WIN98 SE Shutting down for no reason

wrote:
On Dec 19, 5:06 pm, "FromTheRafters"
wrote:
Which is exactly why reseating connectors is a good troubleshooting
procedure. I was always told not to use an eraser or steel wool to
clean contacts (it only wears them out more than need be. A cotton
swab with the proper solvent will remove crud).

Cleaning contacts means diagnostic information provided by a
normally corroded contact is lost. Curing symptoms only makes a defect
harder to locate. If contacts create a failure, then corrosion helps
find the real defect.

Notice what was posted previously and again. First collect facts
before reseating, etc. Reseating does not fix it. Reseating is
simply another symptom to track down a real defect. Unfortunately
some techs remain so naive as to even use an eraser or, as you said,
wear down the contacts in some naive belief that dirty means failure.

Connector technology is well proven and well understood from
manufacturer app notes. Educating technicians is part of the job.
Smarter technicians catch on quick. Connectors are self cleaning. If
cleaning with solvent stops a failure, the defect still exists.

Cited was an example that any tech with experience long ago
understood. GM had computer failures due to cheap connectors.
Replacing computers would (temporarily) fix the failure - clean the
contacts. Even inferior connectors are self cleaning.

A connector inside a computer, fixed by cleaning, means a problem
exists elsewhere. Technicians cleaning contacts with an eraser (or
even worse, steel wool - which quickly got one tech removed) need
reeducation or a new assignment. We could never put up with 'eraser
repairs' because anything that went out the door had to always work.

More amazing is a fifty some year old tech whose only technical
proof is a hysterical declaration - a denial and nothing more. Who
still did not learn basic connector concepts. Who would post replies
that are attacks rather than explain the science. He still has not
explained by why routine design permits corrosion problems without
failure. Proper design makes connector corrosion irrelevant.

Any design must work fine with normal contact corrosion. And then
that corrosion is eliminated by reseating.

Let's see. How many ohms resistance is created by corrosion? How
many volts exist across that corroded contact? Why did others not
provide numbers? Those who are curing problems with an eraser just
know. Junk science requires one to know without any numbers. It
looks dirty. Therefore cleaning with an eraser must fix it. Classic
junk science reasoning. Where does he provide any numbers to prove
his point? Oh. He just knows that contacts need routine cleaning
with an eraser.

Yes, making a change to modify symptoms of a failure will help
isolate the actual defect. However, once that change is made, other
important facts are lost. A naive tech will fix it by cleaning
contact with an eraser rather than first locate the defect. If it
works, then an eraser fixed it? Nonsense. That is a first thing
that a first year tech learns to not do - waste time cleaning
contacts.

If reseating does not clean contacts, then we have a serious design
defect AND eliminate that connector on everything. Every minimally
acceptable connector is self cleaning. Amazing that a tech that old
still did not learn these basic concepts. Junk science is alive and
well.
Perhaps you could explain this: a problem is 'corrected' by removing and
reseating the ram sticks. It is still working fine 2 yrs down the road. What
did doing the above hurt, except saving the user a 'bunch' of money?
You're idea of taking everything to a tech is bs.

wrote:
On Dec 19, 5:06 pm, "FromTheRafters"
wrote:
Which is exactly why reseating connectors is a good troubleshooting
procedure. I was always told not to use an eraser or steel wool to
clean contacts (it only wears them out more than need be. A cotton
swab with the proper solvent will remove crud).

Cleaning contacts means diagnostic information provided by a
normally corroded contact is lost. Curing symptoms only makes a defect
harder to locate. If contacts create a failure, then corrosion helps
find the real defect.

Notice what was posted previously and again. First collect facts
before reseating, etc. Reseating does not fix it. Reseating is
simply another symptom to track down a real defect. Unfortunately
some techs remain so naive as to even use an eraser or, as you said,
wear down the contacts in some naive belief that dirty means failure.

Connector technology is well proven and well understood from
manufacturer app notes. Educating technicians is part of the job.
Smarter technicians catch on quick. Connectors are self cleaning. If
cleaning with solvent stops a failure, the defect still exists.

Cited was an example that any tech with experience long ago
understood. GM had computer failures due to cheap connectors.
Replacing computers would (temporarily) fix the failure - clean the
contacts. Even inferior connectors are self cleaning.

A connector inside a computer, fixed by cleaning, means a problem
exists elsewhere. Technicians cleaning contacts with an eraser (or
even worse, steel wool - which quickly got one tech removed) need
reeducation or a new assignment. We could never put up with 'eraser
repairs' because anything that went out the door had to always work.

More amazing is a fifty some year old tech whose only technical
proof is a hysterical declaration - a denial and nothing more. Who
still did not learn basic connector concepts. Who would post replies
that are attacks rather than explain the science. He still has not
explained by why routine design permits corrosion problems without
failure. Proper design makes connector corrosion irrelevant.

Any design must work fine with normal contact corrosion. And then
that corrosion is eliminated by reseating.

Let's see. How many ohms resistance is created by corrosion? How
many volts exist across that corroded contact? Why did others not
provide numbers? Those who are curing problems with an eraser just
know. Junk science requires one to know without any numbers. It
looks dirty. Therefore cleaning with an eraser must fix it. Classic
junk science reasoning. Where does he provide any numbers to prove
his point? Oh. He just knows that contacts need routine cleaning
with an eraser.

Yes, making a change to modify symptoms of a failure will help
isolate the actual defect. However, once that change is made, other
important facts are lost. A naive tech will fix it by cleaning
contact with an eraser rather than first locate the defect. If it
works, then an eraser fixed it? Nonsense. That is a first thing
that a first year tech learns to not do - waste time cleaning
contacts.

If reseating does not clean contacts, then we have a serious design
defect AND eliminate that connector on everything. Every minimally
acceptable connector is self cleaning. Amazing that a tech that old
still did not learn these basic concepts. Junk science is alive and
well.
Perhaps you could explain this: a problem is 'corrected' by removing and
reseating the ram sticks. It is still working fine 2 yrs down the road. What
did doing the above hurt, except saving the user a 'bunch' of money?
You're idea of taking everything to a tech is bs.

Answer the question placed, don't walk to another just because you presume
to have found support... why do you believe environmentally controlled areas
are required for server farms and the like?

DOCUMENT the supposed claim you made. Submit it to EVERY IT professional and
professional maintenance party across the world... they would LOVE to have
you proof that connectors are self cleaning and they needn't bother with
regular maintenance regarding them. The FAA and military would also love
this exacting proof.

--
MEB
http://peoplescounsel.org
a Peoples' counsel

Windows diagnostics/Security/Networking:
http://peoplescounsel.org/ref/windows-main.htm
_ _
~~
wrote in message
...
On Dec 19, 5:06 pm, "FromTheRafters"
wrote:
Which is exactly why reseating connectors is a good troubleshooting
procedure. I was always told not to use an eraser or steel wool to
clean contacts (it only wears them out more than need be. A cotton
swab with the proper solvent will remove crud).

Cleaning contacts means diagnostic information provided by a
normally corroded contact is lost. Curing symptoms only makes a defect
harder to locate. If contacts create a failure, then corrosion helps
find the real defect.

Notice what was posted previously and again. First collect facts
before reseating, etc. Reseating does not fix it. Reseating is
simply another symptom to track down a real defect. Unfortunately
some techs remain so naive as to even use an eraser or, as you said,
wear down the contacts in some naive belief that dirty means failure.

Connector technology is well proven and well understood from
manufacturer app notes. Educating technicians is part of the job.
Smarter technicians catch on quick. Connectors are self cleaning. If
cleaning with solvent stops a failure, the defect still exists.

Cited was an example that any tech with experience long ago
understood. GM had computer failures due to cheap connectors.
Replacing computers would (temporarily) fix the failure - clean the
contacts. Even inferior connectors are self cleaning.

A connector inside a computer, fixed by cleaning, means a problem
exists elsewhere. Technicians cleaning contacts with an eraser (or
even worse, steel wool - which quickly got one tech removed) need
reeducation or a new assignment. We could never put up with 'eraser
repairs' because anything that went out the door had to always work.

More amazing is a fifty some year old tech whose only technical
proof is a hysterical declaration - a denial and nothing more. Who
still did not learn basic connector concepts. Who would post replies
that are attacks rather than explain the science. He still has not
explained by why routine design permits corrosion problems without
failure. Proper design makes connector corrosion irrelevant.

Any design must work fine with normal contact corrosion. And then
that corrosion is eliminated by reseating.

Let's see. How many ohms resistance is created by corrosion? How
many volts exist across that corroded contact? Why did others not
provide numbers? Those who are curing problems with an eraser just
know. Junk science requires one to know without any numbers. It
looks dirty. Therefore cleaning with an eraser must fix it. Classic
junk science reasoning. Where does he provide any numbers to prove
his point? Oh. He just knows that contacts need routine cleaning
with an eraser.

Yes, making a change to modify symptoms of a failure will help
isolate the actual defect. However, once that change is made, other
important facts are lost. A naive tech will fix it by cleaning
contact with an eraser rather than first locate the defect. If it
works, then an eraser fixed it? Nonsense. That is a first thing
that a first year tech learns to not do - waste time cleaning
contacts.

If reseating does not clean contacts, then we have a serious design
defect AND eliminate that connector on everything. Every minimally
acceptable connector is self cleaning. Amazing that a tech that old
still did not learn these basic concepts. Junk science is alive and
well.

Answer the question placed, don't walk to another just because you presume
to have found support... why do you believe environmentally controlled areas
are required for server farms and the like?

DOCUMENT the supposed claim you made. Submit it to EVERY IT professional and
professional maintenance party across the world... they would LOVE to have
you proof that connectors are self cleaning and they needn't bother with
regular maintenance regarding them. The FAA and military would also love
this exacting proof.

--
MEB
http://peoplescounsel.org
a Peoples' counsel

Windows diagnostics/Security/Networking:
http://peoplescounsel.org/ref/windows-main.htm
_ _
~~
wrote in message
...
On Dec 19, 5:06 pm, "FromTheRafters"
wrote:
Which is exactly why reseating connectors is a good troubleshooting
procedure. I was always told not to use an eraser or steel wool to
clean contacts (it only wears them out more than need be. A cotton
swab with the proper solvent will remove crud).

Cleaning contacts means diagnostic information provided by a
normally corroded contact is lost. Curing symptoms only makes a defect
harder to locate. If contacts create a failure, then corrosion helps
find the real defect.

Notice what was posted previously and again. First collect facts
before reseating, etc. Reseating does not fix it. Reseating is
simply another symptom to track down a real defect. Unfortunately
some techs remain so naive as to even use an eraser or, as you said,
wear down the contacts in some naive belief that dirty means failure.

Connector technology is well proven and well understood from
manufacturer app notes. Educating technicians is part of the job.
Smarter technicians catch on quick. Connectors are self cleaning. If
cleaning with solvent stops a failure, the defect still exists.

Cited was an example that any tech with experience long ago
understood. GM had computer failures due to cheap connectors.
Replacing computers would (temporarily) fix the failure - clean the
contacts. Even inferior connectors are self cleaning.

A connector inside a computer, fixed by cleaning, means a problem
exists elsewhere. Technicians cleaning contacts with an eraser (or
even worse, steel wool - which quickly got one tech removed) need
reeducation or a new assignment. We could never put up with 'eraser
repairs' because anything that went out the door had to always work.

More amazing is a fifty some year old tech whose only technical
proof is a hysterical declaration - a denial and nothing more. Who
still did not learn basic connector concepts. Who would post replies
that are attacks rather than explain the science. He still has not
explained by why routine design permits corrosion problems without
failure. Proper design makes connector corrosion irrelevant.

Any design must work fine with normal contact corrosion. And then
that corrosion is eliminated by reseating.

Let's see. How many ohms resistance is created by corrosion? How
many volts exist across that corroded contact? Why did others not
provide numbers? Those who are curing problems with an eraser just
know. Junk science requires one to know without any numbers. It
looks dirty. Therefore cleaning with an eraser must fix it. Classic
junk science reasoning. Where does he provide any numbers to prove
his point? Oh. He just knows that contacts need routine cleaning
with an eraser.

Yes, making a change to modify symptoms of a failure will help
isolate the actual defect. However, once that change is made, other
important facts are lost. A naive tech will fix it by cleaning
contact with an eraser rather than first locate the defect. If it
works, then an eraser fixed it? Nonsense. That is a first thing
that a first year tech learns to not do - waste time cleaning
contacts.

If reseating does not clean contacts, then we have a serious design
defect AND eliminate that connector on everything. Every minimally
acceptable connector is self cleaning. Amazing that a tech that old
still did not learn these basic concepts. Junk science is alive and
well.

On Dec 21, 3:11*am, "MEB" meb@not wrote:
Answer the question placed, don't walk to another just because you presume
to have found support... why do you believe environmentally controlled areas
are required for server farms and the like?

So you cannot dispute that electronic design includes normal
corrosion in the design parameters. Normal contact corrosion must
never cause electronics failures. AND that connectors are also self
cleaning - making corrosion further irrelevant. OK. Now we have an
agreement.

Do you walk into a server farm in white lab coats, hair nets, or
masks? That was required in environmentally controlled areas such as
aerospace equipment that also must work just fine in far harsher
environments. So what? What is your point other than to ignore what
every responsible connector manufacturer provides in spec sheets and
application notes? A tech caught cleaning contacts with an eraser to
fix electronics will quickly find himself in retraining or
reassigned.

Cleaning contacts with an eraser provides the OP with no useful
solution to his problem. But it does promote another classic urban
myth.

On Dec 21, 3:11*am, "MEB" meb@not wrote:
Answer the question placed, don't walk to another just because you presume
to have found support... why do you believe environmentally controlled areas
are required for server farms and the like?

So you cannot dispute that electronic design includes normal
corrosion in the design parameters. Normal contact corrosion must
never cause electronics failures. AND that connectors are also self
cleaning - making corrosion further irrelevant. OK. Now we have an
agreement.

Do you walk into a server farm in white lab coats, hair nets, or
masks? That was required in environmentally controlled areas such as
aerospace equipment that also must work just fine in far harsher
environments. So what? What is your point other than to ignore what
every responsible connector manufacturer provides in spec sheets and
application notes? A tech caught cleaning contacts with an eraser to
fix electronics will quickly find himself in retraining or
reassigned.

Cleaning contacts with an eraser provides the OP with no useful
solution to his problem. But it does promote another classic urban
myth.

In message
,
writes:
Yes, as stated previously, some problems such as a missing AC noise
filters will not be detected. But those are outside the context of
this discussion. Multimeter also will not detect a stolen
presidential election. That too is outside the context. A
multimeter will always locate a supply defect that causes the system
to not boot or work reliably. The meter either says the supply is
perfectly good - move on to other suspect. Or the meter says this
power supply is defective - must be replaced. Both are definitive
answers.

Meter will identify a supply producing low (or high) voltages, if the
supply is doing so steadily. No argument there.

What it will _not_ identify is a supply with an _intermittent_ fault -
either a loose connection, or one that only shows up under certain
circumstances of load (such as a drive starting up for the 12V line,
though that's less critical).

As stated both from what is being measured AND from decades of
experience, that 'fast pulse' is detected if the supply is
insufficient. There is no one fast pulse. If that fast pulse exists.
It exists repeatedly. Any fast pulse that a power supply can (and
must) handle is made irrelevant by low pass filters. If the fast
pulse is problematic, then a multimeter will report it. If a power
supply is properly sized, the fast pulse is normal and acceptable
operation - does not create 'out of spec' voltages - does not crash a
computer.

If by "fast pulse" you mean a sudden surge of demand, then yes, _if_ you
can make the computer do whatever it is doing (to cause that surge) when
it fails, and do it repeatedly and rapidly, then yes, you might see the
problem on your meter. However, a lot of the time, (a) it is something
that might only occur after half an hour's use, (b) the user may very
well not know what exactly the computer was doing when it fails - it may
happen when unattended. For such a rare occurrence, even if you happen
to be looking at a meter when it happens, if the failure causes the
computer to reboot or freeze - which it often does - then the demand
taken by the computer will change, usually dropping to a low level
(except on the 12V line if it causes a reboot which may make drives
spin), thus removing the cause of the problem. Until next time.

Of course, if you test the power supply by disconnecting it from the
computer and connecting it to a load box that draws the maximum rated
current on all rails, and while doing that measure the voltages, this
won't apply, but most repair facilities don't have such a load (not
least because it's a big thing). [It would also not _completely_ test
the supply because it wouldn't test _dynamic_ regulation, but it _would_
at least make sure any failure is not due to the supply just not being
able to supply its rated load.]

Experience AND the concepts behind this diagnostic procedure
demonstrate why smarter (better trained) techs use the meter and avoid
shotgunning. If the meter was not detecting all relevant failures,

You keep using this term "shotgunning", but you haven't said what you
mean by it yet.

then why was it performing reliably - identifying defects 100% of the
time?

Because you've been lucky, or have blamed other failures on something
else?

How to find those failing (bulging) capacitors before capacitors
cause computer failure? The multimeter. Yes, a multimeter would not
detect those defective capacitors before the defect became apparent.
No measureable defect existed yet. So yes, one can complain that
'what will become defective' capacitors were not detected by the
meter. But the discussion here is how to find currently defective
power supplies; also before a defect becomes worse and crashes a
computer. Locating a defect in but minutes or seconds. Solving

If the caps in the supply have gone that far, then yes. (They don't all
bulge; caps can lose capacity and show no outward sign.)

computer problems maybe 5 times faster than what can be achieved by
shotgunning.

That term again (-:

Clearly, the meter is a superior solution compared to the so popular
method called shotgunning - for multiple reasons already provided.

And again

2.5 digits (analog meter) is insufficient. 3.5 digits (a standard
multimeter) is necessary because of what is being measured.

I very much doubt you'd be able to read an analogue meter to 2.5 digits!
Except near full-scale, and probably with a large mirror scale. I'd say
2.5 digits - as long as the first isn't 0, i. e. as long as the meter
has enough ranges - is enough for this purpose; the noise on a PC's
power rails means anything finer is pointless anyway. However, it is
academic - as you say, meters you can get are probably mostly 3 digits
anyway nowadays.

Shotgunning - replacing parts on spec? I don't now what that

Now you see how I feel ... (-:

means. If a power supply is out of spec (relevant parameters), then a
multimeter should discover it. Shotgunning is replacing a power
supply regardless of whether the original supply is or is not 'in

Ah, at last; I _thought_ that was what you meant, but nice to have it
confirmed.

spec'. Shotgunning is fixing something only on wild speculation.
Computer does not start. Wild speculations – automatically blame and
replace the power supply – is shotgunning.
[]
I see - I suppose it is by analogy of shooting with a shotgun, on the
basis that one pellet might hit.

(The expression I used above - "on spec" - is a [perhaps British?]
expression meaning much the same; I'm not sure where it comes from -
possibly "speculation", certainly not "specification".)

Now, connectors. I wasn't going to join in this argument, but I can't
resist! My main job function at the moment involves supporting old
equipment (mostly avionics, not PCs), often decades old.

AMP, by the way, are now part of Tyco electronics (not to be confused
with Tyco).

Connectors _are_ self-cleaning _to some extent_: the wiping action of
them being inserted and removed will often clean off some of whatever is
preventing them from making good contact. Once made and static, this of
course does not apply. _Usually_, _if_ the mating surfaces were springy
and clean enough in the first place, they should remain making good
contact. However, springiness does reduce in time, and of course there
_is_ vibration in PCs.

Removing and reseating _will_ often scrape through the contamination and
restore a good contact. Whether this has solved the problem or just "got
it working" is a matter for argument; it _hasn't_ fixed the source of
corrosion if there is one (a smoker in the room?), it _hasn't_ restored
and spring pressure that is missing. Conversely, it _may_ make the unit
continue to work for the remaining life of the equipment - it depends
how sure the customer wants to be, and how much they are willing to pay.
I would make sure they are told that that was what was done, in the case
of a PC. (For avionics, quite apart from the safety considerations
anyway, I'd want to check on the springiness, because they are subject
to a lot more vibration. But if that means replacement of the connector,
that means quite a bit of downtime - assuming a new connector of the
same type can still be obtained at all.)

Removing and reseating _will_ also wear the contacts - as will cleaning
them (whether with wire wool or an eraser; a solvent is better).
However, the benefit may outweigh the disadvantage; if the contamination
can clearly be seen and removed, this probably _is_ a solution.

Connectors are _designed_ to be re-made several times. For DIN41612 and
D-type connectors, it basically depends on the thickness of the gold
plating: for the highest grade (either III or I, I can't remember which
way it goes), they have about 5 microns of gold, which is good for
several hundred mating cycles. The lowest grade are just gold "flashed",
and generally those are considered to have no more than about 5 cycles
in them. For things not intended to be plugged in and out many times,
grade II or the lowest is probably adequate, and certainly cheaper (by
quite a bit). Things like peripheral cards for PCs probably have such a
coating, I don't know; one would hope that the _socket_ connections have
a thicker coating, likewise the drive connectors and any other part that
is likely to be disturbed several times in its lifetime. I have also
seen cards (especially memory modules) that have "bright tin" finish
rather than gold; again, this is valid for some uses.

Basically, what I'm saying is: let's be sensible about this, and not get
on our high horses and say "anybody who does xxx should be
shot/sacked/whatever". All techniques - plugging/unplugging, cleaning
(though not I would say with wire brush!), and even the dreaded
replacing-on-spec ("shotgunning"?) have their place. There are times
when "get it working no matter what you do - but quickly" applies; there
are times when "get it working but spend as little as possible" applies;
and there are times - few, I suspect - when "_fix_ it, and tell me what
was wrong so I can reduce the chance of it happening again" applies.
--
J. P. Gilliver. UMRA: 1960/1985 MB++G.5AL(+++)IS-P--Ch+(p)Ar+T[?]H+Sh0!:`)DNAf
** http://www.soft255.demon.co.uk/G6JPG-PC/JPGminPC.htm for ludicrously
outdated thoughts on PCs. **

"Forget computers; it's hard enough getting humans to pass the Turing test."
- David Bedno

In message
,
writes:
Yes, as stated previously, some problems such as a missing AC noise
filters will not be detected. But those are outside the context of
this discussion. Multimeter also will not detect a stolen
presidential election. That too is outside the context. A
multimeter will always locate a supply defect that causes the system
to not boot or work reliably. The meter either says the supply is
perfectly good - move on to other suspect. Or the meter says this
power supply is defective - must be replaced. Both are definitive
answers.

Meter will identify a supply producing low (or high) voltages, if the
supply is doing so steadily. No argument there.

What it will _not_ identify is a supply with an _intermittent_ fault -
either a loose connection, or one that only shows up under certain
circumstances of load (such as a drive starting up for the 12V line,
though that's less critical).

As stated both from what is being measured AND from decades of
experience, that 'fast pulse' is detected if the supply is
insufficient. There is no one fast pulse. If that fast pulse exists.
It exists repeatedly. Any fast pulse that a power supply can (and
must) handle is made irrelevant by low pass filters. If the fast
pulse is problematic, then a multimeter will report it. If a power
supply is properly sized, the fast pulse is normal and acceptable
operation - does not create 'out of spec' voltages - does not crash a
computer.

If by "fast pulse" you mean a sudden surge of demand, then yes, _if_ you
can make the computer do whatever it is doing (to cause that surge) when
it fails, and do it repeatedly and rapidly, then yes, you might see the
problem on your meter. However, a lot of the time, (a) it is something
that might only occur after half an hour's use, (b) the user may very
well not know what exactly the computer was doing when it fails - it may
happen when unattended. For such a rare occurrence, even if you happen
to be looking at a meter when it happens, if the failure causes the
computer to reboot or freeze - which it often does - then the demand
taken by the computer will change, usually dropping to a low level
(except on the 12V line if it causes a reboot which may make drives
spin), thus removing the cause of the problem. Until next time.

Of course, if you test the power supply by disconnecting it from the
computer and connecting it to a load box that draws the maximum rated
current on all rails, and while doing that measure the voltages, this
won't apply, but most repair facilities don't have such a load (not
least because it's a big thing). [It would also not _completely_ test
the supply because it wouldn't test _dynamic_ regulation, but it _would_
at least make sure any failure is not due to the supply just not being
able to supply its rated load.]

Experience AND the concepts behind this diagnostic procedure
demonstrate why smarter (better trained) techs use the meter and avoid
shotgunning. If the meter was not detecting all relevant failures,

You keep using this term "shotgunning", but you haven't said what you
mean by it yet.

then why was it performing reliably - identifying defects 100% of the
time?

Because you've been lucky, or have blamed other failures on something
else?

How to find those failing (bulging) capacitors before capacitors
cause computer failure? The multimeter. Yes, a multimeter would not
detect those defective capacitors before the defect became apparent.
No measureable defect existed yet. So yes, one can complain that
'what will become defective' capacitors were not detected by the
meter. But the discussion here is how to find currently defective
power supplies; also before a defect becomes worse and crashes a
computer. Locating a defect in but minutes or seconds. Solving

If the caps in the supply have gone that far, then yes. (They don't all
bulge; caps can lose capacity and show no outward sign.)

computer problems maybe 5 times faster than what can be achieved by
shotgunning.

That term again (-:

Clearly, the meter is a superior solution compared to the so popular
method called shotgunning - for multiple reasons already provided.

And again

2.5 digits (analog meter) is insufficient. 3.5 digits (a standard
multimeter) is necessary because of what is being measured.

I very much doubt you'd be able to read an analogue meter to 2.5 digits!
Except near full-scale, and probably with a large mirror scale. I'd say
2.5 digits - as long as the first isn't 0, i. e. as long as the meter
has enough ranges - is enough for this purpose; the noise on a PC's
power rails means anything finer is pointless anyway. However, it is
academic - as you say, meters you can get are probably mostly 3 digits
anyway nowadays.

Shotgunning - replacing parts on spec? I don't now what that

Now you see how I feel ... (-:

means. If a power supply is out of spec (relevant parameters), then a
multimeter should discover it. Shotgunning is replacing a power
supply regardless of whether the original supply is or is not 'in

Ah, at last; I _thought_ that was what you meant, but nice to have it
confirmed.

spec'. Shotgunning is fixing something only on wild speculation.
Computer does not start. Wild speculations – automatically blame and
replace the power supply – is shotgunning.
[]
I see - I suppose it is by analogy of shooting with a shotgun, on the
basis that one pellet might hit.

(The expression I used above - "on spec" - is a [perhaps British?]
expression meaning much the same; I'm not sure where it comes from -
possibly "speculation", certainly not "specification".)

Now, connectors. I wasn't going to join in this argument, but I can't
resist! My main job function at the moment involves supporting old
equipment (mostly avionics, not PCs), often decades old.

AMP, by the way, are now part of Tyco electronics (not to be confused
with Tyco).

Connectors _are_ self-cleaning _to some extent_: the wiping action of
them being inserted and removed will often clean off some of whatever is
preventing them from making good contact. Once made and static, this of
course does not apply. _Usually_, _if_ the mating surfaces were springy
and clean enough in the first place, they should remain making good
contact. However, springiness does reduce in time, and of course there
_is_ vibration in PCs.

Removing and reseating _will_ often scrape through the contamination and
restore a good contact. Whether this has solved the problem or just "got
it working" is a matter for argument; it _hasn't_ fixed the source of
corrosion if there is one (a smoker in the room?), it _hasn't_ restored
and spring pressure that is missing. Conversely, it _may_ make the unit
continue to work for the remaining life of the equipment - it depends
how sure the customer wants to be, and how much they are willing to pay.
I would make sure they are told that that was what was done, in the case
of a PC. (For avionics, quite apart from the safety considerations
anyway, I'd want to check on the springiness, because they are subject
to a lot more vibration. But if that means replacement of the connector,
that means quite a bit of downtime - assuming a new connector of the
same type can still be obtained at all.)

Removing and reseating _will_ also wear the contacts - as will cleaning
them (whether with wire wool or an eraser; a solvent is better).
However, the benefit may outweigh the disadvantage; if the contamination
can clearly be seen and removed, this probably _is_ a solution.

Connectors are _designed_ to be re-made several times. For DIN41612 and
D-type connectors, it basically depends on the thickness of the gold
plating: for the highest grade (either III or I, I can't remember which
way it goes), they have about 5 microns of gold, which is good for
several hundred mating cycles. The lowest grade are just gold "flashed",
and generally those are considered to have no more than about 5 cycles
in them. For things not intended to be plugged in and out many times,
grade II or the lowest is probably adequate, and certainly cheaper (by
quite a bit). Things like peripheral cards for PCs probably have such a
coating, I don't know; one would hope that the _socket_ connections have
a thicker coating, likewise the drive connectors and any other part that
is likely to be disturbed several times in its lifetime. I have also
seen cards (especially memory modules) that have "bright tin" finish
rather than gold; again, this is valid for some uses.

Basically, what I'm saying is: let's be sensible about this, and not get
on our high horses and say "anybody who does xxx should be
shot/sacked/whatever". All techniques - plugging/unplugging, cleaning
(though not I would say with wire brush!), and even the dreaded
replacing-on-spec ("shotgunning"?) have their place. There are times
when "get it working no matter what you do - but quickly" applies; there
are times when "get it working but spend as little as possible" applies;
and there are times - few, I suspect - when "_fix_ it, and tell me what
was wrong so I can reduce the chance of it happening again" applies.
--
J. P. Gilliver. UMRA: 1960/1985 MB++G.5AL(+++)IS-P--Ch+(p)Ar+T[?]H+Sh0!:`)DNAf
** http://www.soft255.demon.co.uk/G6JPG-PC/JPGminPC.htm for ludicrously
outdated thoughts on PCs. **

"Forget computers; it's hard enough getting humans to pass the Turing test."
- David Bedno

Where did you find agreement with your claim in what was presented by me...

EVERY critical server area is environmentally controlled, due, in part, to
the massive and diverse types of pollutants now found in common air;
that's.the air across the planet, from pole to pole. Highly corrosive and
destructive; prior theoretical ideas or even proven mechanical
characteristics once thought as viable for protective qualities or lacking
need for such, are now found as needing protection.
From centuries old marble statues and paintings, to old rock and man-made
structures; all are being destroyed by the air borne pollutants. They need
no more than minimal exposure to *common air*.
That an engineer, such as you claim to be, would suggest that connectors
can be counted on to be *self cleaning* is absurd, moreover, shows how
out-of-touch with the real world many engineers are. Texts and teachings
which ignore these developments are at the very least wrong, and when
dealing with actual designed goods and/or structures and/or something as
simple as an electrical connector, are likely dangerous and hazardous, and,
when taken with human activities, life threatening. This would be CRITICAL,
certainly such in the airline industry, and definitely in the military.
But one can NOT stop there. Buildings, bridges and other structures which
man uses are ALL subject to these pollutants' destructive properties found
in the common air which we all breath and surrounds EVERYTHING [unless in a
vacuum].

When I suggested using an erasure, it wasn't the scrub the area with
repeated and destructive penetrating action, but a light application,
because one also finds air borne pollutants which contain dielectric
properties or cause such via chemical reaction. Standard thermal expansion
and contraction of connections can, over the course of time, allow this
film-like activity to break once viable connections.

So take some time and look OUTSIDE that box in which you have placed
yourself. There are two *worlds*, one in which we live, and the second -
theoretical or *the dreams of man*. Living in the real world and attempting
to understand it is difficult, but certainly provides more protection.
Living in a dream world may be more comfortable, but certainly provides
opportunity for abuse, dangerous situations, and false sense of security.

--
MEB
http://peoplescounsel.org
a Peoples' counsel

Windows diagnostics/Security/Networking:
http://peoplescounsel.org/ref/windows-main.htm
_ _
~~
wrote in message
...
On Dec 21, 3:11 am, "MEB" meb@not wrote:
Answer the question placed, don't walk to another just because you presume
to have found support... why do you believe environmentally controlled
areas
are required for server farms and the like?

So you cannot dispute that electronic design includes normal
corrosion in the design parameters. Normal contact corrosion must
never cause electronics failures. AND that connectors are also self
cleaning - making corrosion further irrelevant. OK. Now we have an
agreement.

Do you walk into a server farm in white lab coats, hair nets, or
masks? That was required in environmentally controlled areas such as
aerospace equipment that also must work just fine in far harsher
environments. So what? What is your point other than to ignore what
every responsible connector manufacturer provides in spec sheets and
application notes? A tech caught cleaning contacts with an eraser to
fix electronics will quickly find himself in retraining or
reassigned.

Cleaning contacts with an eraser provides the OP with no useful
solution to his problem. But it does promote another classic urban
myth.

Where did you find agreement with your claim in what was presented by me...

EVERY critical server area is environmentally controlled, due, in part, to
the massive and diverse types of pollutants now found in common air;
that's.the air across the planet, from pole to pole. Highly corrosive and
destructive; prior theoretical ideas or even proven mechanical
characteristics once thought as viable for protective qualities or lacking
need for such, are now found as needing protection.
From centuries old marble statues and paintings, to old rock and man-made
structures; all are being destroyed by the air borne pollutants. They need
no more than minimal exposure to *common air*.
That an engineer, such as you claim to be, would suggest that connectors
can be counted on to be *self cleaning* is absurd, moreover, shows how
out-of-touch with the real world many engineers are. Texts and teachings
which ignore these developments are at the very least wrong, and when
dealing with actual designed goods and/or structures and/or something as
simple as an electrical connector, are likely dangerous and hazardous, and,
when taken with human activities, life threatening. This would be CRITICAL,
certainly such in the airline industry, and definitely in the military.
But one can NOT stop there. Buildings, bridges and other structures which
man uses are ALL subject to these pollutants' destructive properties found
in the common air which we all breath and surrounds EVERYTHING [unless in a
vacuum].

When I suggested using an erasure, it wasn't the scrub the area with
repeated and destructive penetrating action, but a light application,
because one also finds air borne pollutants which contain dielectric
properties or cause such via chemical reaction. Standard thermal expansion
and contraction of connections can, over the course of time, allow this
film-like activity to break once viable connections.

So take some time and look OUTSIDE that box in which you have placed
yourself. There are two *worlds*, one in which we live, and the second -
theoretical or *the dreams of man*. Living in the real world and attempting
to understand it is difficult, but certainly provides more protection.
Living in a dream world may be more comfortable, but certainly provides
opportunity for abuse, dangerous situations, and false sense of security.

--
MEB
http://peoplescounsel.org
a Peoples' counsel

Windows diagnostics/Security/Networking:
http://peoplescounsel.org/ref/windows-main.htm
_ _
~~
wrote in message
...
On Dec 21, 3:11 am, "MEB" meb@not wrote:
Answer the question placed, don't walk to another just because you presume
to have found support... why do you believe environmentally controlled
areas
are required for server farms and the like?

So you cannot dispute that electronic design includes normal
corrosion in the design parameters. Normal contact corrosion must
never cause electronics failures. AND that connectors are also self
cleaning - making corrosion further irrelevant. OK. Now we have an
agreement.

Do you walk into a server farm in white lab coats, hair nets, or
masks? That was required in environmentally controlled areas such as
aerospace equipment that also must work just fine in far harsher
environments. So what? What is your point other than to ignore what
every responsible connector manufacturer provides in spec sheets and
application notes? A tech caught cleaning contacts with an eraser to
fix electronics will quickly find himself in retraining or
reassigned.

Cleaning contacts with an eraser provides the OP with no useful
solution to his problem. But it does promote another classic urban
myth.