[SOLVED] High current capacitor

[kripacharya]

Is the ESR the only factor involved in selecting a capacitor for high current capability ?

If so, which type of capacitor would be best for current ~15-20A @1000Vpk freq < 1Mhz ?

 

[FvM]

At this high AC voltage level, dielectric losses must be taken into consideration as well. Suitable power electronics capacitors (there aren't so much available in this range at all) should have respective information in the datasheets.

You should specify a capacitor value and an actual operation frequency to give an idea of possible candidates.

freq < 1Mhz is a bit general, similar to speed below speed of light...

 

[THINKER_KAM_MAN]

Is the ESR the only factor involved in selecting a capacitor for high current capability ?

If so, which type of capacitor would be best for current ~15-20A @1000Vpk freq < 1Mhz ?

Just a suggestion

ESR, is not the only criteria
You need to consider value,
Voltage rating,
Reliability,
Temperature,
Dielectric type,
Dielectric Loss,
Temperature coefficient,
Pulse currant capability,
Ripple current handling,
Case size,
Availability,
Applications.


What value capacitor do you need.
looks like you need a transmitter type capacitor
(door Knob type)

 

[kripacharya]

At this high AC voltage level, dielectric losses must be taken into consideration as well. Suitable power electronics capacitors (there aren't so much available in this range at all) should have respective information in the datasheets.

You should specify a capacitor value and an actual operation frequency to give an idea of possible candidates.

freq < 1Mhz is a bit general, similar to speed below speed of light...

Doesn't ESR include the effects of dielectric losses ?
I'm looking for caps of the range 10nF to 200nF range.

And a freq of < 1 Mhz is more like speed below speed of sound, rather than speed of light. A Freq spec of < 50GHz would be more like your light comparison surely ?

 

[D.A.(Tony)Stewart]

There are only 2 losses in caps series and parallel leakage. The dissipation factor is sometimes used instead for ESR. Motor caps and supply ripple caps are rated in Amps. Loe ESR caps are sometimes rated in Amps, but usually have lower ESR than those rated in Amps. The tan delta is sometimes used for dissipation.

Then there is stray lead inductance or ESL which determines the self resonant freq. (SRF)

Caps will self destruct as a short from overvoltage unless self-healing type plastic.

Surge suppression caps are rated for kV pulses with various rise time and durations such as 1, 10us. They are usually rated for line voltage in classes like X1 or Y2 and made from metallized plastic film.

 

[FvM]

Doesn't ESR include the effects of dielectric losses ?

If you have a single ESR number to characterize a capacitor, no. If you refer to a frequency dependent ESR specification in a capacitor datasheet, yes.

You also have usually curves of permissible AC current respectively AC voltage versus frequency which consider both loss mechanisms.

And a freq of < 1 Mhz is more like speed below speed of sound, rather than speed of light.

My comment wasn't meant literally. I was wondering if 1MHz is a realistic number in combination with the other parameters. Can we read 1000 Vpk as 700 Vrms, or is it a pulse voltage specification? Superimposed DC?

From the capacitors I know, polypropylene "snubber" types are possible candidates. But your capacitance, frequency and voltage ranges have no or only a small overlap, so it would be better to refer to a specific example.

 

[THINKER_KAM_MAN]

Is the ESR the only factor involved in selecting a capacitor for high current capability ?

If so, which type of capacitor would be best for current ~15-20A @1000Vpk freq < 1Mhz ?

just a suggestion

ok, value 10nf to 200nf

what is the application?
and actual frequency.
.........
is for SMPSU,
RF AMP output CAP,
power factor correction,
TV LOPT,
supply decoupling,
input coupling dc blocker.
capacitance power supply

and radio waves travel at the speed of light
in space regard less of the frequency. but lower through cables

is this true: it takes light approx
1sec to travle 300*10^8 mtr
then
frequency of light = (300*10^8)/(300*10^8) = 1Hz

 

[kripacharya]

frequency is in the range 100Khz to 250Khz.
This cap will be used as part of a series LC circuit driven by MOSFETs which deliver a square wave of upto 50v pk-pk, and it is terminated in a power-line coupler where the reflected impedance is low.

Simulations show that the voltage & currents across the cap can go very high at the resonance frequency.

The Q is not too high, and the cap range falls in the range 10nF to 100nF or so.

 

[FvM]

frequency is in the range 100Khz to 250Khz

The specification is approaching feasible numbers.





Care must be taken when comparing capacitor ratings. Some manufacturers are calculating it for 10 K temperature rise (e.g. Wima), others up to 40 K (Kemet). At the end of the day you have to qualify capacitors in your application circuit based on measurements.

 

[THINKER_KAM_MAN]

frequency is in the range 100Khz to 250Khz.
This cap will be used as part of a series LC circuit driven by MOSFETs which deliver a square wave of upto 50v pk-pk, and it is terminated in a power-line coupler where the reflected impedance is low.

Simulations show that the voltage & currents across the cap can go very high at the resonance frequency.

The Q is not too high, and the cap range falls in the range 10nF to 100nF or so.

some application, sounds good.
well there are a special type of capacitors
for power applications
called "Power ceramic capacitors"

https://www.rell.com/products/Capacitors/Ceramic-RF-Power-Capacitor.html

they can range up to 100 KV rating.
however the capacitance values are up to 3nf
unless you have one specially made for you. because of the voltage ratings, they can be large.
you could also use door knob capacitors
connected in parallel to obtain the value
they are cheaper.
Teflon high voltage power capacitors might also do.

 

[kripacharya]

The specification is approaching feasible numbers.
....

LoL !! Thanks FvM !

- - - Updated - - -

..
you could also use door knob capacitors...

door knobs are wayyy too large for our use. I was looking for something like those metalized polypropelene caps. Thanks for the pointers though.

 

[THINKER_KAM_MAN]

LoL !! Thanks FvM !

- - - Updated - - -



door knobs are wayyy too large for our use. I was looking for something like those metalized polypropelene caps. Thanks for the pointers though.

well the high voltages and currents your are considering requires this type of capacitor.

if you equipment is just for a quick test you might getaway with
the standard types of capacitors.
but if your requirements is for continuous operation
you will find that the end termination of the standard
capacitor will fail because of the pulse currents.

further more they can't handle 50KV
vacuum capacitors would be required for that.

i wonder if your simulation is correct, you know what simulators
are like.

 

[FvM]

further more they can't handle 50KV

the 50 kV requirement is your idea, I think.

We are talking about 50 or 100 price factors between standard power electronic foil capacitors and RF HV capacitors. The latter are dedicated for applications like RF transmitters. General power electronic can hopefully use standard types.

 

[D.A.(Tony)Stewart]

Is the ESR the only factor involved in selecting a capacitor for high current capability ?

If so, which type of capacitor would be best for current ~15-20A @1000Vpk freq < 1Mhz ?

Can you please fix your specified requirements?
and yes ESR and voltage rating is the most important spec to prevent self destruction of capacitors.


your spec implies high inductance, otherwise, 20A*1000Vp=20 kilowatts (Peak)
If this is a surge dump waveform to be applied to a snubber, then say so, as some inductance is required between Source and snubber otherwise >1MHz is required.

The closest small cap to an ideal one is made of Mica or PTFE.

E.g. Mica Cornell Dubilier CDV16 Mica 3000pF
Highlights
• Handles up to 9.0 amps rms continuous current
• Very low ESR from 10 to 100 MHz
• Low, notch-free impedance to 1GHz
• Stable: no capacitance change with (V), (t), and (f) • Very high Q at UHF/VHF frequencies
• Tape and reeling available
• dV/dt capability up to 275,000 V/μs • 1,500 amps peak current capability

 

[THINKER_KAM_MAN]

well i did not mean that he require 50KV caps just that
they can do 50KV.
and sunnyskyguy seems to have catch on to where i am going
because one needs a special capacitor as his initial requirements
where @ 1000V and big amps.

 

[kripacharya]

Can you please fix your specified requirements?
and yes ESR and voltage rating is the most important spec to prevent self destruction of capacitors.


your spec implies high inductance, otherwise, 20A*1000Vp=20 kilowatts (Peak)
If this is a surge dump waveform to be applied to a snubber, then say so, as some inductance is required between Source and snubber otherwise >1MHz is required.

The closest small cap to an ideal one is made of Mica or PTFE.

E.g. Mica Cornell Dubilier CDV16 Mica 3000pF
Highlights
• Handles up to 9.0 amps rms continuous current
• Very low ESR from 10 to 100 MHz
• Low, notch-free impedance to 1GHz
• Stable: no capacitance change with (V), (t), and (f) • Very high Q at UHF/VHF frequencies
• Tape and reeling available
• dV/dt capability up to 275,000 V/μs • 1,500 amps peak current capability

Most of my refined specs are already in this thread.
A cap has a 90o phase diff between V & I. Which means for ideal cap the power is ZERO Kilowatts.

However the ESR & other losses cause the power to be non-zero, and this is where the type of cap is important. This is what this thread is about.

 

[FvM]

A cap has a 90° phase diff between V & I. Which means for ideal cap the power is ZERO Kilowatts.

However the ESR & other losses cause the power to be non-zero, and this is where the type of cap is important. This is what this thread is about.

I was under the impression that the thread is about selection of suitable capacitors for a specific application rather than capacitor theory.

Different capacitor types that may serve your application have been suggested. It's your matter to decide what's suitable and affordable.

 

[kripacharya]

I was under the impression that the thread is about selection of suitable capacitors for a specific application rather than capacitor theory.

Different capacitor types that may serve your application have been suggested. It's your matter to decide what's suitable and affordable.

It is about app specific cap selection. Thanks for all your inputs. I've been able to put together a prototype which seems to work fine. Currents i think are not that high yet probably because of the ESR/ losses and the fact that I have only loaded it with 1/4th max load. Clearly the o/p impedance is not zero, and hence voltage drops with higher loads.

I was just clarifying to SunnySkyguy some possible misunderstanding he might have had.

thanks guys !!

 

[bill66656]

As I know, the impedance of the capacitor at the switching frequency and it's harmonics should be less than that of the paralleled port, for the unwanted high frequency ac signal can be absorbed by the capacitor.

I didn't know the high current rate would affect the criteria. Maybe it won't.