why are electrolytic caps so much worse than in 1981?

T

Terry Given

Guest
Hi All,

my copy of Eugene Hnateks "design of solid-state power supplies, 2nd ed"
arrived last week. Its a great book, there is a wealth of information on
self-oscillating dc-dc converters (Graham, you would probably be quite
interested in it). On p. 280 he states:

"The lowest ESR in aluminium [capacitor] construction is what is known
as the 'stacked foil' type; with all other styles it is typically an
order of magnitude higher. Representative values would be 0.5 milliOhms
for a stacked foil device and 5 to 20 milliOhms for other styles"

and on p.585 he states:

"Stacked foil exhibits inductance values as low as 2nH and provides
superior ripple-curent ratings"


Wow, 2mOhms ESR, 2nH ESL - thats one hell of an electrolytic capacitor.
He repeats the < 1mOhm remark in several places throughout the book, so
its probably not a typo. Where can I get these puppies :)

Cheers
Terry
 
"Terry Given" <my_name@ieee.org> wrote in message
news:AQQVd.6473$1S4.687012@news.xtra.co.nz...
Harry Dellamano wrote:
"Terry Given" <my_name@ieee.org> wrote in message
news:YzQVd.6465$1S4.686735@news.xtra.co.nz...

Hi All,

my copy of Eugene Hnateks "design of solid-state power supplies, 2nd ed"
arrived last week. Its a great book, there is a wealth of information on
self-oscillating dc-dc converters (Graham, you would probably be quite
interested in it). On p. 280 he states:

"The lowest ESR in aluminium [capacitor] construction is what is known as
the 'stacked foil' type; with all other styles it is typically an order
of magnitude higher. Representative values would be 0.5 milliOhms for a
stacked foil device and 5 to 20 milliOhms for other styles"

and on p.585 he states:

"Stacked foil exhibits inductance values as low as 2nH and provides
superior ripple-curent ratings"


Wow, 2mOhms ESR, 2nH ESL - thats one hell of an electrolytic capacitor.
He repeats the < 1mOhm remark in several places throughout the book, so
its probably not a typo. Where can I get these puppies :)

Cheers
Terry

You can get them at 100C, the ESR anyway, but the 2nH ESL, only in
hybrids.
Cheers,
Harry



Hi Harry,

I suspected as much. All his work seemed to be done for NASA, so $$ was
probably not much of a concern. I havent seen an ESR that low though, and
I have looked at a lot of big caps over the last decade or so. Got any
links?

Cheers
Terry
Click to expand...
Got links in freezer, bring some beer and I will cook them up.
Mosfets will hit 1.0m Ohms before any caps do. If this was in 1981 he was
off by at least two orders of magnitude. Maybe his Wheatstone Bridge got
wet.
Terry, keep submitting, you've got some good stuff.
Regards,
Harry
 
Harry Dellamano wrote:
"Terry Given" <my_name@ieee.org> wrote in message
news:AQQVd.6473$1S4.687012@news.xtra.co.nz...

Harry Dellamano wrote:

"Terry Given" <my_name@ieee.org> wrote in message
news:YzQVd.6465$1S4.686735@news.xtra.co.nz...


Hi All,

my copy of Eugene Hnateks "design of solid-state power supplies, 2nd ed"
arrived last week. Its a great book, there is a wealth of information on
self-oscillating dc-dc converters (Graham, you would probably be quite
interested in it). On p. 280 he states:

"The lowest ESR in aluminium [capacitor] construction is what is known as
the 'stacked foil' type; with all other styles it is typically an order
of magnitude higher. Representative values would be 0.5 milliOhms for a
stacked foil device and 5 to 20 milliOhms for other styles"

and on p.585 he states:

"Stacked foil exhibits inductance values as low as 2nH and provides
superior ripple-curent ratings"


Wow, 2mOhms ESR, 2nH ESL - thats one hell of an electrolytic capacitor.
He repeats the < 1mOhm remark in several places throughout the book, so
its probably not a typo. Where can I get these puppies :)

Cheers
Terry

You can get them at 100C, the ESR anyway, but the 2nH ESL, only in
hybrids.
Cheers,
Harry



Hi Harry,

I suspected as much. All his work seemed to be done for NASA, so $$ was
probably not much of a concern. I havent seen an ESR that low though, and
I have looked at a lot of big caps over the last decade or so. Got any
links?

Cheers
Terry


Got links in freezer, bring some beer and I will cook them up.
Click to expand...
?! now I'm confused. Still, if I'd have said I have a few maggot-packs
in the freezer you would no doubt be confused too. Aint dialects fun :)

Mosfets will hit 1.0m Ohms before any caps do. If this was in 1981 he was
off by at least two orders of magnitude. Maybe his Wheatstone Bridge got
wet.
Click to expand...
LOL. the bit that surprised me was it was repeated several times.

I have worked with some 150uF 900Vdc 1.2nH (not a typo) film caps
though. They were a special, designed by our resident albanian/serbian
croat (interesting fella), and were fantastic. Based on RBs idea, I
thought up a way of reducing the inductance on a big electrolytic - down
to the 2-3nH region. Hard to convince a cap manufacturer to make them
though.

Terry, keep submitting, you've got some good stuff.
Regards,
Harry
Click to expand...
<blushes>
I really enjoy this NG, I have learned an awful lot of really useful
stuff, and "met" some wicked smart people. If any of you are ever in NZ,
look me up.

Cheers
Terry
 
"Harry Dellamano" <harryd@tdsystems.org> wrote in message
news:DHRVd.39495$uc.31108@trnddc01...
Got links in freezer, bring some beer and I will cook them up.
Mosfets will hit 1.0m Ohms before any caps do. If this was in 1981 he
was
off by at least two orders of magnitude. Maybe his Wheatstone Bridge got
wet.
Terry, keep submitting, you've got some good stuff.
Regards,
Harry
Click to expand...
I've got a couple of 1.6 ?? milliohm electros at work - 47000 uF 16V, I
think they are. Got them from Farnell at AU $50 each when we needed to test
our milliohm meter at low values. They are not particularly big.

Roger
 
Hello Terry,

Got links in freezer, bring some beer and I will cook them up.


?! now I'm confused. Still, if I'd have said I have a few maggot-packs
in the freezer you would no doubt be confused too. Aint dialects fun :)
Click to expand...

Since Google doesn't have an English - English translation mode: I
believe you guys in NZ speak more of a UK style and then "links" might
be "bangers". There is a subtle difference though. Our links are usually
a bit more meaty and harder.

Regards, Joerg

http://www.analogconsultants.com
 
Joerg wrote:
Hello Terry,

Got links in freezer, bring some beer and I will cook them up.



?! now I'm confused. Still, if I'd have said I have a few maggot-packs
in the freezer you would no doubt be confused too. Aint dialects fun :)



Since Google doesn't have an English - English translation mode: I
believe you guys in NZ speak more of a UK style and then "links" might
be "bangers". There is a subtle difference though. Our links are usually
a bit more meaty and harder.

Regards, Joerg
Click to expand...
Hi Joerg,

now owuldnt an english-english translator be a cute bit of software, eh
wot? We dont use the term bangers, but most would recognise it. sossys,
snags, snarlers..... most of which are really just sausage-shaped bread
in a casing. I had 10kg of incredibly spicy Boerwoerst (sp?) made a
month or so ago, by a butcher up near Kopu. Absolutely delicious, he
uses his own meat (ooh er, sounds a bit rude) - the butchery is on his
farm. Damn they are good, and cheaper than the crap the stupormarket sells.

Cheers
Terry
 
Terry Given wrote:

now owuldnt an english-english translator be a cute bit of software, eh
wot? We dont use the term bangers, but most would recognise it.
Click to expand...
The problem isn't the bits you recognize/recognise, but the ones
that are invisible, such as "table the discussion" having different
meanings or the use of "E-Z" which can be spoken as "ee-zee (easy)
or as "ee-zed (eased)".

BTW, I do not agree that electrolytic caps are worse than they were
in 1981. Perhaps the cheap ones are, but the mil-spec ones have not
changed for the worse.
 
legg wrote:
On Fri, 04 Mar 2005 16:26:46 +1300, Terry Given <my_name@ieee.org
wrote:


Hi All,

my copy of Eugene Hnateks "design of solid-state power supplies, 2nd ed"
arrived last week. Its a great book, there is a wealth of information on
self-oscillating dc-dc converters (Graham, you would probably be quite
interested in it). On p. 280 he states:

"The lowest ESR in aluminium [capacitor] construction is what is known
as the 'stacked foil' type; with all other styles it is typically an
order of magnitude higher. Representative values would be 0.5 milliOhms
for a stacked foil device and 5 to 20 milliOhms for other styles"

and on p.585 he states:

"Stacked foil exhibits inductance values as low as 2nH and provides
superior ripple-curent ratings"


Wow, 2mOhms ESR, 2nH ESL - thats one hell of an electrolytic capacitor.
He repeats the < 1mOhm remark in several places throughout the book, so
its probably not a typo. Where can I get these puppies :)


Hnatek is not talking about electrolytic parts here.

Aluminium is used as an electrode material in polycarbonate, polyester
and polystyrene film capacitors as well. It can be a deposited film or
simple foil.

Stacked film refers to polycarbonate film caps first marketed by
Siemens and TRW. These are identifiable by their markedly sharp-edged
box-shaped bodies, as they are cut from larger continuous 'rods' of
laminated stock, after schoopage is applied. They differ from
conventional construction, which is involves individual part rolling
(and flattening, in many cases to reduce real estate use).

Leads are stitch welded to the schoopage, or they can be reflowed.

A protective coating of epoxy is sometimes applied to reduce
delamination.

http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32560_64.pdf
http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32572_73.pdf
http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32591_94.pdf

http://www.paktron.com/

Some interesting work has been done identifying apparent standing-wave
behavior in this type of construction - dependant on physical size,
mounting terminal method and orientation - at frequencies where most
other film caps are just stray inductance.

RL
Click to expand...
Ah, that explains it nicely. I have looked at using Capsticks before,
but avoided them on the grounds of cost.

Perusing the siemens data, none seem to get much better than 10mOhms at
resonance. The capsticks go down to about 2.5mOhms.

So Hnateks numbers are still wildly optimistic, but being out by an
order of magnitude is a lot better than being out by two orders of
magnitude (he says, stating the blindingly obvious. Still, I need the
typing practice :)

Cheers
Terry
 
On Sat, 05 Mar 2005 15:37:40 +1300, Terry Given <my_name@ieee.org>
wrote:



http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32560_64.pdf
http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32572_73.pdf
http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32591_94.pdf

http://www.paktron.com/

Some interesting work has been done identifying apparent standing-wave
behavior in this type of construction - dependant on physical size,
mounting terminal method and orientation - at frequencies where most
other film caps are just stray inductance.

RL


Ah, that explains it nicely. I have looked at using Capsticks before,
but avoided them on the grounds of cost.

Perusing the siemens data, none seem to get much better than 10mOhms at
resonance. The capsticks go down to about 2.5mOhms.
Click to expand...
Paktron data also documents an apparent frequency-dependance of ESR,
that makes me hold my nose. ESR is unlikely to behave in this manner,
so just what ARE they measuring for their published specs?

RL
 
legg wrote:
On Sat, 05 Mar 2005 15:37:40 +1300, Terry Given <my_name@ieee.org
wrote:




http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32560_64.pdf
http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32572_73.pdf
http://www.epcos.com/inf/20/20/db/fc_05/MKT_B32591_94.pdf

http://www.paktron.com/

Some interesting work has been done identifying apparent standing-wave
behavior in this type of construction - dependant on physical size,
mounting terminal method and orientation - at frequencies where most
other film caps are just stray inductance.

RL


Ah, that explains it nicely. I have looked at using Capsticks before,
but avoided them on the grounds of cost.

Perusing the siemens data, none seem to get much better than 10mOhms at
resonance. The capsticks go down to about 2.5mOhms.



Paktron data also documents an apparent frequency-dependance of ESR,
that makes me hold my nose. ESR is unlikely to behave in this manner,
so just what ARE they measuring for their published specs?

RL
Click to expand...
I would expect some frequency dependance (ie skin/proximity effect), but
thats the opposite of what the curves show. Paktron are not the only
ones to display such plots of ESR with frequency, I have analagous
Hitachi AIC cap plots, and seem to recall seeing others too.

perhaps its a frequency-dependant loss mechanism, and a
parallel-to-series conversion to get to "ESR"

Cheers
Terry
 
John Woodgate wrote:
I read in sci.electronics.design that legg <legg@nospam.magma.ca> wrote
(in <a2kk21p7m96feidff4i1rm5ogfkbdqrgqg@4ax.com>) about 'why are
electrolytic caps so much worse than in 1981?', on Sun, 6 Mar 2005:


Paktron data also documents an apparent frequency-dependance of ESR,
that makes me hold my nose. ESR is unlikely to behave in this manner, so
just what ARE they measuring for their published specs?


They are probably using a model that is somewhat too simple, maybe just
C R and L in series. To get an ESR that is independent of frequency, the
model needs at least a resistor in parallel with the C to model
dielectric loss.
Click to expand...
I think this must be it (see my last post). Ripple current ratings for
electrolytics are generally higher at HF than LF, often by around 30% or
so. Thus the losses must be higher at LF, which would lead to the
observed "ESR" curve using the P-S approach.

Whether or not the same is true for film caps, beats me.

Cheers
Terry
 
On Sun, 6 Mar 2005 07:11:41 +0000, John Woodgate
<jmw@jmwa.demon.contraspam.yuk> wrote:

I read in sci.electronics.design that legg <legg@nospam.magma.ca> wrote
(in <a2kk21p7m96feidff4i1rm5ogfkbdqrgqg@4ax.com>) about 'why are
electrolytic caps so much worse than in 1981?', on Sun, 6 Mar 2005:

Paktron data also documents an apparent frequency-dependance of ESR,
that makes me hold my nose. ESR is unlikely to behave in this manner, so
just what ARE they measuring for their published specs?

They are probably using a model that is somewhat too simple, maybe just
C R and L in series. To get an ESR that is independent of frequency, the
model needs at least a resistor in parallel with the C to model
dielectric loss.
Click to expand...
A Paktron article includes a graph of ESR vs frequency for a number of
types. Supposedly this data is collected and interpreted using the
same methods and equipment.

http://www.paktron.com/techarticles/3.99A/3.99A.pdf

I think the effect in film types at lowering frequencies is related
somehow to dielectric absorption, which can have effects at very low
frequencies that are a bit spooky.

Some dielectrics have very low absoption, though, without altering the
ESR vs frequency plots published, so it's a mystery to me. A parallel
R in a film cap model would likely have too large an order of
magnitude to affect it.

RL
 
I read in sci.electronics.design that legg <legg@nospam.magma.ca> wrote
(in <a2kk21p7m96feidff4i1rm5ogfkbdqrgqg@4ax.com>) about 'why are
electrolytic caps so much worse than in 1981?', on Sun, 6 Mar 2005:

Paktron data also documents an apparent frequency-dependance of ESR,
that makes me hold my nose. ESR is unlikely to behave in this manner, so
just what ARE they measuring for their published specs?
Click to expand...
They are probably using a model that is somewhat too simple, maybe just
C R and L in series. To get an ESR that is independent of frequency, the
model needs at least a resistor in parallel with the C to model
dielectric loss.
--
Regards, John Woodgate, OOO - Own Opinions Only.
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
 
In article <%WsWd.7054$1S4.741270@news.xtra.co.nz>,
Terry Given <my_name@ieee.org> wrote:

I would expect some frequency dependance (ie skin/proximity
effect), but thats the opposite of what the curves show. Paktron
are not the only ones to display such plots of ESR with
frequency, I have analagous Hitachi AIC cap plots, and seem to
recall seeing others too.

perhaps its a frequency-dependant loss mechanism, and a
parallel-to-series conversion to get to "ESR"
Click to expand...
Cyril Bateman's articles in Electronics World mention
and explain ESR versus frequency.

EW April 1995, page 287. "Power dissipation in capacitors".

EW December 1997, page 998. "Understanding capacitors".

--
Tony Williams.
 
Tony Williams wrote:
In article <%WsWd.7054$1S4.741270@news.xtra.co.nz>,
Terry Given <my_name@ieee.org> wrote:


I would expect some frequency dependance (ie skin/proximity
effect), but thats the opposite of what the curves show. Paktron
are not the only ones to display such plots of ESR with
frequency, I have analagous Hitachi AIC cap plots, and seem to
recall seeing others too.


perhaps its a frequency-dependant loss mechanism, and a
parallel-to-series conversion to get to "ESR"


Cyril Bateman's articles in Electronics World mention
and explain ESR versus frequency.

EW April 1995, page 287. "Power dissipation in capacitors".

EW December 1997, page 998. "Understanding capacitors".
Click to expand...
Hi Tony,

can you please scan & post it to ABSE? or provide a summary?

Cheers
Terry
 
In article <FrpXd.7972$1S4.847136@news.xtra.co.nz>,
Terry Given <my_name@ieee.org> wrote:

Tony Williams wrote:
EW December 1997, page 998. "Understanding capacitors".

can you please scan & post it to ABSE? or provide a summary?
Click to expand...
Post2: Chunk of CB's text from EW December 1997, page 1002.

I have stressed that esr is frequency dependant, but does
it really change by a significant amount, or am I simply
being pedantic?

Consider the case of a high quality 10nF polystyrene
foil/film capacitor. I selected such a device as one of
the standard capacitors when building my capacitance
bridge. All measurements were taken using a Wayne Kerr
precision component analyser. with the test voltage set
to 1V. Table 1.

These results show clearly how esr values do change
significantly with frequency, for this capacitor.
Many writers on this topic have confused these esr and
tsr terms. Obviously they differ substantially, except
at that frequency when the capacitor is self resonant.

Since correct understanding of esr is essential to avoid
over-stressing capacitors, I make no apology for labouring
the point.

CB's Table 1, page 998.
~~~~~~~~~~~~~~~~~~~~~~~
Frequency Capacitance (nF) Tan(delta) 'Q' ESR(ohms)

100Hz 9.9982 0.00010 9000 17.0
1KHz 9.9988 0.00005 20000 0.8
10KHz 9.9986 0.00015 6000 0.26
100KHz 10.0000 0.0005 3000 0.05

--
Tony Williams.
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top