Holy Moly -- Residual Electricity????

[Androcles]

: > : >
: > >> What caused the high temperature?
: >
: > > Being in a hot place.
: >
: > Then the failure occurred BEFORE the capacitor went bad.
: >
:
: Um, what? You mean just the fact that I *try* to
: operate a capacitor in a poorly air-conditioned room
: implies it has failed?

It is your claim that the failure has an environmental cause.
Do not operate capacitors in a poorly air-conditioned room,
they will fail (according to Poe). Um, yes.

 

[Androcles]

: > On Wed, 14 Nov 2007 10:07:49 -0800, Randy Poe <[email protected]>
: > wrote:
: >
: > >> What caused the high temperature?
: >
: > >Being in a hot place.
: >
: > > - Randy
: >
: > It would have to be pretty hot of at least above 120F tho I've seen
: > some hardware listed in the 140's F as maximum. My guess is the 2nd
: > applies here. Now the problem is what causes the high temperature to
: > begin with?? I have doubts about a bad fan but perhaps or maybe a
: > short circuit????
:
: This all seems incidental to the discussion.

Bullshit, it is your claim that capacitor failure has an environmental
cause, that is most pertinent to the discussion you contributed to.

 

[w_tom]

Then the failure occurred BEFORE the capacitor went bad.

Temperature has what relevance to the problem? Unfortunately, many
want to blame temperature when other functions are not understood.

Computers must work just fine even when room temperature is above
100 degree F. In fact, testing a computer in a 100 degree room is one
way to find defects because the defect results in failure.

Electrolytic capacitors inside power supplies are typically rated
for higher temperatures. One failure mode is Effective Series
Resistance or ESR (not Parallel Resistance). If chemistry inside the
electrolytic is failing, then series resistance increases. Higher
resistance means capacitor gets hotter - a feedback cycle that
eventually causes capacitor failure.

But this is not relevant to a power supply operating on 240 volts AC
when the power supply is rated for 240 volts.

Electrolytic failure is most often a manufacturing defect inside the
capacitor or excessive voltage applied to that capacitor. Capacitors
are routinely selected to withstand temperatures well above standard
room temperature. Its temperature rating is usually printed on that
electrolytic. Long before accepting what was posted, view those
numbers yourself.

 

[M.I.5¾]

Sounds like a short circuit to me. Perhaps he's connecting
them with the wrong polarity of the "certain type".

It's not a short circuit as such, just a lower resistance in parallel with
the capacitor. It's enough to prevent the switching circuit from switching.

 

[M.I.5¾]

How do you dream up all these ridiculous things?"M.I.5¾"

Two charts of electrolytic capacitor failure modes. Check out
Table 2 here:
http://industrial.panasonic.com/www-data/pdf/ABA0000/ABA0000TE4.pdf

Operating at high temperature is shown to cause failure
by increase in leakage current.

or Figure 2.10 here:
http://etd.gatech.edu/theses/available/etd-04082007-083102/unrestricted/imam_afroz_m_200705_phd.pdf

Operating at high temperature is shown to cause failure
by loss of effective resistance (i.e. increased leakage).

-----------

Indeed. In general electrolytic capacitors are rated at either 85°C or
105°C. The former is obviously cheaper than the latter and that's often
what gets used as such supplies don't run that warm. However, even if
operated below 85°C, the former type are considerably more unreliable than
the latter.

 

[M.I.5¾]

How do you dream up all these ridiculous things?

Apparently well to known to nearly everyone - but not you.

 

[Androcles]

:
: : > Sounds like a short circuit to me. Perhaps he's connecting
: > them with the wrong polarity of the "certain type".
: >
:
: It's not a short circuit as such, just a lower resistance in parallel

If you connect a copper wire or a solder splash across the cap
you'd be connecting a lower resistance in parallel, even copper
wire has some resistance. That's a short circuit by definition.

 

[M.I.5¾]

Temperature has what relevance to the problem? Unfortunately, many
want to blame temperature when other functions are not understood.

Computers must work just fine even when room temperature is above
100 degree F. In fact, testing a computer in a 100 degree room is one
way to find defects because the defect results in failure.

Electrolytic capacitors inside power supplies are typically rated
for higher temperatures. One failure mode is Effective Series
Resistance or ESR (not Parallel Resistance). If chemistry inside the
electrolytic is failing, then series resistance increases. Higher
resistance means capacitor gets hotter - a feedback cycle that
eventually causes capacitor failure.

Both ESR and EPR are present in every capacitor made. ESR is usually quoted
on spec sheets because every capacitor has this built in (it's a fact of the
way they are designed). EPR, hopefully is not present to any extent that
affects the operation of the capacitor. Although the ESR can rise, it is
not a particularly common failure mode and is usually more due to mechanical
construction than chemistry.

But this is not relevant to a power supply operating on 240 volts AC
when the power supply is rated for 240 volts.

On the contrary, it is relevant to any circuit using capacitors.

Electrolytic failure is most often a manufacturing defect inside the
capacitor or excessive voltage applied to that capacitor. Capacitors
are routinely selected to withstand temperatures well above standard
room temperature. Its temperature rating is usually printed on that
electrolytic. Long before accepting what was posted, view those
numbers yourself.

All capacitor made are not equal - it's a fact of manufacturing. About half
the capacitors will work better than they were designed to, and the other
half will not work as well. A few will have manufacturing features that
lead to some failure mode or other. An increase in the leakage (or a
reduction in the EPR) is but one such mode of failure (but the most common,
at least among electrolytic capacitors). In general, once the capacitor
starts to leak excessively, the leakage will get worse as time progresses.
It is a temperature dependant phenomenon.

It is of course possible to manufacture capacitors that conform more closely
to the intended design and have far fewer examples of failure (and indeed
such are made). However, if they were used to build computer power
supplies, nobody would be able to afford them. It's all a compromise
between reliability and cost.

 

[Androcles]

:
: : > How do you dream up all these ridiculous things?"M.I.5¾"
: >
: > : >
: >
: >
: > >: > >> What type of capacitor has to cool off before it accepts a charge?
: >
: > > A: A faulty one.
: >
: > > It is well a known problem particularly among certain types of
: > > electrolytic capacitors. The usual problem is that the EPR (Effective
: > > parallel resistance) of the capacitor falls alarmingly as it warms up
: > > rendering it ineffective as a capacitor.
:
: Two charts of electrolytic capacitor failure modes. Check out
: Table 2 here:
: http://industrial.panasonic.com/www-data/pdf/ABA0000/ABA0000TE4.pdf
:
: Operating at high temperature is shown to cause failure
: by increase in leakage current.
:
: or Figure 2.10 here:
:
http://etd.gatech.edu/theses/available/etd-04082007-083102/unrestricted/imam_afroz_m_200705_phd.pdf
:
: Operating at high temperature is shown to cause failure
: by loss of effective resistance (i.e. increased leakage).
:
: -----------
:
: Indeed. In general electrolytic capacitors are rated at either 85°C or
: 105°C. The former is obviously cheaper than the latter and that's often
: what gets used as such supplies don't run that warm. However, even if
: operated below 85°C, the former type are considerably more unreliable than
: the latter.

Oh, so the latter break down at temperatures below 85°C because
they are less reliable than the former cheaper ones... very logical.

 

[M.I.5¾]

It would have to be pretty hot of at least above 120F tho I've seen
some hardware listed in the 140's F as maximum. My guess is the 2nd
applies here. Now the problem is what causes the high temperature to
begin with?? I have doubts about a bad fan but perhaps or maybe a
short circuit????

The particular temperatures we are talking here are not unusual
temperatures, but the usual temperature excursions experienced by components
in any circuit while it is operating. Any operating circuit has current
flowing in its parts and this generates heat. The capacitors we are
concerned with are not experiencing heat outside of their design range, but
rather the capacitors are unable tolerate even this design heat because of
adverse manufacturing tolerances or possibly even that he capacitor has been
subject to a transient event that has reduced its ability to withstand such
heat (for example a voltage spike that partially breaks down the dielectric
layer).

 

[Unknown]

BS

Apparently well to known to nearly everyone - but not you.

 

[w_tom]

All capacitor made are not equal - it's a fact of manufacturing. About half
the capacitors will work better than they were designed to, and the other
half will not work as well.

Not quite correct. Almost all work at least as good as they were
designed for. Some will work far better. Cofidence levels - all
capacitors must at least meet specs - what they are intended to do.

The primary source of heating is ESR. ESR increases as the capactor
fails causing even more heat. And all that is completely irrelevant
to the OPs question. That computer must work just fine even when the
70 degree room rises to above 100 degree F. Again, what is the
temperature printed on that capacitor's label?

 

[M.I.5¾]

:
: : > Sounds like a short circuit to me. Perhaps he's connecting
: > them with the wrong polarity of the "certain type".
: >
:
: It's not a short circuit as such, just a lower resistance in parallel

If you connect a copper wire or a solder splash across the cap
you'd be connecting a lower resistance in parallel, even copper
wire has some resistance. That's a short circuit by definition.

So?

I doubt that the OP has done that.

 

[M.I.5¾]

Not quite correct. Almost all work at least as good as they were
designed for. Some will work far better. Cofidence levels - all
capacitors must at least meet specs - what they are intended to do.

You really don't know anything do you?

Capacitors are produced is such large quantities that the conformance to
specification is done by testing a sample of each batch only. It is a
fundamental requirement of sample testing that you have to accept that items
that do not conform to specification will be accepted. If you cannot accept
that then you can't do sample testing. Also, if you have a look at any
sample testing tables, you will see that in most of them failures are
allowed, but the batch can still be accepted. There are tables that allow
no failures, but that does not prevent failures from occuring in the
majority (untested) part of the batch.

The BS6001 series of standards would be a good starting point.

The primary source of heating is ESR. ESR increases as the capactor
fails causing even more heat. And all that is completely irrelevant
to the OPs question. That computer must work just fine even when the
70 degree room rises to above 100 degree F.

ESR is but one source of heating and usually (or should be) the most
insignificant. The primary sources of heating are all the heat disipating
components around the capacitor. ESR does not necessarily increase as the
capacitor fails, but a reduced EPR can also be a source of internally
generated heat, but only if the rest of the circuit permits enough current
to flow for it to become a factor.

Again, what is the temperature printed on that capacitor's label?

I have already addressed this point. Try reading the rest of the thread.

 

[M.I.5¾]

:
: : > How do you dream up all these ridiculous things?"M.I.5¾"
: >
: > : >
: >
: >
: > >: > >> What type of capacitor has to cool off before it accepts a charge?
: >
: > > A: A faulty one.
: >
: > > It is well a known problem particularly among certain types of
: > > electrolytic capacitors. The usual problem is that the EPR
(Effective
: > > parallel resistance) of the capacitor falls alarmingly as it warms
up
: > > rendering it ineffective as a capacitor.
:
: Two charts of electrolytic capacitor failure modes. Check out
: Table 2 here:
: http://industrial.panasonic.com/www-data/pdf/ABA0000/ABA0000TE4.pdf
:
: Operating at high temperature is shown to cause failure
: by increase in leakage current.
:
: or Figure 2.10 here:
:
http://etd.gatech.edu/theses/available/etd-04082007-083102/unrestricted/imam_afroz_m_200705_phd.pdf
:
: Operating at high temperature is shown to cause failure
: by loss of effective resistance (i.e. increased leakage).
:
: -----------
:
: Indeed. In general electrolytic capacitors are rated at either 85°C or
: 105°C. The former is obviously cheaper than the latter and that's often
: what gets used as such supplies don't run that warm. However, even if
: operated below 85°C, the former type are considerably more unreliable
than
: the latter.

Oh, so the latter break down at temperatures below 85°C because
they are less reliable than the former cheaper ones... very logical.

How you managed to arrive at that conclusion is anyone's guess.

 

[Androcles]

:
: : >
: > : > :
: > : : > : > Sounds like a short circuit to me. Perhaps he's connecting
: > : > them with the wrong polarity of the "certain type".
: > : >
: > :
: > : It's not a short circuit as such, just a lower resistance in parallel
: >
: > If you connect a copper wire or a solder splash across the cap
: > you'd be connecting a lower resistance in parallel, even copper
: > wire has some resistance. That's a short circuit by definition.
: >
:
: So?

So it is a short circuit of "the certain type" "as such", contrary to your
hand-waving waffle. What's the 5¾ for? Not your hat size, surely?

 

[Androcles]

:
: : >> All capacitor made are not equal - it's a fact of manufacturing. About
: >> half
: >> the capacitors will work better than they were designed to, and the
other
: >> half will not work as well.
: >
: > Not quite correct. Almost all work at least as good as they were
: > designed for. Some will work far better. Cofidence levels - all
: > capacitors must at least meet specs - what they are intended to do.
: >
:
: You really don't know anything do you?
:
: Capacitors are produced is such large quantities that the conformance to
: specification is done by testing a sample of each batch only. It is a
: fundamental requirement of sample testing that you have to accept that
items
: that do not conform to specification will be accepted. If you cannot
accept
: that then you can't do sample testing. Also, if you have a look at any
: sample testing tables,

Bwahahahahahah!
You really don't know anything about Quality Control, do you?

Do these words mean anything to you: Mean, Variance, Standard Deviation?

Tables! Got any sample testing chairs to go with them, size 5¾ ?
HAHAHAHAHAHA!

 

[Androcles]

:
: : >
: > : > :
: > : : > : > How do you dream up all these ridiculous things?"M.I.5¾"
: > : >
: > : > : > : >
: > : >
: > : >
: > : > >: > : > >> What type of capacitor has to cool off before it accepts a
charge?
: > : >
: > : > > A: A faulty one.
: > : >
: > : > > It is well a known problem particularly among certain types of
: > : > > electrolytic capacitors. The usual problem is that the EPR
: > (Effective
: > : > > parallel resistance) of the capacitor falls alarmingly as it warms
: > up
: > : > > rendering it ineffective as a capacitor.
: > :
: > : Two charts of electrolytic capacitor failure modes. Check out
: > : Table 2 here:
: > : http://industrial.panasonic.com/www-data/pdf/ABA0000/ABA0000TE4.pdf
: > :
: > : Operating at high temperature is shown to cause failure
: > : by increase in leakage current.
: > :
: > : or Figure 2.10 here:
: > :
: >
http://etd.gatech.edu/theses/available/etd-04082007-083102/unrestricted/imam_afroz_m_200705_phd.pdf
: > :
: > : Operating at high temperature is shown to cause failure
: > : by loss of effective resistance (i.e. increased leakage).
: > :
: > : -----------
: > :
: > : Indeed. In general electrolytic capacitors are rated at either 85°C
or
: > : 105°C. The former is obviously cheaper than the latter and that's
often
: > : what gets used as such supplies don't run that warm. However, even if
: > : operated below 85°C, the former type are considerably more unreliable
: > than
: > : the latter.
: >
: > Oh, so the latter break down at temperatures below 85°C because
: > they are less reliable than the former cheaper ones... very logical.
: >
:
: How you managed to arrive at that conclusion is anyone's guess.

It's quite simple. Here's a table, you like tables.

------------------- Former ------------------ Latter ----------------
< 85°C reliable unreliable

85°C unreliable reliable

_____________________________________________

 

[M.I.5¾]

:
: : >> All capacitor made are not equal - it's a fact of manufacturing.
About
: >> half
: >> the capacitors will work better than they were designed to, and the
other
: >> half will not work as well.
: >
: > Not quite correct. Almost all work at least as good as they were
: > designed for. Some will work far better. Cofidence levels - all
: > capacitors must at least meet specs - what they are intended to do.
: >
:
: You really don't know anything do you?
:
: Capacitors are produced is such large quantities that the conformance to
: specification is done by testing a sample of each batch only. It is a
: fundamental requirement of sample testing that you have to accept that
items
: that do not conform to specification will be accepted. If you cannot
accept
: that then you can't do sample testing. Also, if you have a look at any
: sample testing tables,

Bwahahahahahah!
You really don't know anything about Quality Control, do you?

Probably more than you do sunshine.

Do these words mean anything to you: Mean, Variance, Standard Deviation?

Certainly do. Thats how sampling tables get produced in the first place.

 

[M.I.5¾]

:
: : >
: > : > :
: > : : > : > How do you dream up all these ridiculous things?"M.I.5¾"
: > : >
: > : > : > : >
: > : >
: > : >
: > : > >: > : > >> What type of capacitor has to cool off before it accepts a
charge?
: > : >
: > : > > A: A faulty one.
: > : >
: > : > > It is well a known problem particularly among certain types of
: > : > > electrolytic capacitors. The usual problem is that the EPR
: > (Effective
: > : > > parallel resistance) of the capacitor falls alarmingly as it
warms
: > up
: > : > > rendering it ineffective as a capacitor.
: > :
: > : Two charts of electrolytic capacitor failure modes. Check out
: > : Table 2 here:
: > : http://industrial.panasonic.com/www-data/pdf/ABA0000/ABA0000TE4.pdf
: > :
: > : Operating at high temperature is shown to cause failure
: > : by increase in leakage current.
: > :
: > : or Figure 2.10 here:
: > :
: >
http://etd.gatech.edu/theses/available/etd-04082007-083102/unrestricted/imam_afroz_m_200705_phd.pdf
: > :
: > : Operating at high temperature is shown to cause failure
: > : by loss of effective resistance (i.e. increased leakage).
: > :
: > : -----------
: > :
: > : Indeed. In general electrolytic capacitors are rated at either 85°C
or
: > : 105°C. The former is obviously cheaper than the latter and that's
often
: > : what gets used as such supplies don't run that warm. However, even
if
: > : operated below 85°C, the former type are considerably more
unreliable
: > than
: > : the latter.
: >
: > Oh, so the latter break down at temperatures below 85°C because
: > they are less reliable than the former cheaper ones... very logical.
: >
:
: How you managed to arrive at that conclusion is anyone's guess.

It's quite simple. Here's a table, you like tables.

------------------- Former ------------------ Latter ----------------
< 85°C reliable unreliable
_____________________________________________

Don't be pillock all your life, have a day off occasionally. I neither said
that nor inferred it. I stated that even when operated below 85°C,
capacitors rated at 85°C are more unreliable than those rated at 105°C.