I have always been a user of teflon and polystyrene caps in the signal pathe (for over 20yrs now) I recently built a ClassD audio amp which uses SMD caps on the input. These are niobium oxide. They have a very high DA (41) and compared to almost any film, would be considered a poor choice, yet I do not hear problems. They are also polarized. Since they are SMDs replacing them with film types is problematical. It could be done by removing the SMDs and putting the input caps at at the input connectors (xlrs) I hesitate to do this as the amp as I said, sound fine. Is it worth to try this? Has any one else done with a switching amp? Thanks for any help or suggestions.
DA does not cause distortion, as it is a linear phenomenon. It slightly modifies the phase shift at LF and subsonic frequencies, which is likely to be inaudible. Some people find that high DA caps sound bad, but that could be a correlation rather than causation (i.e. some dielectrics with high DA are also non-linear - the non-linearity is the problem not the DA).
The manufacturer's information claims that niobium oxide dielectric properties do not vary with bias voltage. I don't know to what extent that claim is true, but it seems to fit with your findings that no distortion is generated.
The manufacturer's information claims that niobium oxide dielectric properties do not vary with bias voltage. I don't know to what extent that claim is true, but it seems to fit with your findings that no distortion is generated.
Looking at http://www.newark.com/pdfs/techarticles/avx/AVX_NIOBIUM_OXIDE_TECHNOLOGY_ROADMAP.pdf
The key feature is no piezoelectric effects, which plague high value ceramics. This implies that the capacitance is not modulated by bias voltage.
The key feature is no piezoelectric effects, which plague high value ceramics. This implies that the capacitance is not modulated by bias voltage.
I had always been of the opinion that a high DA produced a smearing effect as small amounts of the signal were released over time, thus producing a less than ideal transient response. I am now forced to question the audibility of this judging by the results obtained by the use of niobium oxide caps. Any comments on this aspect? Thanks
It is good to question. Your description of DA seems culled from the folklore of the last 30yr. and not your personal experience.
EDIT - It would be good to read the information more carefully dielectric constant is not DA and piezoelectricity is mechanical. I don't see data for voltage coefficient of capacitance or DA.
Last edited:
Actually Scott, I have been using teflon caps for over twenty years since Jerry Foreman of Precision Audio Supply introduced me to them. I have always found them to improve audible clarity and transient response when used to replace even polyprop and foil caps, much less mylar caps, so I can speak from some empirical experience. When I went back to graduate school, I had a year in which component auditioning played an important part. Of course, it might not have been DA that I was hearing. There could have been other factors responsible for what I was hearing, but my testing involved a number of different circuits but all in the signal path and I always preferred the teflon or polystyrene cap. Scott I certainly have the utmost regard for your credentials. Could you perhaps point to other factors that I may have ovelooked. Thanks
Jacco I do not think I am being fooled into thinking that teflon are better because they are bigger or more expensive. But then again I could be fooling myself. That is why I started this thread as a new experience has defied twenty years of listening experience using original masters of classical music which I engineered. Regards
I need to get with the times- wasn't aware that there was any standard system for reporting DA. You may have read about the bridge residual tests I've done on caps on my site. It would be interesting to run those same tests on the niobium oxide and see how they compare.
IMO, cap audibility depends entirely on where the cap is in the circuit. For most of my purposes I doubt I could hear a difference between polystyrene, polypropylene and maybe Teflon. They're just too identical in every property.
IMO, cap audibility depends entirely on where the cap is in the circuit. For most of my purposes I doubt I could hear a difference between polystyrene, polypropylene and maybe Teflon. They're just too identical in every property.
Those are the units of dielectric constant ("x" relative to vacuum). 41 has no meaning with respect to DA.
I don't hear these things in general, so maybe I can't help you. I just thought it was a good thing to question.
Conrad - The spec quoted for DA is MIL-C-19978
Last edited:
Thanks Scott for the explanation. I had always assumed there was a direct correlation between dialectric constant and DA. It is a good day I have learned something! BTW John Curl, whose expertise I also respect is a big fan of teflon and polystyrene caps and uses them whenever possible. I guess even experts can agree to disagree. Kindest regards and thanks again for the input.
Don't believe everything you read. The 'release over time' is seen in DC (or near DC) systems such as sample-and-hold, and can be modelled with extra parallel CR circuits. All these do is slightly change the LF response by a negligible amount.john dozier said:
Given that a DC signal is 'smeared' it seems so logical that the same thing would happen to music. This is an occasion where intuition leads people astray. DA simple stores and then releases the long term signal average. What is the long term average of a music signal? Well, if there is any AC coupling anywhere earlier in the path (including the microphone itself) the long term average of music is zero. So DA stores and returns zero. Not very harmful?
So for best SQ one should use ribbon mics with dc coupled preamp, a dc coupled ADC, a dc coupled CD mastering system, dc coupled CD player, and dc coupled amp? Know of any recording company using the above? The closest I have come is Sheffield Labs with their LA Philharmonic recordings. Regards
A ribbon mike has, in effect, AC coupling because it is a velocity transducer.
I wonder if you misunderstood my point? Any AC coupling means long-term average is zero. Lots of AC coupling, as is the norm, means that not only is there no DC but there is little subsonic either so DA has even less effect. DA itself really is harmless for audio. The issue is whether some common dielectrics with high DA also have other, harmful, features such as non-linearity.
I wonder if you misunderstood my point? Any AC coupling means long-term average is zero. Lots of AC coupling, as is the norm, means that not only is there no DC but there is little subsonic either so DA has even less effect. DA itself really is harmless for audio. The issue is whether some common dielectrics with high DA also have other, harmful, features such as non-linearity.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.