Low Pass Filter Capacitors

[simon]

I want to try some capacitors in my DAC buffer stage to make a low pass filter. I'm looking for 100pF at a minimum of 250V rating.

A quick look on Farnell suggests ceramic, polypropylene, polyester and silver mica are available - do some types of capacitor lend themselves to this application more than others? Or does it not really matter?

[Neal]

COG (NPO) Ceramics can be very good but personally I'd go for the Polyprop then Polyester...the micas can be a bit variable and TBH no better than Polyprop but at much greater expense. my 2p worth!

[simon]

Thanks Neal, polyprop or ceramic would probably have been my choice so it's good to get an opinion. Going to be a struggle getting to a £20 order for free delivery for just 4 small caps

[Mike H]

My personal preferrence for small values audio use is polystyrene. Hi-Fi Collective have 630V ones.

http://www.hificollective.co.uk/compone ... yrene.html

In my own experience ceramic discs can sound nasty also polyester if not specifically for audio (either), mica may be OK though, but depends as Neal says. Mylar might be OK too but again depends where you're using them.

[Mike H]

Oh of course you may not get mylar or polyester is such small values.

Just leaves ceramic, mica, polystyrene then.

[Nick]

At that size I would try silver mica

[simon]

They're a couple of quid each, not exactly expensive compared to everything else in it!

[Mike H]

Quite so.

Try one of each

[simon]

This DIY lark. Nothing's ever finished really, is it?!

[coolboz]

I think ceramic is good.

[Nick]

I think ceramic is good.

What dialectric have you used?

[Mike H]

Found it again!!!



This page I knew from years ago but then it disappeared, about choosing capacitors.


http://www.reliablecapacitors.com/pickcap.htm

Note texts about ceramics.

[Neal]

The trouble with that Mike is it appears to lump all ceramics in together. A more in depth and better study of capacitor performance was done by Bateman in 2002 where it was noted Ceramic COG / NPO types had excellent distortion performance.

[Neal]

Quote: Ceramic capacitors.
`Ceramic' covers an extremely wide range of dielectrics. In the seventies the Erie Company produced more than fifty different capacitor ceramic formulations, sub-divided as Class 1 (non-polar) or Class 2 (polar) according to the materials used.
Class 1 ceramics do not contain Barium Titanate, so have a low `k' value. The best known is C0G. With its controlled temperature coefficient of zero ± 30 ppm, it was originally called NP0 by the Erie Corporation. It is non-polar and has a small dielectric absorption coefficient. From my tests it has almost no measurable harmonic distortion. COG ceramic is more stable with time and temperature than mica capacitors and from my tests COG can produce less distortion. see Fig. 5
C0G ceramic provides the most stable capacitance value, over long time periods and temperature excursions, of all easily obtained capacitor dielectrics. It is frequently used as a capacitance transfer standard in calibration laboratories. Yet as a small disc capacitor it costs only pennies. Assembled as a multilayer, it can provide capacitances of 100 nF and above, rated for 100 volts working, and much higher voltages for smaller capacitances.
Other Class 1 ceramics, sometimes called `low k', provide increased capacitance within a controlled temperature coefficient, e.g. P100, N750 etc. in ppm. These also are non-polar and exhibit little dielectric absorption. I have tested up to N750, sometimes called U2J, and found very low distortion.
Class 2 ceramics do include Barium Titanate. It produces a very high dielectric constant, with `k' values ranging from a few hundred to several thousands depending on other additives used. Class 2 ceramic is strongly polar, its capacitance varies with applied voltage and temperature. It exhibits an easily measured dielectric absorption, which increases with `k' value.
Popular Class 2 ceramics include the X7R, W5R, BX capacitor grades and the exceptionally high `k' Z5U. These do produce
extremely large measured distortions, so are not suited for use in the signal path of an audio system.
see Fig. 6

See Part three for Figures to above...

[Mike H]

Yes there is a difference between low K and high K, I thought that was explained in the Picking Capacitors article, but perhaps not perhaps I saw it somewhere else instead