Capacitor discharge question

[Chuck]

On Fri, 11 Jan 2008 12:13:14 -0800 (PST), Winfield Hill
<hill@rowland.org> wrote:

Winfield Hill wrote:
and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).
^^^^
9 Hz

Thanks for the additional information, Winfield.

My point was only that because ESR is frequency-dependent and the
frequency at which 3 ohms was measured is unstated, 3 ohms may not be
valid for the analysis.

It is still not clear to me how we know the ESR is not less than 3 for
the OP's time constant. None of the posts seemed to address this. If I
understand your analysis, you have taken 3 ohms as the actual ESR at
the frequency of interest, just as the 0.1 ohm resistor and the 90 uF
capacitor values were taken as actual.

I'm open to "recalibration".

Chuck


----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

 

[Phil Allison]

"Chuck the Cuntheaded Moron from Hell "

Thanks for the additional information, Winfield.


My point was only that because ESR is frequency-dependent

** BULLSHIT - you did not have any point.

You are pathetic, lying pile of autistic shit who cannot even read.

and the
frequency at which 3 ohms was measured is unstated, 3 ohms may not be
valid for the analysis.

** Some folk here know the ESR behaviour of electros cos we deal with them
all the time.

YOU fucking do NOT !!

So PISS OFF !!!

It is still not clear to me how we know the ESR is not less than 3 for
the OP's time constant.

** See above - you congenital imbecile.

IGNORANCE is NOT knowledge.

I'm open to "recalibration".

** A bullet in the head would do that nicely.



........ Phil

 

[Winfield]

Chuck wrote:

Winfield Hill wrote:
Winfield Hill wrote:
and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).
^^^^
9 Hz

Thanks for the additional information, Winfield.

My point was only that because ESR is frequency-dependent and the
frequency at which 3 ohms was measured is unstated, 3 ohms may not
be valid for the analysis.

No, that's not exactly right.

It is still not clear to me how we know the ESR is not less than
3 for the OP's time constant. None of the posts seemed to address
this. If I understand your analysis, you have taken 3 ohms as the
actual ESR at the frequency of interest, just as the 0.1 ohm
resistor and the 90 uF capacitor values were taken as actual.

I'm open to "recalibration".

OK, I'll explain. This is what we see if we measure and
analyze many electrolytic caps. ** First, if you examine
datasheets, you'll see esr is usually specified at 100kHz.
In fact, there's a broad region where the ESR changes very
little, e.g., from 0.48 to 0.40 ohms from 1kHz to 200kHz,
for a 68uF 350V electrolytic I measured this afternoon.
From 5kHz and up, the nearly-constant ESR is well below
1/Xc, and this shows a that single value at 100kHz is a
genuinely-useful parameter. (I apologize for not posting
a graph to show this better - we'll do that in AoE 3rd-ed.)

I can tell you, a 1-to-200kHz relatively-flat esr frequency
range is what we generally what we see when measuring small
electrolytics. We have to take the OP's 3 ohms for his part.

** Second, as we go down in frequency, where does Xc take
over from esr? f = 1 / 2pi C Resr = 1 / 2pi 63uF 0.44-ohms
= 5.7kHz for the "68uF" 350V capacitor I measured. Now, to
get into the dielectric series-resistance loss region, shown
in the QuadTech document you referenced, we have to go down
another factor of 50 to 100 in frequency from there, e.g.,
to below 60Hz. In fact, my "68uF" cap has a loss resistance
of 1/69 Xc at 60Hz, and I have to go all the way down to 5Hz
to reach the 1/100 dielectric loss that we expect to see for
an electrolytic. So, clearly there are dramatically-different
regions for electrolytic capacitors, and we can generalize
about them, and most of the time the dielectric losses are
really not much of an issue, being at very low frequencies.

** For example, consider my 68uF cap at 120Hz, the operating
frequency for a bridge-rectifier storage cap. Here the esr
measures about 0.7 ohms, not a whole lot higher than its 0.45
ohms in the 5 to 20kHz region. But consider, in a rectifier
storage capacitor situation, with a short charging-conduction
time, say 1/5 of a cycle's peak, we're really talking about
5*120 = 600Hz. Here I measured an esr of about 0.5 ohms, or
nearly as low as a datasheet-frequency 100kHz esr = 0.4-ohms.

In conclusion, we can safely rely on a single reported value
for capacitor esr, and not worry about whatever dielectric
losses might be at frequencies far below f = 1 / 2pi C Resr.

 

[Tom Bruhns]

On Jan 11, 11:19 am, Winfield Hill <h...@rowland.org> wrote:

Chuck wrote:
Phil Allison wrote:
"Chuck the Cunthead Moron from Hell "

If I understand your question, there is insufficient
information to provide an answer.
ESR is not simply the capacitor's series resistance.

** Shame how it basically is just that.

If it were, then the simple analysis would work.

** Is that nonsense offered as proof ??

Simple = false ?

In fact, ESR is measured at some (unspecified) frequency.

** But as the OP's cap is 60uF at 600volt with and estimated
ESR of 3 ohms - it is most likely to be an electro.

ESR, or whatever you might choose to call it, of a
capacitor discharged as you describe is likely to be
different.

** No it ain't, you fuckwit asshole.

You could assume ESR is invariant wrt energy (likely true)
and perform a simple test on the capacitor at much lower
voltage levels.

** Shame how the OP's Q was about the survival of the
FUCKING RESISTOR

Ah, Phil. Pickings must be slim today. ;-) Usually, your
technical comments are impressive.

For some insight into ESR and actual series resistance,
take a quick look here:
http://www.lowesr.com/QT_LowESR.pdf

This is sci.electronics.design remember.

Chuck

I really hate to side with Phil, especially since
he liberally trashes me whenever it suits his fancy,
but you are off base on this one Chuck, IMHO.

As stated, a 3-ohm esr would be little high for a
60uF electrolytic capacitor, but not excessively so
- 60uF isn't a very large cap. A quick calculation
shows that 3 ohms of esr exceeds 1/Xc above 884Hz,
and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).

Next, we know Mook is discharging the 90uF cap into
a short, basically, so the time constant is 3-ohms
* 90uF = 270us, and f = 1 / 2pi 270us = 590Hz, and
that's well above 90Hz, so we can safely assume the
entire loss is from the so-called "esr" resistance
(the one Phil was referring to), which is more or
less independent of frequency over a wide range of
frequencies above 200Hz or so, for this capacitor.

So I would say, yes, there is enough information to
answer Mook's question. We only have to assume that
the dielectric-loss D is less than about 0.02, which
is pretty safe - I've never measured an electrolytic
that bad and certainly not any worse than that.

So Win, if you drop 0.1 ohms suddenly across the hypothetical 60uF cap
with 3 ohms ESR charged to 600V, what does the terminal voltage of the
cap look like as a function of time? What do you suppose the chances
are of internal arcing in such a resistor of unknown construction
capable of handling a 5 joule impulse? ;-)

It might be more interesting to drop that resistor across a 60uF
ceramic cap, or even a polyprop, charged to 600 volts.

Cheers,
Tom

 

[Mook Johnson]

"Winfield Hill" <hill@rowland.org> wrote in message
news:9ded2a69-ed71-4788-9459-c5f11758c86e@q39g2000hsf.googlegroups.com...

Chuck

I really hate to side with Phil, especially since
he liberally trashes me whenever it suits his fancy,
but you are off base on this one Chuck, IMHO.

As stated, a 3-ohm esr would be little high for a
60uF electrolytic capacitor, but not excessively so
- 60uF isn't a very large cap. A quick calculation
shows that 3 ohms of esr exceeds 1/Xc above 884Hz,
and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).

Next, we know Mook is discharging the 90uF cap into
a short, basically, so the time constant is 3-ohms
* 90uF = 270us, and f = 1 / 2pi 270us = 590Hz, and
that's well above 90Hz, so we can safely assume the
entire loss is from the so-called "esr" resistance
(the one Phil was referring to), which is more or
less independent of frequency over a wide range of
frequencies above 200Hz or so, for this capacitor.

So I would say, yes, there is enough information to
answer Mook's question. We only have to assume that
the dielectric-loss D is less than about 0.02, which
is pretty safe - I've never measured an electrolytic
that bad and certainly not any worse than that.

Here are some specifics on the capacitor. The 60uF capacitor is actually a
module consisting of 60 indvidual capacitors. These are the capacitors used
in the module
http://www.vishay.com/docs/40072/134d.pdf The one we used is the 100uF
125VDC @25C in the T case.


The capacitor module is 6 parallel strings of 10 series connected capacitors
to make a module that is 1250V @ 25c and 750C @ 200C capacitor bank that is
60uF total.

This application will see 200C temperature so that severely limits the type
of capacitor used to this large monstrosity.

thanks for the input and if you have any more insight I'm all ears...well
eyes.

 

[Winfield]

Winfield wrote:

OK, I'll explain. This is what we see if we measure and
analyze many electrolytic caps. ** First, if you examine
datasheets, you'll see esr is usually specified at 100kHz.
In fact, there's a broad region where the ESR changes very
little, e.g., from 0.48 to 0.40 ohms from 1kHz to 200kHz,
for a 68uF 350V electrolytic I measured this afternoon.
From 5kHz and up, the nearly-constant ESR is well below
1/Xc, and this shows a that single value at 100kHz is a
genuinely-useful parameter. (I apologize for not posting
a graph to show this better - we'll do that in AoE 3rd-ed.)

Correction: "the nearly-constant ESR is well above 1/Xc,"

 

[Winfield]

Tom Bruhns wrote:

Winfield Hill wrote:

I really hate to side with Phil, especially since
he liberally trashes me whenever it suits his fancy,
but you are off base on this one Chuck, IMHO.

As stated, a 3-ohm esr would be little high for a
60uF electrolytic capacitor, but not excessively so
- 60uF isn't a very large cap. A quick calculation
shows that 3 ohms of esr exceeds 1/Xc above 884Hz,
and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).

Next, we know Mook is discharging the 90uF cap into
a short, basically, so the time constant is 3-ohms
* 90uF = 270us, and f = 1 / 2pi 270us = 590Hz, and
that's well above 90Hz, so we can safely assume the
entire loss is from the so-called "esr" resistance
(the one Phil was referring to), which is more or
less independent of frequency over a wide range of
frequencies above 200Hz or so, for this capacitor.

So I would say, yes, there is enough information to
answer Mook's question. We only have to assume that
the dielectric-loss D is less than about 0.02, which
is pretty safe - I've never measured an electrolytic
that bad and certainly not any worse than that.

So Win, if you drop 0.1 ohms suddenly across the hypothetical 60uF cap
with 3 ohms ESR charged to 600V, what does the terminal voltage of the
cap look like as a function of time? What do you suppose the chances
are of internal arcing in such a resistor of unknown construction
capable of handling a 5 joule impulse? ;-)

It might be more interesting to drop that resistor across a 60uF
ceramic cap, or even a polyprop, charged to 600 volts.

My 68uF 350V cap has an esr of about 0.44 ohms at 5kHz,
which is its internal time-constant frequency. If we
assume a simple series R+C circuit, charged to say 324V,
an instantaneous 0.1-ohm load would create a 600A peak
discharge current, and initially 264 volts would appear
across the internal series resistance. Shall I try it?
I have some IGBTs that can handle the 600A switching.

 

[legg]

On Sat, 12 Jan 2008 06:34:37 -0600, "Mook Johnson" <mook@mook.net>
wrote:

"Winfield Hill" <hill@rowland.org> wrote in message
news:9ded2a69-ed71-4788-9459-c5f11758c86e@q39g2000hsf.googlegroups.com...
Chuck

I really hate to side with Phil, especially since
he liberally trashes me whenever it suits his fancy,
but you are off base on this one Chuck, IMHO.

As stated, a 3-ohm esr would be little high for a
60uF electrolytic capacitor, but not excessively so
- 60uF isn't a very large cap. A quick calculation
shows that 3 ohms of esr exceeds 1/Xc above 884Hz,
and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).

Next, we know Mook is discharging the 90uF cap into
a short, basically, so the time constant is 3-ohms
* 90uF = 270us, and f = 1 / 2pi 270us = 590Hz, and
that's well above 90Hz, so we can safely assume the
entire loss is from the so-called "esr" resistance
(the one Phil was referring to), which is more or
less independent of frequency over a wide range of
frequencies above 200Hz or so, for this capacitor.

So I would say, yes, there is enough information to
answer Mook's question. We only have to assume that
the dielectric-loss D is less than about 0.02, which
is pretty safe - I've never measured an electrolytic
that bad and certainly not any worse than that.


Here are some specifics on the capacitor. The 60uF capacitor is actually a
module consisting of 60 indvidual capacitors. These are the capacitors used
in the module
http://www.vishay.com/docs/40072/134d.pdf The one we used is the 100uF
125VDC @25C in the T case.


The capacitor module is 6 parallel strings of 10 series connected capacitors
to make a module that is 1250V @ 25c and 750C @ 200C capacitor bank that is
60uF total.

This application will see 200C temperature so that severely limits the type
of capacitor used to this large monstrosity.

thanks for the input and if you have any more insight I'm all ears...well
eyes.

I think that you're asking for trouble using tantalum caps in pulsed
power circuits. I'm not sure what options there are at 200C. Placing
the capacitive energy storage in the hostile environment, if there are
alternatives, could be asking for trouble.

You don't mention bleeder equalization parts, or other frequency
compensating components that I would expect to see in a
series-parallel connected module of this sort. This argues against
predictable pulse performance.

RL

 

[legg]

On Fri, 11 Jan 2008 23:32:06 -0800 (PST), Tom Bruhns <k7itm@msn.com>
wrote:
<snip>

So I would say, yes, there is enough information to
answer Mook's question. We only have to assume that
the dielectric-loss D is less than about 0.02, which
is pretty safe - I've never measured an electrolytic
that bad and certainly not any worse than that.

So Win, if you drop 0.1 ohms suddenly across the hypothetical 60uF cap
with 3 ohms ESR charged to 600V, what does the terminal voltage of the
cap look like as a function of time? What do you suppose the chances
are of internal arcing in such a resistor of unknown construction
capable of handling a 5 joule impulse? ;-)

The internal construction can, at least, be assumed to be distributed
evenly throughout the body of the part, by distinctive layer. This is
ideal for power surge absorption.

Arcing? Under what influence?

What I've observed, is physical movement, as the inductance is forced
to attempt to reduce dI/dT. A 'jumping' of the electrolytic element
within the case. It's no easy matter to get anywhere near an ideal
innitial instantaneous current peak, nor for that matter, to measure
it.

RL

 

[Phil Allison]

" Chuck the Cunthead Moron from Hell "

For some insight into ESR and actual series resistance, take a quick
look here:

http://www.lowesr.com/QT_LowESR.pdf

** As with most links posted in lieu of an actual point - this one is
also utterly irrelevant to the OP's issue.




........ Phil

 

[Winfield Hill]

Winfield Hill wrote:

and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).
^^^^

9 Hz

 

[Winfield Hill]

Chuck wrote:

Phil Allison wrote:
"Chuck the Cunthead Moron from Hell "

If I understand your question, there is insufficient
information to provide an answer.
ESR is not simply the capacitor's series resistance.

** Shame how it basically is just that.

If it were, then the simple analysis would work.

** Is that nonsense offered as proof ??

Simple = false ?

In fact, ESR is measured at some (unspecified) frequency.

** But as the OP's cap is 60uF at 600volt with and estimated
ESR of 3 ohms - it is most likely to be an electro.

ESR, or whatever you might choose to call it, of a
capacitor discharged as you describe is likely to be
different.

** No it ain't, you fuckwit asshole.

You could assume ESR is invariant wrt energy (likely true)
and perform a simple test on the capacitor at much lower
voltage levels.

** Shame how the OP's Q was about the survival of the
FUCKING RESISTOR

Ah, Phil. Pickings must be slim today. ;-) Usually, your
technical comments are impressive.

For some insight into ESR and actual series resistance,
take a quick look here:
http://www.lowesr.com/QT_LowESR.pdf

This is sci.electronics.design remember.

Chuck

I really hate to side with Phil, especially since
he liberally trashes me whenever it suits his fancy,
but you are off base on this one Chuck, IMHO.

As stated, a 3-ohm esr would be little high for a
60uF electrolytic capacitor, but not excessively so
- 60uF isn't a very large cap. A quick calculation
shows that 3 ohms of esr exceeds 1/Xc above 884Hz,
and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).

Next, we know Mook is discharging the 90uF cap into
a short, basically, so the time constant is 3-ohms
* 90uF = 270us, and f = 1 / 2pi 270us = 590Hz, and
that's well above 90Hz, so we can safely assume the
entire loss is from the so-called "esr" resistance
(the one Phil was referring to), which is more or
less independent of frequency over a wide range of
frequencies above 200Hz or so, for this capacitor.

So I would say, yes, there is enough information to
answer Mook's question. We only have to assume that
the dielectric-loss D is less than about 0.02, which
is pretty safe - I've never measured an electrolytic
that bad and certainly not any worse than that.

 

[Chuck]

On Sat, 12 Jan 2008 00:22:28 +1100, "Phil Allison"
<philallison@tpg.com.au> wrote:

"Chuck the Cunthead Moron from Hell "

If I understand your question, there is insufficient information to
provide an answer.


** ROTFLMAO !!!!!!!!!

There are ZERO questions posted on usenet with enough info to even know
what the fuck OPs are on about.



ESR is not simply the capacitor's series resistance.


** Shame how it basically is just that.


If it were, then
the simple analysis would work.


** Is that nonsense offered as proof ??

Simple = false ?



In fact, ESR is measured at some (unspecified) frequency.


** But as the OP's cap is 60uF at 600volt with and estimated ESR of 3
ms - it is most likely to be an electro.

Means absolutely fuck all to a trolling, posturing, ASD fucked moron like
YOU - right ?


ESR, or whatever you might choose to call it,
of a capacitor discharged as you describe is likely to be different.


** No it ain't, you fuckwit asshole.


You could assume ESR is invariant wrt energy (likely true) and perform
a simple test on the capacitor at much lower voltage levels.


** Shame how the OP's Q was about the survival of the

FUCKING RESISTOR


YOU ASSSS




........ Phil

Ah, Phil. Pickings must be slim today. ;-) Usually, your technical
comments are impressive.

For some insight into ESR and actual series resistance, take a quick
look here:

http://www.lowesr.com/QT_LowESR.pdf

This is sci.electronics.design remember.

Chuck



----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

 

[Phil Allison]

"Chuck the Cunthead Moron from Hell "

If I understand your question, there is insufficient information to
provide an answer.

** ROTFLMAO !!!!!!!!!

There are ZERO questions posted on usenet with enough info to even know
what the fuck OPs are on about.

ESR is not simply the capacitor's series resistance.

** Shame how it basically is just that.

If it were, then
the simple analysis would work.

** Is that nonsense offered as proof ??

Simple = false ?

In fact, ESR is measured at some (unspecified) frequency.

** But as the OP's cap is 60uF at 600volt with and estimated ESR of 3
ms - it is most likely to be an electro.

Means absolutely fuck all to a trolling, posturing, ASD fucked moron like
YOU - right ?

ESR, or whatever you might choose to call it,
of a capacitor discharged as you describe is likely to be different.

** No it ain't, you fuckwit asshole.

You could assume ESR is invariant wrt energy (likely true) and perform
a simple test on the capacitor at much lower voltage levels.

** Shame how the OP's Q was about the survival of the

FUCKING RESISTOR


YOU ASSSS




......... Phil

 

[Chuck]

On Thu, 10 Jan 2008 21:59:28 -0600, "Mook Johnson" <mook@mook.net>
wrote:

Lets say there is a 60uF cap charged to 600V. The energy in that caps

E = CV^2 which is ~ 22 joules

If the cap has 3 ohms of ESR and the output was shorted across a .1 ohm
resistor. How much of the neergy would be dissipated in teh cap vs. in the
resistor?

Would it be as simple as 0.1/3.1 = ~ 3% of the total energyin the resistor?
I cna get my hands on some resistors rated for 3-5 joules that will fit in
the confines of space and wondering if they could be destroyed by the
capacitor spark test (less than 5 times).

If I understand your question, there is insufficient information to
provide an answer.

ESR is not simply the capacitor's series resistance. If it were, then
the simple analysis would work. In fact, ESR is measured at some
(unspecified) frequency. ESR, or whatever you might choose to call it,
of a capacitor discharged as you describe is likely to be different.

You could assume ESR is invariant wrt energy (likely true) and perform
a simple test on the capacitor at much lower voltage levels.

Chuck

----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

 

[Phil Allison]

"Mook Johnson"

Lets say there is a 60uF cap charged to 600V. The energy in that caps

E = CV^2 which is ~ 22 joules

** Its 11 joules, actually.

If the cap has 3 ohms of ESR

** Very poor grade cap.

Not designed for rapid discharge or shorting.

and the output was shorted across a .1 ohm resistor. How much of the
neergy would be dissipated in teh cap vs. in the resistor?

** The ratio of resistances is gonna be close to the mark.

Would it be as simple as 0.1/3.1 = ~ 3% of the total energyin the
resistor? I cna get my hands on some resistors rated for 3-5 joules that
will fit in the confines of space and wondering if they could be destroyed
by the capacitor spark test (less than 5 times).

** Low ohm resistors may not like high pulse voltages as the gap in the
spiralling may arc over at the moment of applying a 600 volt charged cap.

It will soon go open if this happens.



........ Phil

 

[John Popelish]

Mook Johnson wrote:

Lets say there is a 60uF cap charged to 600V. The energy in that caps

E = CV^2 which is ~ 22 joules

Actually, there is a divide by two in that formula, also, so
only half that.

If the cap has 3 ohms of ESR and the output was shorted across a .1 ohm
resistor. How much of the neergy would be dissipated in teh cap vs. in the
resistor?

Would it be as simple as 0.1/3.1 = ~ 3% of the total energyin the resistor?
I cna get my hands on some resistors rated for 3-5 joules that will fit in
the confines of space and wondering if they could be destroyed by the
capacitor spark test (less than 5 times).

Instantaneous power dumped into a resistor is related to the
current by P=I^2*R, so since the current through both series
connected resistors is similar at all times, the total
energy dumped into the series resistors must be proportional
to their individual resistance.

--
Regards,

John Popelish

 

[Mook Johnson]

Lets say there is a 60uF cap charged to 600V. The energy in that caps

E = CV^2 which is ~ 22 joules

If the cap has 3 ohms of ESR and the output was shorted across a .1 ohm
resistor. How much of the neergy would be dissipated in teh cap vs. in the
resistor?

Would it be as simple as 0.1/3.1 = ~ 3% of the total energyin the resistor?
I cna get my hands on some resistors rated for 3-5 joules that will fit in
the confines of space and wondering if they could be destroyed by the
capacitor spark test (less than 5 times).

 

[Tom Bruhns]

On Jan 12, 9:22 am, legg <l...@nospam.magma.ca> wrote:

On Fri, 11 Jan 2008 23:32:06 -0800 (PST), Tom Bruhns <k7...@msn.com
wrote:
snip



So I would say, yes, there is enough information to
answer Mook's question. We only have to assume that
the dielectric-loss D is less than about 0.02, which
is pretty safe - I've never measured an electrolytic
that bad and certainly not any worse than that.

So Win, if you drop 0.1 ohms suddenly across the hypothetical 60uF cap
with 3 ohms ESR charged to 600V, what does the terminal voltage of the
cap look like as a function of time? What do you suppose the chances
are of internal arcing in such a resistor of unknown construction
capable of handling a 5 joule impulse? ;-)

The internal construction can, at least, be assumed to be distributed
evenly throughout the body of the part, by distinctive layer. This is
ideal for power surge absorption.

Arcing? Under what influence?

What I've observed, is physical movement, as the inductance is forced
to attempt to reduce dI/dT. A 'jumping' of the electrolytic element
within the case. It's no easy matter to get anywhere near an ideal
innitial instantaneous current peak, nor for that matter, to measure
it.

RL

Well, PottyMouth Humorboy suggested that "low ohm resistors ... may
arc over" with a 600V pulse applied. Perhaps, but I'd rather expect
that (1) with the effective series R and L involved, the resistor
would never see close to 600V, and (2) resistors that can handle that
much energy are likely not constructed in a way that would have a
problem with a 600V impulse (though the latter should of course be
confirmed and not blindly assumed).

Cheers,
Tom

 

[Tom Bruhns]

On Jan 12, 5:48 am, Winfield <winfieldh...@yahoo.com> wrote:

Tom Bruhns wrote:
Winfield Hill wrote:
I really hate to side with Phil, especially since
he liberally trashes me whenever it suits his fancy,
but you are off base on this one Chuck, IMHO.

As stated, a 3-ohm esr would be little high for a
60uF electrolytic capacitor, but not excessively so
- 60uF isn't a very large cap. A quick calculation
shows that 3 ohms of esr exceeds 1/Xc above 884Hz,
and we can estimate that ordinary esr loss exceeds
the dielectric loss at about 1/100 of that frequency,
or 90Hz (the factor of 100 is for D = 0.01, etc.).

Next, we know Mook is discharging the 90uF cap into
a short, basically, so the time constant is 3-ohms
* 90uF = 270us, and f = 1 / 2pi 270us = 590Hz, and
that's well above 90Hz, so we can safely assume the
entire loss is from the so-called "esr" resistance
(the one Phil was referring to), which is more or
less independent of frequency over a wide range of
frequencies above 200Hz or so, for this capacitor.

So I would say, yes, there is enough information to
answer Mook's question. We only have to assume that
the dielectric-loss D is less than about 0.02, which
is pretty safe - I've never measured an electrolytic
that bad and certainly not any worse than that.

So Win, if you drop 0.1 ohms suddenly across the hypothetical 60uF cap
with 3 ohms ESR charged to 600V, what does the terminal voltage of the
cap look like as a function of time? What do you suppose the chances
are of internal arcing in such a resistor of unknown construction
capable of handling a 5 joule impulse? ;-)

It might be more interesting to drop that resistor across a 60uF
ceramic cap, or even a polyprop, charged to 600 volts.

My 68uF 350V cap has an esr of about 0.44 ohms at 5kHz,
which is its internal time-constant frequency. If we
assume a simple series R+C circuit, charged to say 324V,
an instantaneous 0.1-ohm load would create a 600A peak
discharge current, and initially 264 volts would appear
across the internal series resistance. Shall I try it?
I have some IGBTs that can handle the 600A switching.

Well, if it's easy. I was thinking I could use a little SCR to do
something similar; long ago (well before power mosfets) I used them to
generate reasonably fast rise pulses -- basically a pulse-forming
network switch, as used in pulsed magnetron radars.

Cheers,
Tom