If this can help, there is what i have from various sources about DA of various Caps in ascending order:
MLO 0.0015%
Teflon <0.01%
Polystyrene 0.05%
Multi-Layer Glass 0.05%
Poly-Propylene 0.05% to 0.1
Multi-Layer Glass 0.05%
Polyphenylene sulfide film capacitors (PPS) 0.05 to 0.1%
Polyester film capacitors (PET) 0.2 to 0.5%
Poly-Carbonate 0.35%
Mica 0.3% - 0.7%
NP0 Ceramic 0.6%
Polyethylene naphthalate film capacitors (PEN) 1.0 to 1.2%
Tantalum electrolytic capacitors with solid electrolyte 2 to 3 %
X7R ceramic capacitors 2.5% (Ceramic Discs or Multilayers)
Aluminium electrolytic capacitors with non-solid electrolyte 10 to 15%
MLO 0.0015%
Teflon <0.01%
Polystyrene 0.05%
Multi-Layer Glass 0.05%
Poly-Propylene 0.05% to 0.1
Multi-Layer Glass 0.05%
Polyphenylene sulfide film capacitors (PPS) 0.05 to 0.1%
Polyester film capacitors (PET) 0.2 to 0.5%
Poly-Carbonate 0.35%
Mica 0.3% - 0.7%
NP0 Ceramic 0.6%
Polyethylene naphthalate film capacitors (PEN) 1.0 to 1.2%
Tantalum electrolytic capacitors with solid electrolyte 2 to 3 %
X7R ceramic capacitors 2.5% (Ceramic Discs or Multilayers)
Aluminium electrolytic capacitors with non-solid electrolyte 10 to 15%
Last edited:
The NP0 caps that Groner measured had the lowest distortion of anything iirc. So that calls into further question the linkage of DA and distortion. What would be useful: a measurement of voltage coefficient for each dielectric. Of course it is almost certainly as much or more about construction as about dielectric for film caps.
why isn't anyone doing the basic reading?
many questions are addressed already
really try reading some of the links - even when the math and physics makes your eyes glaze over its often possible to glean some useful things from the text, graphs
the Pease Capacitor Soakage paper shows a detailed model of a 1.0 uF Mylar cap measured at varying time scales with Bob's custom test circuit, plots several dielectrics over decades of time resolution
the Pease article is getting harder to find free and complete - my archive.org link has lost the last half of the paper
the Kundert Modeling Dielectric Absorption in Capacitors paper gives a graph of dielectric permittivity mechanisms, more models, more abstract "frac-pole" continuous models rather than the parallel RC branch model Pease shows
as a linear phenom DA shouldn't care about dielectric thickness
but nonlinear distortion mechanisms do care about V/m in the dielectric - one of Bateman's takeaway points in the Capacitor Sound articles is using higher V rated Caps if measured distortion is important to you
many questions are addressed already
really try reading some of the links - even when the math and physics makes your eyes glaze over its often possible to glean some useful things from the text, graphs
the Pease Capacitor Soakage paper shows a detailed model of a 1.0 uF Mylar cap measured at varying time scales with Bob's custom test circuit, plots several dielectrics over decades of time resolution
the Pease article is getting harder to find free and complete - my archive.org link has lost the last half of the paper
the Kundert Modeling Dielectric Absorption in Capacitors paper gives a graph of dielectric permittivity mechanisms, more models, more abstract "frac-pole" continuous models rather than the parallel RC branch model Pease shows
as a linear phenom DA shouldn't care about dielectric thickness
but nonlinear distortion mechanisms do care about V/m in the dielectric - one of Bateman's takeaway points in the Capacitor Sound articles is using higher V rated Caps if measured distortion is important to you
Last edited:
That is especially true with polar types found in consumer audio equipment. Much much less so with films.
I didnt check any mica or npo or tiny values of any type. Just film against the popular use of polar caps in audio.
-RNM
Last edited:
Yes, audio is "asymmetrical" is fuzzy thinking. Take the simplest arrangement, a capacitor paralleled by a series R/C, it does the same thing. Charge it for a long time, short it, and open it. The second capacitor retains some charge and you see the result of that.
This subject has been around the bend before same inapt analogies come up, we're only a page away from DA as "memory/echos".
"Can we model a high DA cap by a normal one paralleled with an other one of high value and high ESR ?"
This does not imply your perfect //RC does not distort audio signals (non linearity) neither.
Last edited:
Thickness, IMO, would show up in the simple DA model as higher series R value with the Cda.
With polar caps, the higher volatge measure better when you measure the Z vs freq.... they act more towards the ideal cap...... straighter, extended line on log-log. I recommended higher volatge and higher temp since day one because of the Z curve data. Do those poor Z curves of low volatage and low temp polar caps indicate higher distortion?
THx-RNMarsh
Last edited:
yes the linearity also applies to the thickness argument - if you want the same value C with thicker dielectric then you have to have proportionately more area - if everything is Linear then you should end up in the same place
just like the 4 R in paralleled 2x( series 2x) give the same terminal value of R (or more generally Z)
just like the 4 R in paralleled 2x( series 2x) give the same terminal value of R (or more generally Z)
Last edited:
And even worse when it is the dog chasing his tail and gets distracted hearing the echo of his master's voice.
I see PMA answered my question, good. DA is not an indicator of
THD. I think jcx's estimates of the response distortion of 100th's or less dB in most cases is a good start, certainly for a large number of ordinary capacitors.
Harmonic. When I write "distortion", I mean non-linear distortion. In case I want to distinguish both, I write "linear distortion" (amplitude, phase, time response) and "non-linear distortion" (new harmonic components created). Initial change of time response is not any non-linear distortion.
- Status
- Not open for further replies.