Burn-in? What's it all about anyway?

  • Thread starter Thread starter Ruel Smith
  • Start date Start date
R

Ruel Smith

Russ said:
This is probably going to sound like a dumb question to most, but I have
to
ask it anyway. I've heard of 'burn-in' from years past, but I always
thought of it as a testing process. Can a motherboard improve
operationally
with a little bit of crunch/processing time? Can this initial working
period be helpful in 'seating' things a bit?
Click to expand...

As far as I know, burn-in is only to test the stability of an overclocked
system. If it can handle the intense computations for a sustained length of
time with locking up, crashing, or getting too hot, you're good to go.
Maybe you can tweak it a little more until it doesn't fair so well and back
it off to the highest good setting you had.

However, in my experience, that type of testing doesn't guarantee anything.
I've had what I thought were rock solid systems that would freeze when I
tried to print or when a certain game was launched.
 
This is probably going to sound like a dumb question to most, but I have to
ask it anyway. I've heard of 'burn-in' from years past, but I always
thought of it as a testing process. Can a motherboard improve operationally
with a little bit of crunch/processing time? Can this initial working
period be helpful in 'seating' things a bit?

The reason I'm asking is that I had a miserable time getting a new
motherboard to work at 166 Mhz FSU. I sent it back to the vendor and
received the replacement Friday night. When I first started using it,
almost immediately I experienced instances where the mouse froze up but
windows did not. The scanner was just getting garbage. Software was
starting up all by itself. I also had a few lockups while processing video
files (encoding). I thought 'here we go again', and was pretty disappointed
(again) after only 4 hours. After a little bit of shut down time, I started
it up again before I went to bed. It's been running for 36 hours since
then, encoding video without any problems at all. Even my scanner works
properly now when it didn't at first. Incredible! Everything seems to be
running just fine now.

It doesn't make any sense to me but it's almost like the motherboard/CPU
needed a short 'break-in' period to operate properly. Truth, or is it just
me looking for a (il)logical 'splanation? :)

Thanks...
 
haven't done one since 1990 even then I think it was more 'lets see
if the darn thing will run for 'X' amount of time rather than a 'burn
in'. Some wrench jockey probably came up with it... ;^) sounds more
like an 'engine' thing to me.
 
Russ said:
This is probably going to sound like a dumb question to most, but I have to
ask it anyway. I've heard of 'burn-in' from years past, but I always
thought of it as a testing process.
Click to expand...

You're correct. "Burn in" refers to operating components under stress
(often elevated temperature, hence the term) to induce infant mortality
failures, which one hopes won't occur. If the device gets through this then
it should fall in the 'mature' portion of the reliability curve, assuming
one didn't damage it during the test (always a point of controversy). Note
that it does not 'alter' the SUCCESSFUL device: it doesn't fail. Its the
ones you throw away that one could say 'changed' in some way.

Basically, the term 'burn in' has been hijacked by the computer hobby
field, mainly from a misunderstanding of its meaning, to 'explain' the
'unexplained': usually some unexpected 'positive' result.

('Run in', in the mechanical world, is an entirely different thing where
the parts literally 'grind' themselves 'to fit'.)
Can a motherboard improve operationally
with a little bit of crunch/processing time? Can this initial working
period be helpful in 'seating' things a bit?
Click to expand...

It is possible for heatsink thermal compound to settle, especially if it
was improperly applied, and connector contacts do 'creep' with temperature
changes. But, Murphy's Law, corollary 16: these things almost always
migrate to the worse condition and seldom the better (Actually, Murphy is
more absolutist about it).

Phase change thermal pads DO need some 'heat' to set properly but it should
do so under semi-normal operation.

The reason I'm asking is that I had a miserable time getting a new
motherboard to work at 166 Mhz FSU. I sent it back to the vendor and
received the replacement Friday night. When I first started using it,
almost immediately I experienced instances where the mouse froze up but
windows did not. The scanner was just getting garbage. Software was
starting up all by itself. I also had a few lockups while processing video
files (encoding). I thought 'here we go again', and was pretty disappointed
(again) after only 4 hours. After a little bit of shut down time, I started
it up again before I went to bed. It's been running for 36 hours since
then, encoding video without any problems at all. Even my scanner works
properly now when it didn't at first. Incredible! Everything seems to be
running just fine now.
Click to expand...

Well, considering its October, one might as well attribute it to ghosts and
goblins as opposed to 'burn in'. I mean, they're both used to 'explain' the
'unexplained', right? <g>

Lord knows what state the software was in after being confused by the
previous board and lord knows if the magical Windows XP repaired itself
(that's a 'feature' of it), assuming we're talking about XP.
It doesn't make any sense to me but it's almost like the motherboard/CPU
needed a short 'break-in' period to operate properly. Truth, or is it just
me looking for a (il)logical 'splanation? :)
Click to expand...

Don't restrict yourself to just one illogical explanation. There's plenty
 
As David Maynard has accurately stated, computer hobbyists
masquerading as experts have distorted the concept of
burn-in. Many have assumed, using word association, that
running diagnostics repeatedly is burn-in. Burn-in was always
about running systems at extremes, such as temperature. For
example, any acceptable computer works just fine in a 100
degree F room. Any minimally acceptable burn-in test would
test everything worst case - including operation in a 100
degree room.

To demonstrate the concept. Many will test memory in a
memory tester and declare it good. Bad test. Test can only
say when memory is bad. However if that memory is tested when
at 100+ degrees F, then intermittent defects are exposed.
Memory that passes diagnostics at 70 degrees can fail when at
100 degrees F. That memory is 100% defective and will
probably be failing months or years later even at 70 degrees.

Burn-in testing finds failures before they happen. Burn-in
testing typically also includes multiple temperature cycling
with system powered off and less extreme temperature cycling
with power applied.

Once had a motherboard that was intermittent. Traced the
problem directly to a cache memory chip by selectively heating
only that chip with a hairdryer. A hairdryer on high heat
(normal temperature for any semiconductor) made the
functioning chip fail during diagnostics. Heat is a powerful
tool to find defective electronics. Others instead try to
cure those symptoms with more fans and big heatsinks.
 
Make sure the power supply is working 100% OK.
Borrow a good supply from a friend to make sure.
 
Russ M. said:
This is probably going to sound like a dumb question to most, but I
have to ask it anyway. I've heard of 'burn-in' from years past, but I
always thought of it as a testing process. Can a motherboard improve
operationally with a little bit of crunch/processing time? Can this
initial working period be helpful in 'seating' things a bit?
Click to expand...

My understanding coincides with the dictionary definition.

http://dictionary.reference.com/search?q=burn-in period

burn-in period
1. A factory test designed to catch systems with marginal
components before they get out the door; the theory is that
burn-in will protect customers by outwaiting the steepest part
of the bathtub curve (see infant mortality).

Systems are different than components. Testing a system at component
extremes probably isn't a good idea. Big companies probably run systems
at higher than normal temperatures and maybe at slightly faster speeds,
to determine whether the system fails (like voltage level or timing
problems), to determine whether components are sensitive enough and fast
enough to continue doing their job, not looking for whether individual
components catastrophically fail.

In other words, my guess is that you do not need to stress components in
a system to near failure (near their maximum ratings) in order to
determine whether the system is high-quality.

I suppose nowadays electronic design automation (EDA) programs can take
care of that system failure stuff, given correctly specified components
(however marginal).
 
Russ said:
This is probably going to sound like a dumb question to most, but I have to
ask it anyway. I've heard of 'burn-in' from years past, but I always
thought of it as a testing process. Can a motherboard improve operationally
with a little bit of crunch/processing time? Can this initial working
period be helpful in 'seating' things a bit?

The reason I'm asking is that I had a miserable time getting a new
motherboard to work at 166 Mhz FSU.
Click to expand...

Did you change the motherboard but not reinstall the OS?
 
Did you change the motherboard but not reinstall the OS?
I didn't reinstall the OS. Could that be the reason it works after an
initial period be because the OS was 'adjusting' to problems and
self-correcting?

I actually tried to do a maintenance re-install today, but because I
installed service packs previously, it doesn't recognize it as Win2K from
the master CD. I may have to do a fresh install and re-install all my
progs.

Any thoughts about this?
 
Russ said:
I didn't reinstall the OS. Could that be the reason it works after an
initial period be because the OS was 'adjusting' to problems and
self-correcting?

I actually tried to do a maintenance re-install today, but because I
installed service packs previously, it doesn't recognize it as Win2K from
the master CD. I may have to do a fresh install and re-install all my
progs.

Any thoughts about this?
Click to expand...

Installing service packs doesn't make it 'not win2k'. Your CD should find it.

Did you say 'no' to the first set of 'repair' questions (repair console or
recover floppy) and go on as if doing a fresh install? It didn't then find
an existing Win2K installation and ask if you wanted to repair it?

If not then something is seriously corrupted. Try a chkdsk.
 
Russ said:
I didn't reinstall the OS. Could that be the reason it works after an
initial period be because the OS was 'adjusting' to problems and
self-correcting?
Click to expand...

I had something similar happen. After all I did to try and troubleshoot
the problem and corrective measures, I was about to reformat. Then, it
worked even though I did nothing new. This was after SP2 (and removing
the unneccessary files) and changing video cards. I still don't know
what got it working again. It's very frustrating.
 
This is probably going to sound like a dumb question to most, but I have to
ask it anyway. I've heard of 'burn-in' from years past, but I always
thought of it as a testing process. Can a motherboard improve operationally
with a little bit of crunch/processing time? Can this initial working
period be helpful in 'seating' things a bit?
Click to expand...

No.



--
Free Windows/PC help,
http://www.geocities.com/sheppola/trouble.html
remove obvious to reply
email (e-mail address removed)
Free songs to download and,"BURN" :O)
http://www.soundclick.com/bands/8/nomessiahsmusic.htm
 
so get out your hair dryers and test that board.............
Click to expand...

JAD, heh maybe not so bad idea... :-))))

see: ... [link] another alternative way ... (on my site under
comp/Testing Stability & under Some CPU Burn-in
[link] pre"burn-in" it in the pot on the stove! :-))
 
Back
Top