch and replaced her power switch
with a cheap doorbell switch from Home Depot. The computer's been rock
stable ever since.
|
The first
Athlon/Asus to go bad was actually a Duron/Asus, and when it went
intermittent I threw it out. Too bad, I'd love to go back and try it
without case switches. Oh well, at least 3 of the 5 are still running.
But...
Another
one that went bad is still sitting, with CPU and RAM still attached, in
a static proof bag on my bookshelf. One of these days I'll hook it to a
power supply, video card and RAM, load Linux on it, and test for a few
days. Perhaps all it needs is a brand new case.Steve Litt teaches courses on troubleshooting. You can email Steve here.
Step 9: Take Pride
By Steve Litt
It took awhile, and required several missteps, but I solved this
problem. The fact that it was my main business machine made it a high
pressure situation, ripe for troubleshooters mistakes, getting it fixed is a tribute to my troubleshooting ability.
Step 9: Take Pride encompasses two activities:
- How was I brilliant?
- How could I have been even better?
I
was brilliant because I kept Troubleshooters.Com running throughout
this difficult situation, and because I fixed the problem in a
reasonable manner.
That being said, there
was obviously room for improvement. If I'd checked the power and reset
switches right away, this could have been solved in a few minutes
without the purchase of other parts and equipment. Checking the
switches definitely qualifies as appropriate corrective maintenance. So
how can I avoid this mistake next time?
First and most obvious,
by keeping in mind that intermittent hangs, intermittent shutdowns and
intermittent reboots can easily be caused by the switches. This can be
tested by disconnecting the switches, after which the computer must be
started by shorting two pins with a screwdriver.
Unfortunately,
disconnecting the switches isn't an easy diagnostic test.
Typically the switches plug into a group of pins in the lower
forward corner of the motherboard, in an area obscured by the case.
It's hard to see, and hard to manipulate either with your hands or with
a needlenose pliers. Although disconnecting everything is a 10 second
affair, connecting them back again can take 10 minutes if you know
where things go, hours otherwise.
Before approaching this
diagnostic test, prepare by securing a small, bright flashlight, a
pencil and paper. Draw a large diagram of the pin group, pull out one
connector at a time, and write down which pins it was on. Also write
down what is written on the connector you removed, and which direction
the writing was in. If the motherboard is marked, write down the
markings too. By taking an extra 10 minutes to make this diagram now,
you can avoid hours of experimentation to figure out what to put where,
and which polarity.
Bless their hearts, some modern motherboards
come with a pin block that plugs into the pin group and has pins coming
out the other end, so that you can attach the connectors to the pin
block and when they're all connected, connect the pin block to the pin
group. This can save about 9 of the 10 minutes it takes a person with
big, clumsy hands (like me) in connecting the connectors to the pin.
Also, the pin block has connector names very clearly written on it in
fonts twice the size of the motherboard.
Evaluation of My Decisions
The
decisions I made determined the course and duration of the
troubleshoot. It could have lasted an hour, it could have lasted a
month. Let's evaluate some of those decisions:
Ignoring the Intermittent While Sparse
Ignoring
the intermittent while it was happening less than twice a day was a
good choice. Sparse intermittents cause less trouble (because they
happen less often), and they're much harder to troubleshoot. I was able
to work effectively for a month or more while this was happening less
than once a day. As it happened more frequently, I made sure to save my
work frequently.
Plugging into a Surge Protector Instead of the UPS
This
was a very quick, easy test, and a test enabling me to keep on working
while the test for this intermittent was happening. Good move on my
part.Cleaning the Dust Out
Corrective
maintenance. This should have been preventive maintenance -- dust
buildup in a computer reduces the life of all components. This was a
good thing to do right away. It should have been done months ago.Disconnecting, Lubricating, Reconnecting All Cables, Daughtercards and RAM
This
was the first thing I tried. It's corrective maintenance, so I did it
before trying diagnostic tests, as suggested by the Universal
Troubleshooting Process. Although this is time consuming, with an
intermittent it's probably the best use of the time.
Swapping the Video Card
Video
card swaps are usually pretty easy, but configuring X afterwards is
somewhat time consuming. In this case, by the time I swapped the video
card, the computer was auto-rebooting in text mode and even from within
bios setup, so this was a decent move. Certainly, before one concludes
it's the motherboard/CPU, the video card must be swapped so as not to
accidentally throw out a good motherboard and CPU.
If you're using the onboard video card, putting in a video card daughterboard is often a good idea.
It was a fairly quick test, so I'll get a passing grade on this, although in hindsight perhaps I should have delayed it.
Disconnecting Everything from the Motherboard but the Power Supply, RAM, CPU and Video Card
This is standard with any kind of boot problem, and is always time well spent when there's a boot problem. Good move.Wiggling Everything
I
don't remember whether I wiggled all cables and physically manipulated
the motherboard, but if I didn't I should have. That would have led to
the discovery of an intermittent cable, daughtercard, memory or
motherboard. I think I forgot to do this, and I should have remembered.Booting Knoppix
Is
it hardware, or software? By booting Knoppix, I replaced the whole
operating system (the original system was Mandriva 2007). When the
system rebooted unchanged after booting Knoppix, it was pretty clear it
wasn't a software problem, so it was hardware. Good move.Swapping the RAM
This
computer had three RAM sticks. By removing two, and then alternatingly
swapping the one RAM stick, and even moving the RAM between slots, I
ruled out the RAM (unless all three slots and sticks magically developed the same
problem). A quick test that ruled out a major component, this RAM swap
was an excellent move.
Swapping the Power Supply
Swapping
the power supply is a tougher diagnostic than RAM stick swapping, but
it's made easier by not mounting the testing power supply. Instead, all
current power supply connections are disconnected and replaced by
connections from a separate power supply. The problem remained the same.
One
problem with this diagnostic is the computer must be laid on its side
so the heavy, dangling power supply doesn't pull on the motherboard,
disks and fans. That precludes booting from a CD, which in this case
wasn't an issue. It also could conceivably mask a mechanical
intermittent depending on the box's orientation.
Even with those caveats, this 15 minute diagnostic was still a good one.
Forgetting the Case
After
ruling out the software, power supply, cable connections, daughtercard
and memory connections, and the RAM, I assumed I'd narrowed it down to
the motherboard and CPU.
Bzzzzzt, WRONG!
I hadn't yet swapped the case.
A case can be at fault for a number of different reasons:
- Short circuit (perhaps caused by a screw that fell in)
- Grounding problems
- Inability to properly seat daughtercards (usually in a cheap case)
- In-case electronics such as LEDs, speaker, and reset and power switches
I
have big, clumsy hands, poor eyesight, and not a lot of room around the
house to spread out my work. Removing and reinserting the motherboard
and daughtercards isn't a trivial task. Also, there's a chance of
harming the motherboard through an electrical mistake or trying
forcefully to remove it while a screw remains in place. Removing the
motherboard is not trivial.
However, by far the most likely
culprit in the case is the case's electrical system -- switches, LEDs,
and speaker. These can be disconnected much more easily than
motherboard removal. As mentioned, disconnection requires preparation
and documentation, but it's a reasonably quick test. It's a test I
didn't do. It's a test that would have located the root cause in 20
minutes.
In short, I proceeded on an invalid assumption, which
led me to the BIG MISTAKE -- thinking the root cause was in one section
when it was really in another. Proceding on invalid assumptions and
making the big mistake, costing days of time, is a well known
phenominon in troubleshooting. I should have known better.
From
now on, I'll ask myself early whether the switches could be at fault,
and disconnect them to test, being careful to prepare and document. If
necessary, I'll also be prepared to remove the motherboard to totally
rule out the case.
Buying a New Computer
The minute I
supposedly narrowed it down to the motherboard/CPU, I went shopping for
a new computer. A brand new, $800.00 computer. We now know that was
unnecessary, but was it a good decision at the time?
Step 2: Make a Damage Control
includes the need to prevent business interruption. This broken
computer was interrupting my business bigtime. The minute I became
convinced that troubleshooting might take longer than simply installing
an operating system and data on a new computer, I went shopping. Even
though it later turned out to be unnecessary, and even though the old
computer is running once more (in a different case), I'll stand by my
decision. It got me back to work.
I've had a very strong bias toward never having my main computer down, ever since a 1996 computer building debacle in which I cannibalized my daily driver computer to make my new computer,
and was out of business for a week. If you're troubleshooting
technology essential for your business, your first allegiance is to
your business, and if that means buying and swapping a whole box before
troubleshooting the old box, so be it.
Planned Obsolescence?
By Steve Litt
NOTE
The preceding articles described my repair of an intermittent, and my
shopping for new computers. Those days of shopping made me aware of
some disturbing trends in computer hardware. Read on...
|
After 10 hours perusing motherboards and other
devices at Compusa, one thing was evident: Almost nothing from my
intermittent daily driver desktop could be used with today's modern,
"legacy free" motherboards. Most did not have a parallel port to
support my 17ppm, duplexing HP 4050 printer. None supported my Athlon
XP2600+ processor, nor my current RAM sticks. Most had only one, or at
the most two PCI slots, meaning that if the computer's built in network
and sound facilities didn't work with Linux, I was in trouble. Most had
4 pin or 8 pin 12volt power connections my older power supplies
didn't supply.
In other words, if your 3 year old computer's
motherboard breaks, you might as well buy a whole new computer and
throw out the entire old one.
Trouble is, up til now I've been
operating on the assumption of an 8 year life for my computers. The
first 3 years a new computer is my daily driver. Then it becomes an
experimental machine for a year or two, and then gets passed down to
one of the kids, then maybe down again to another one of the kids, and
at maybe 6 years old, when it can no longer run modern operating
systems, it becomes a firewall or lightly used server, or just an
experimental machine. Buying one new computer every 3 years was keeping the
whole family on computers suitable for their use. If a motherboard
broke, I bought a new one and maybe a new CPU and RAM to go with it.
But
now, with motherboards breaking every 4 years and changing connection
types even faster, my hand-me-down techniques weren't working any more.
It angers me that my fleet of Athlon/Asus computers is almost all gone,
while my son still uses the 8 year old Abit/Dual_celeron purchased in
February 2000, and used as my daily driver 2001-2004. The Pentium II300
purchased in December 1997, used as my Windows daily driver from then
until March 2001, still functions as a headless Windows application
appliance to make MS Office files for my customers, and to access old
files built with Windows-only programs (WordPerfect 5.1, Micrografx
Windows Draw, etc).
Something has happened to Mean Time Between
Failures on modern motherboards, and that something isn't good. At
least it's not good for me. It's great for hardware vendors.
Another
CompUSA customer said Microsoft was pushing motherboard makers to
manufacture "legacy free" motherboards. That means no floppy socket, no
IDE sockets, no PCI sockets, no AGP socket, no parallel port, no serial
port, USB instead of PS/2 keyboard and mouse. Of course,
motherboard makers haven't yet murdered all of those longstanding
standards, but it's happening at an ever increasing rate. A web article
that extoles the virtues of a "legacy free" PC (URL in URLs section of
this magazine) states that both Intel and Microsoft have been
encouraging motherboard makers to go "legacy free".
Is this a
quantum shift in computers. I think not. I predict that three years
from now, much of this new stuff will be considered legacy, six years
from now it all will be. The purchase of a hard disk or a daughtercard
will be aimed at an increasingly short window of opportunity to be used
with a motherboard.
The parts interchangeability we've come to
appreciate in computers is disappearing. Why? It's hard to say, but as
a starting point, I'd say "follow the money". Computers break sooner,
and one broken part often necessitates the purchase of a whole new
computer for lack of compatible parts. Follow the money.
Steve Litt teaches courses on troubleshooting. You can email Steve here.
I've Seen the Future, and It's Ugly
By Steve Litt
I've seen what happens when an industry follows the "planned obsolescence" path recently embarked on by the
computer industry. The bicycle industry started on that path in the
early 1990's.
In
the 1980's there were basically two kinds of bikes -- American bikes
and non-American bikes. American bikes were made of thicker walled
tubing with a smaller outer diameter. All tubes of all American bikes
had the same diameter. Most American bikes used simple, one piece
cranks you still see on the Roadmaster bikes at Walmart. Non-American
bikes all had larger tubing, uniform in outer and inner diameter across
brands, brazed frames and three piece cranks which were pretty much all
interchangeable. All bikes had caliper brakes attached at the fork
tops. Brakes were interchangeable. All derailleur bikes had 5 or 6
speed cog sets in the back and manual shifters that could accommodate
any cogset. The Suntour company had modular cogsets such that you
could build your own cogset with any sized gears you wanted.
So
in the 1980's, you could pretty much replace any component with any
other component. Repair and upgrade were trivial. With many consumer
choices, competition kept prices down.
Then, in the early
1990's, they came out with "indexed shifters", which enabled you to
shift gears almost instantly. They were a heck of a lot more
convenient, with a downside not noticed at first -- the shift lever had
to be exactly matched to the derailleur and the rear cogset. All sorts
of different indexed shifters proliferated, but to upgrade any part of
the bicycle transmission, you had to upgrade cogset, derailleur and
shift levers. Oh, and they got a new way to attach the cogset to the
rear wheel, so you might have to replace the rear wheel too.
Move
forward ten years. Every bicycle has different sized tubing, so things
like shift levers can't go on the frame. Every bicycle requires a
specific diameter seatpost, so tall riders can't just buy a long
seatpost -- it's much more complicated. For the rider's convenience,
they began incorporating shift levers on brake levers, so now a
transmission change might mean a change of the braking system too. The
new cantelever and V brakes mean you can't just slam on a caliper brake
anymore. Just
the other day, I saw a bicyle whose seatpost was oval rather than
round. If you need a seatpost for that bicycle, you'd better hope the
bicycle's manufacturer makes it, and you better be prepared to pay
whatever they charge, because they're the only game in town.
When
parts interchangeability is gone, parts and the whole cost more, and
repair and upgrading become a chore. I appreciate the ability to buy a
bunch of components and put them together into just the computer I
like. I pray I'm wrong, but it looks to me like 5-10 years from now
building your own will be overwhelming due to part specificity, and the
only reasonable choice will be to buy whatever computer setup the
marketing guys in the computer companies decide I want.
Computer Buying Decisions
By Steve Litt
There are two types of computer obsolescence:
- Performance obsolescence
- Connection obsolescence
Performance Obsolescence
The
former has always been with us. It's pretty sure if you buy a computer
today, it won't efficiently run the software made three years from now.
Performance obsolescence has always been with us, and there's a good
way to deal with it -- trickle down computing. After 3 years your main
desktop computer goes to one of your kids. Three years later it becomes
a firewall or server appliance. Three years after that it's nine years
old -- you throw it in trash -- it's paid for itself.
No computer lasts in toto
for nine years. Something's going to break during that time. In the
good old days from 1985 to 2005, you just bought a replacement part,
because hardware standards changed very slowly. A 2000 hard disk would
fit a 1990 computer, always assuming you could find a BIOS update to
accommodate the new drive's geometry. A 1990 hard disk would bolt right
into a 2000 computer. A 2005 ATX power supply would exactly replace a
1995 ATX power supply. The converse isn't true, but a 1995 ATX power
supply could have run a 2000 ATX computer.
Bottom line, from
1985 to somewhere early in the 2000s, a broken component could replaced
without replacing a whole slew of other components, so you could keep
your computers going for 9 years.
Connection Obsolescence
If
the only thing that's obsolete is performance, the computer's life can
be extended simply by using it in a less demanding capacity, and if
parts break, you get new ones.
In the 21st Century, however,
motherboards and other components have continually and quickly
undergone radical transformations. ATX motherboards first required 20
pin connectors, then 24 pin connectors, then 24 pin connectors with 4
pin 12 volt auxilliary connectors, or sometimes 8 pin 12 volt
auxilliary connectors.
In this decade, video cards went from PCI
to (horrible) ATX, and now to PCI Express (PCIe). ATX had three
different, incompatible variants. The old PCI slots hosted not only
video cards, but sound cards, network cards, additional hard disk
interfaces -- pretty much everything you'd want to put in a computer.
In 2000, any self respecting motherboard would have 4 or 5 PCI slots,
and maybe 1 or 2 ISA slots to work with 1992 technology. Now in 2008,
you're lucky if your motherboard has more than 1 PCI slot. So those PCI
video, sound, and network cards you've collected over the years are
almost useless. Yet if the motherboard's built in video, network and
sound haven't programmed in compatibility with your operating system
(Linux folks beware), you'll need those daughtercards. Yeah, maybe you
could replace them with PCIe, but good luck finding those. It's ugly.
IDE vs. SATA
In
the last few years, Serial ATA hard and CD drives have replaced IDE (also
called Parallel ATA) hard and CD drives. That 3 year old half gig hard drive
that you spent a fortune on three years ago is still big enough for today, but
it won't fit the motherboard unless the motherboard comes with an IDE port
(they never come with two any more). Worse yet, some motherboards can't boot
off an IDE drive, so you have to go SATA. Worse still, bootable CDs more than a
couple years old (and even some current ones) don't have SATA drivers, so you
can't boot off of them either. I use Windows 98 to access my old data. Windows
98 can't be installed on a new computer. Luckily for me, my 1997 Pentium II300
works still runs perfectly after 11 years. They don't build equipment like that
anymore.
Quality Reduction
As
mentioned, the Pentium II/300 I bought in December 1997 still works
perfectly, serving as a Windows 98 appliance. This computer served as
my daily driver 12/1997-3/2001 -- it's been ridden hard. But they made
electronics more reliable back in those days. You're really lucky to
get a modern motherboard to last 6 years.
One could make the
argument that if it lasted 6 years it's served its purpose, but don't
forget you sometimes need to run legacy software, and legacy software
often doesn't work with modern operating systems or hardware. Ugh!
Purchase Strategy
The
change to rapid connection obsolescence and quicker time to failure
dramatically impacts your purchase strategy. Take hard drives as an
example.
In the old days, you bought the biggest hard drive
available, even if its cost per megabyte or gigabyte was significantly
more. The reason was simple enough -- that huge drive would be modest
in 3 years and paltry in 6. If you'd bought a more moderate hard drive,
it would have been paltry in three years. And of course, you could
easily move the hard drive to a new motherboard when you bought one.
How
different it is now. The world has just upgraded to SATA drives. I'm guessing
within 3 years SATA will be out and something else will be in. The
drive you buy today is likely to suffer connection obsolescense in
three years. You'd be better off buying what you'll need for three
years than going as big as you can.
RAM is similar. That 8GB of
RAM you buy today for your current motherboard will likely be useless
for a motherboard manufactured two years from now. Buy what you need.
But Remember RAM Is a Good Investment
Even
though a motherboard failure might render useless the RAM you buy
today, it's still often a good strategy to go heavy on RAM. As you
upgrade software and operating systems, work on bigger files, and open
more programs simultaneously, you'll consume more semiconductor RAM.
When you exceed the semiconductor RAMs capacity, memory will swap to
the hard disk -- slow at best, crashy at worst. Personally, I'd spend
less on my processor and really load up on RAM, to process at high
speed no matter what the task. |
So
what are your priorities when buying a new computer? As always, buy at
the knee of the price/performance curve. Bleeding edge costs too much
for the added capabilities, and trailing edge saves too little money to
justify the decreased performance.
Assume your new computer will
last only three years. If you think you'll need 500GB of disk space toward the
end of the three year period, don't buy a 1TB drive -- it's likely that
drive will be connection obsolete by that time. Try to get a cheap hard
drive, because now that you'll be buying brand new hard drives every 3
years, the expense mounts up.
To the best of your ability, try
to use the network, video and sound provided by the motherboard. If you
use the latest Windows, that should be no problem. If you use Linux,
even the latest Linux, the free software will need a few months to
reverse engineer the hardware devices for which vendors refuse to
release specifications, so you may need a daughtercard for some of
these functionalities.
Speaking of Linux, depending on your
vendor it might be possible to "try before you buy". Connect a power
supply to a mobo with lots of RAM, connect a DVD drive, a monitor,
network and sound system, and boot a few different leading edge Linux
live CDs or DVDs. If any one of them can get the motherboard's built in
video, sound and network to work, the motherboard is Linux probably
compatable enough -- one way or another you can probably finesse the
latest version of your favorite Linux distribution to work with it.
I've
always recommended motherboards with lots of expansion slots. These
days, you must also make sure you have plenty of "legacy" slots, so
that for the next couple years, you can use your older but still good
disks and daughtercards. This will be true forever more. Remember, that
oh-so-modern PCIe sound card you buy today will be connectionally
obsolete in a few years unless you get a motherboard with a "legacy"
PCIe slot.
With all this obsolescence going on, it might be best
to buy cheaper components. Of couse, you don't want to buy one of those
price point motherboards known to be DOA or go bad soon, but don't buy
the $159.00 model of the good brand so it can give you some special
front side bus or northbridge or southbridge. Get a connectionally
modern processor for less than a hundred bucks. Go heavy on memory, but
not so heavy it will still be extra in 3 years. Size disks to what you
think you'll need in the next three years.
And then, with the money you've saved, plan to use it for only 2 years.
Where to Buy
In
the 1990's I lived in Southern California. Half the people in Monterrey
Park had cousins in Taiwan, ran computer stores, and had a large
variety of equipment straight from Taiwan. If you live in such a place,
those computer stores are where to buy.
Things are a little
tougher for the rest of us. We buy from local stores that buy from
those stores in Monterrey, or from distributors that buy from those
stores in Monterrey Park. Often, local prices are high, and the
selection is sparse.
My friends at Greater Orlando Linux User
Group (GoLUG) maintain fleets of older computers and practice trickle
down computing like I do. They often need to replace a motherboard and
keep the same components. Some of them get an exact replacement on
Ebay. Sounds risky to me, but they say it works out well. According to
them, the main thing is to give the seller complete instructions on how
to package the motherboard or other equipment in order to prevent
static damage and mechanical damage. The motherboard should be in a
truly static proof bag -- the silver ones are the best according to my
friends. Then, package it for mechanical protection. My friends tell me
the seller is often able to throw in a CPU and RAM, so you have
everything you need.
Many times my friends look at a computer
malfunction as an excuse to get a modern new computer. Everything new
-- mobo, RAM, drives, maybe even power supply and case. This is much
easier, but still a challenge. They usually like to get what others
call "legacy" connections -- PCI, IDE, PS/2, AGP, and parallel port.
The local computer store or even the local superstore might not have
what they want, so they often go to NewEgg to order. This sounds risky
to me if you're installing Linux, but some of my friends have always
been able to do what it took to get everything working on these
sight-unseen motherboards. Personally, my experience is that with
desktop motherboards, I can usually get Linux running, although
typically the built in sound card won't work, and maybe the built in
network card won't work either. This is the reason I like to have lots
of expansion slots, including the "legacy" ones.
The alternative
I use most often for my business desktops is to go to a vendor who lets
me "try before I buy". I go to LEK Computers (URL in URLs section) with
a DVD drive, a hard drive, and a whole bunch of up to date live Linux
CDs and Linux install CDs. With the mobo on a static proof envelope,
they put in CPU and RAM, connect it up to my DVD drive, disk, network
and sound system. I boot live CDs and verify that video, sound and
network work. I install Linux on the hard drive. Once the mobo's been
proven to support Linux, I buy THAT mobo, RAM and CPU, all the other
things necessary to make it a computer, pay them $25 to install it (I
have bad eyes and big, clumsy hands), and take it home a few days
later. I don't spend a cent until I know it works.
These are just some of the alternatives in buying a computer. There are probably many others that work.Summary
There
was a time when a six year old motherboard would run a disk drive you
bought today, and vice versa. Those days are gone. Connection standards
are changing too fast, and motherboard mean time to first failure too
short, to depend on more than 3 years service from any component --
motherboard, drive, daughtercard, RAM, case or power supply.
So
buy quality, but buy cheap. Don't pay the extra hundred for the faster
CPU -- in two years an even faster CPU will be dirt cheap. Don't buy
more disk than you'll need in the next three years -- it's likely that
disk will be incompatible with motherboards three years into the
future. Given today's situation, your computer must be though of more
as a single unit, and less as a group of mix and match components. Buy
motherboards with some backward compatibility to stretch the value of
other components.
When it comes to modern computers, buy cheap, buy often, and buy BEFORE your main computer self-destructs.Troubleshooting Professional Needs Theme Ideas
By Steve Litt
This is the eleventh year I've published Troubleshooting Professional Magazine, and I'm running out of theme ideas.
I've written about the
UTP, The Attitude, intermittence, bottleneck analysis, toolsmanship,
and generic problem solving. I've written short stories, dense and dry
treatises, and humor. I can't think of what else to write.
Which is why I need your help. Please email me with topics you'd like to see covered in future Troubleshooting Professional Magazine issues.
Steve Litt is the author of Samba Unleashed. You can email Steve here.
Letters to the
Editor
All letters become the property of the publisher (Steve Litt), and
may
be edited for clarity or brevity. We especially welcome additions,
clarifications,
corrections or flames from vendors whose products have been reviewed in
this
magazine. We reserve the right to not publish letters we deem in
bad taste
(bad language, obscenity, hate, lewd, violence, etc.).
Submit letters to the editor to Steve Litt's email address, and be
sure
the subject reads "Letter to the Editor". We regret that we cannot
return
your letter, so please make a copy of it for future reference.
How to Submit an Article
We anticipate two to five articles per issue, with issues coming out
monthly.
We look for articles that pertain to the Troubleshooting Process, or
articles
on tools, equipment or systems with a Troubleshooting slant. This can
be
done as an essay, with humor, with a case study, or some other literary
device.
A Troubleshooting poem would be nice. Submissions may mention a
specific product,
but must be useful without the purchase of that product. Content must
greatly
overpower advertising. Submissions should be between 250 and 2000 words
long.
Any article submitted to Troubleshooting Professional Magazine must
be
licensed with the Open Publication License, which you can view at
http://opencontent.org/openpub/.
At your option you may elect the option to prohibit substantive
modifications.
However, in order to publish your article in Troubleshooting
Professional
Magazine, you must decline the option to prohibit commercial use,
because
Troubleshooting Professional Magazine is a commercial publication.
Obviously, you must be the copyright holder and must be legally able
to
so license the article. We do not currently pay for articles.
Troubleshooters.Com reserves the right to edit any submission for
clarity
or brevity, within the scope of the Open Publication License. If you
elect
to prohibit substantive modifications, we may elect to place editors
notes
outside of your material, or reject the submission, or send it back for
modification.
Any published article will include a two sentence description of the
author,
a hypertext link to his or her email, and a phone number if desired.
Upon
request, we will include a hypertext link, at the end of the magazine
issue,
to the author's website, providing that website meets the
Troubleshooters.Com
criteria for links
and that the
author's
website first links to Troubleshooters.Com. Authors: please understand
we
can't place hyperlinks inside articles. If we did, only the first
article
would be read, and we can't place every article first.
Submissions should be emailed to Steve Litt's email address, with
subject
line Article Submission. The first paragraph of your message should
read
as follows (unless other arrangements are previously made in writing):
Copyright (c) 2001 by <your name>. This
material
may be distributed only subject to the terms and conditions set forth
in
the Open Publication License, version Draft v1.0, 8 June 1999
(Available
at http://www.troubleshooters.com/openpub04.txt/ (wordwrapped for
readability
at http://www.troubleshooters.com/openpub04_wrapped.txt). The latest
version
is presently available at http://www.opencontent.org/openpub/).
Open Publication License Option A [ is | is not]
elected,
so this document [may | may not] be modified. Option B is not elected,
so
this material may be published for commercial purposes.
After that paragraph, write the title, text of the article, and a
two
sentence description of the author.
Why not Draft v1.0, 8 June 1999 OR LATER
The Open Publication License recommends using the word "or later" to
describe
the version of the license. That is unacceptable for Troubleshooting
Professional
Magazine because we do not know the provisions of that newer version,
so
it makes no sense to commit to it. We all hope later versions will be
better,
but there's always a chance that leadership will change. We cannot take
the
chance that the disclaimer of warranty will be dropped in a later
version.
Trademarks
All trademarks are the property of their respective owners.
Troubleshooters.Com
(R) is a registered trademark of Steve Litt.
URLs Mentioned in this Issue