0.2-2uF cap with lowest ESR, concrete ESR values?

[njiruk]

Hi,

I'm looking for a capacitor between 0.2 and 2 uF with ESR as low as possible.
Electrolytic polymer capacitor's are not available for such low capacities, and I found several sources stating that ceramic's would have lower ESR.
This page:

http://www.kemet.com/kemet/web/home...0CBE2E5238A759A385256A8700515CCD?OpenDocument

says that polymer aluminum's have lowest ESR of electrolyts, but still are 3 times the ESR of ceramic capacitors.
And here:

http://www.e-sonic.com/whatsnew/emag2004_3/emag_Passive_nic.pdf

it is stated that ceramic chip capacitor's (MLCC's) have "ultra low esr", especially Class 1 (NP0/C0G).

That sounds good, but in practice I have difficulties finding concrete ESR values, and even more to find ESR values in the 100Hz-2kHz range, used in my application. This can drastically change...

What type of capacitor will guarantee lowest ESR and would there be any catalog with precise ESR information?

Thanks in advance!
{urijn

 

[itsmoked]

Lowest ESR ones I've ever come across:

http://us.sanyo.com/Electronic-Devices-Components/OS-CON

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

What purpose are you going to use the capacitor for? There's little point in suggesting something designed for a heavy-current power application if you want a signal-level filter capacitor. ESR can not be the only parameter of importance - size, voltage rating, current handling, cost...?

 

[njiruk]

@ScottyUK: Your question makes sense...

-size is not of importance.
-voltage rating and current handling neither, as I can use the oscillator at any amplitude.
-I'd like not to surpass 3€/piece

@itsmoked: the sanyo os-con's look interesting indeed, but I'm curious about the specs of ceramic variants...

To illustrate my application:

It is for a highly resonating audio series RLC oscillator, (with f_res = about 1200 Hz).
Logically I wont use a resistor, the R value will be the sum of the inductor's DCR and capacitor's ESR values.

-> given that w0 = 1/sqrt(LC) and Q = sqrt(L/C) / R
and R = DCR + ESR.
Apart from materials and design, I believe we can globally state that:
DCR = proportional to the inductor's cable length, so as L=N^2, roughly: DCR ~ sqrt(L)
ESR = inversely proportional to the capacity: ESR ~ 1/C
This gives: Q ~ sqrt(L/C) / (sqrt(L)+1/C)

So that I conclude that I should chose L/C as high as possible. I've found this confirmed using several LC sets of existing components.

I found a cheap choke of 5mH with DCR=45mOhm (no risk for saturation as the voltage rating is not of big importance for my application).

That's why I'd like to find a capacitor in that capacity range.

 

[geekEE]

If you're using ceramic caps in an audio circuit, watch out for microphonic effects. I've built amps that ended up amplifying the sound of me tapping on the PCB with a screwdriver. This effect is much worse if the source of your audio is a relatively high impedance.

If size is not important, how about paralleling multiple lower-value caps to get an effectively lower ESR?

Glenn

 

[njiruk]

The piezo-electric effect, I see... Well, I guess that's not a big concern as it's a test that I can perform in silent and "hissless" conditions...
I'd think paralleling caps gives the same results as taking a similar type of twice the capacity, no?
But, should I opt for MLCC's rather then polymers, concerning ESR? No concrete ESR values (preferably in the audio frequency domain) available?

 

[ScottyUK]

You might consider the polypropylene film types often used in power electronics. ESR will typically be very low, ESL somewhat higher than for a ceramic type. Does ESL bother you? I suspect it may not be a dominant effect at 1200Hz but certainly something to be aware of as frequencies rise.

A quick look in Kemet's range got the following possibles:
[li]2uF with 2m[Ω] ESR at 100kHz for one of the power types designed to piggyback onto an IGBT module, with an ESL of 43nH.[/li]
[li]2uF with 1.5m[Ω] ESR at 100kHz for one of the axial cans designed for GTO protection, with an undisclosed ESL.[/li]

 

[BrianG]

Why are you so concerned about ESR in an audio oscillator circuit? The "cheap choke of 5mH" that you have found will probably have a greater effect on the actual resonant frequency, due to its inductance tolerance, than any effect that a few milliohms difference in ESR of the capacitor will have. As you are apparently building a "series RLC oscillator" and not a filter the overall Q is of less importance than having sufficient loop gain to maintain stable oscillation.

 

[njiruk]

That sounds great!
They even appear to be not that expensive...
ESL is still high enough indeed.
I didn't find ESR information though. Generally the dissipation factor @ 120Hz is specified.
Logically I can presume ESR = tan(d)/(2*pi*120*C), right?

Strange, when I compare ESR at 120Hz and the ESR calculated with the dissipation factor with this formula, for example in the Nichicon FPCAP catalog (

http://www.nichicon.co.jp/english/products/pdf/2012fpcap_catalog_all.pdf),

for some capacitor's it's coherent, for others I obtain a ratio of thousand...!?

I'll maybe just buy some components and measure ESR myself...

 

[njiruk]

@BrianG: that's right, as in a simple RLC-circuit R=ESR+DCR... However, I'm actually building a "discrete audio transmission line", where various frequencies will travel through, here the resistive part of the capacitor becomes important, especially at high frequencies.

 

[Comcokid]

You posted that the capacitor is for a highly resonating audio series RLC oscillator so you should select the capacitor for properties related to maximizing the Q of the resonant circuit. You need to concentrate of the capacitors Dielectric Absorption properties. I would not rely on the ESR of the capacitor for the R of your circuit.

ESR is a catch-all measurement for capacitors that is generally used when selecting capacitors for filtering (power supply noise/ripple). Dissipation Factor is an older figure-of-merit that lumps ESR and ESL together in one number.

If your aplication was HF to UHF the best choice would be NP0 ceramics. Since your application is in the audio range, I would use a polystyrene, polypropylene, or Teflon film (very low DF). Given your cost target I would forget about Teflon.

 

[njiruk]

@Comcokid : Initially I said a RLC oscillator, to keep things simple, but given the development of the post, I explained the purpose into more detail in my previous post... It is not really the Q of a single RLC, but the "modal Q-factors", so the ESR and the inductor's DCR becomes independent.
But the relation of dielectric absorption to the capacitor's R is (the main part of) ESR.
Maybe you want to state that this absorption factor is more directly connected to Q as it is independent of C? But I find ESR more practical in my mathematical approaches...
Dissipation factor rather lumps ESR and wC together (

http://en.wikipedia.org/wiki/Dissipation_factor).

Thanks for the audio related tip, I'll try polypropylenes soon!

For the moment I'm more concerned about the parasitic capacitance of the 5mH choke..!