What is considered low ESL for HV Caps?

J

JT

Guest
Hi,

I'm looking for the lowest possible ESL for a 1000VDC 50uF cap. So far
I've tracked down a 16nH and 30nH part. Are these much lower than,
lets say, I would find on an ordinary electrolytic?

Cheers,

JT
 
I read in sci.electronics.design that JT <jt.steenkamp@usask.ca> wrote
(in <1124739854.980618.324200@g49g2000cwa.googlegroups.com>) about 'What
is considered low ESL for HV Caps?', on Mon, 22 Aug 2005:
Hi,

I'm looking for the lowest possible ESL for a 1000VDC 50uF cap. So far
I've tracked down a 16nH and 30nH part. Are these much lower than,
lets say, I would find on an ordinary electrolytic?

You wouldn't get a 1000 V electrolytic; the top whack is about 600 V.
Click to expand...
A 50 uF 1000 V part can't be very small, and the physical size sets a
lower limit to the inductance.
--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
 
Yes, but is 16-30 nH pretty standard or does it fall into a category of
fast capacitors? I know manufacturing techniques and dielectrics play
a role. All I really need is a benchmark to go by.
 
On 22 Aug 2005 12:44:15 -0700, "JT" <jt.steenkamp@usask.ca> wrote:

Hi,

I'm looking for the lowest possible ESL for a 1000VDC 50uF cap. So far
I've tracked down a 16nH and 30nH part. Are these much lower than,
lets say, I would find on an ordinary electrolytic?
Click to expand...
The values of inductance quoted typically reflect limitations of the
capacitor's mechanical construction. They may be smaller than
mechanical strays inadvertently present in hardware where they are
typically used, in the voltage range that you refer to.

HV caps of most types are more likely to show unanticipated
restrictions on permissible dV/dT, ie pulsed current. These are not
small-signal components, after all.

Optimal capacitor performance over a very wide frequency range usually
requires more than one capacitor type, paralleled, if self-resonance
effects are a threat.

It is unlikely that the 50uF 1KV part you mention is an electrolytic -
perhaps you might offer a part number or other reference.

RL
 
JT wrote:
Hi,

I'm looking for the lowest possible ESL for a 1000VDC 50uF cap. So far
I've tracked down a 16nH and 30nH part. Are these much lower than,
lets say, I would find on an ordinary electrolytic?

Cheers,

JT
Click to expand...
No, not if Cornell-Dubilier is telling the truth. Their aluminum
electrolytics are less than 2nH.

http://www.cde.com/ and click on the "Application Guide (revised)" link in
about the center of the page. You'll get a helpful document. Look at the top
left of page 2.185 in that document to verify my statement above.

Good luck.

John
 
John - KD5YI wrote:
JT wrote:

Hi,

I'm looking for the lowest possible ESL for a 1000VDC 50uF cap. So far
I've tracked down a 16nH and 30nH part. Are these much lower than,
lets say, I would find on an ordinary electrolytic?

Cheers,

JT


No, not if Cornell-Dubilier is telling the truth. Their aluminum
electrolytics are less than 2nH.

http://www.cde.com/ and click on the "Application Guide (revised)" link
in about the center of the page. You'll get a helpful document. Look at
the top left of page 2.185 in that document to verify my statement above.

Good luck.

John
Click to expand...
Actually, I just read more of the document and on page 2.193 I found...

INDUCTANCE
Inductance is the equivalent series inductance, and it is relatively
independent of both frequency and temperature.
Typical values range from 2 to 8 nH for SMT types, 10 nH
to 30 nH for radial-leaded types, 20 to 50 nH for screw-terminal
types, and up to 200 nH for axial-leaded types.
These low values are achieved by tab location and intrinsic,
low inductance of the dielectric contact geometry. The
capacitor element has typical inductance of less than 2 nH.

The 2nH is sort of misleading since it is pretty hard to use a capacitor
element without the terminals and other gear it usually comes with.

Anyway, there you go.

John
 
John - KD5YI wrote:
John - KD5YI wrote:

JT wrote:

Hi,

I'm looking for the lowest possible ESL for a 1000VDC 50uF cap. So far
I've tracked down a 16nH and 30nH part. Are these much lower than,
lets say, I would find on an ordinary electrolytic?

Cheers,

JT


No, not if Cornell-Dubilier is telling the truth. Their aluminum
electrolytics are less than 2nH.

http://www.cde.com/ and click on the "Application Guide (revised)"
link in about the center of the page. You'll get a helpful document.
Look at the top left of page 2.185 in that document to verify my
statement above.

Good luck.

John


Actually, I just read more of the document and on page 2.193 I found...

INDUCTANCE
Inductance is the equivalent series inductance, and it is relatively
independent of both frequency and temperature.
Typical values range from 2 to 8 nH for SMT types, 10 nH
to 30 nH for radial-leaded types, 20 to 50 nH for screw-terminal
types, and up to 200 nH for axial-leaded types.
These low values are achieved by tab location and intrinsic,
low inductance of the dielectric contact geometry. The
capacitor element has typical inductance of less than 2 nH.

The 2nH is sort of misleading since it is pretty hard to use a capacitor
element without the terminals and other gear it usually comes with.

Anyway, there you go.

John
Click to expand...
That saves me from writing my $0.02 then :)

bloody misleading really, that bit on page 2.185.

IME big electros are around the 20-30nH range. An n-tuple well
paralleled can reduce that of course, but often uses series caps to get
a high enough coltage rating :(

I do have some 450Vdc 250uF film caps with 2-3nH *total* ESL.
Custom-made, using a very sneaky trick. I could tell you how, but then
I'd have to kill you. Oh, they only have one screw terminal (oops, said
too much :)

Astonishingly expensive, renting a capacitor manufacturers factory for a
week :)

Cheers
Terry
 
Thanks for the help gentlemen. In case anyone's interested, the caps
I tracked down are:

Vishay: GMKPg 1000-42uF
Elcon: UL30BL0050

Thanks to this discussion, I'm also considering the Cornell-dublier
SCRN film-paper series. These are considered good for high-frequency
yet feature screw terminals, which puts them in the slower league
according to their application guide. Thus, if 20-50nH is still
small, I'm yet to see a cap, at my required spec, whichever
dielectric, with a higher ESL.
 
On 25 Aug 2005 14:37:45 -0700, "JT" <jt.steenkamp@usask.ca> wrote:

Thanks for the help gentlemen. In case anyone's interested, the caps
I tracked down are:

Vishay: GMKPg 1000-42uF
Elcon: UL30BL0050

Thanks to this discussion, I'm also considering the Cornell-dublier
SCRN film-paper series. These are considered good for high-frequency
yet feature screw terminals, which puts them in the slower league
according to their application guide. Thus, if 20-50nH is still
small, I'm yet to see a cap, at my required spec, whichever
dielectric, with a higher ESL.
Click to expand...
These are polypropylene.

RL
 
"legg" ...
These are polypropylene.
Click to expand...

** Yep - the OP never suggested otherwise.


Hint: He is almost certainly an SET, audio amp freak.




.......... Phil
 
That's correct. I only mentioned electrolytic as a dielectric I could
compare with. I'm very well aware of what the caps are made of.

But, their purpose is far from audio. You've fingered the wrong freak.
;)
 
On 26 Aug 2005 08:01:44 -0700, "JT" <jt.steenkamp@usask.ca> wrote:

That's correct. I only mentioned electrolytic as a dielectric I could
compare with. I'm very well aware of what the caps are made of.

But, their purpose is far from audio. You've fingered the wrong freak.
Click to expand...
Audio?

A misunderstanding - from your initial comparison of 'found' parts to
"ordinary electrolytics". I assumed that you were discussing 'special'
electrolytics - with abnormal voltage ratings. No other dielectrics or
physical constructions were mentioned.

Inductance of a carefully terminated film capacitor (full schoopage),
is a function of its lead spacing and body distance from the test
points used. This forms the physical loop. Coaxial construction offers
a minimal lead spacing, but is pricey.

RL
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top