As a first order approximation, using perfect C's and R's. But this is actually quite funny, come to think of it.
By introducing imperfect capacitors in this CR-ladder, you can model nonlinearities in DA! Lets imagine a CR-ladder with the junkiest X7R caps possible. Main distortion mechanism in these caps is that their capacitance is a function of the voltage applied. In the case of X7R caps, capacitance goes down as field strength goes up. This would model for a dielectric material where DA moves in the same direction - field strength up, DA down.
I hope you agree with me, that a CR-ladder either built or simmed with imperfect capacitors, would distort an AC signal applied to it. So, it is possible to model a nonlinear dielectric material by extending the first order approximation, and the result is distortion.
This thought experiment, of course, does not prove that nonlinear DA exists in any material. But it does show how it would lead to distortion.
Last edited:
It's asserted, but there's no data. Seems to me that this connection would maximize H3.
That guy was Austrian meteorologist Julius von Hann, not a relative of Mr. Hamming.
< Julius von Hann - Wikipedia, the free encyclopedia >
Edit: I'm 2nd!
Well it is single ended.
The Pass'ter
...
Simplicity is not the only reason for the use of the single-ended topology.
The characteristic of a single-ended gain stage is the most musically
natural.
Its asymmetry is similar to the compression / rarefaction
characteristic of air, where for a given displacement slightly higher
pressure is observed on a positive (compression) than on a negative
(rarefaction).
Air itself is observed to be a single-ended medium, where
the pressure can become very high, but never go below 0. The harmonic
distortion of such a medium is second harmonic, the least offensive
variety.
...
[broken up for your reading pleasure]
http://www.firstwatt.com/pdf/art_zen_amp.pdf
Ol'Skhul '94!
PS I'm sensing that biased, large value electros (Silmic II, mmm) are a good solution.
PPS No wonder I love all my simple, single-ended pre and PWR amp cicuits! (OK, I've gone too far ...)
Last edited:
Well it took you some time, but it's a good one
The only thing that took some time was getting back here to read your post. I went searching for that specific image a minute after reading it.
se
Back before Richard Marsh led us to servos, we used various caps for filtering and DC blocking, especially in feedback loops of amps and preamps. Once we went to servos, by 1980, then we could remove the TANTALUM caps from the feedback loop, and direct couple throughout. That is what most engineers do today, when looking for best overall performance. In the feedback loop, the nominal DC offset would be the DC imbalance of the input differential pair, usually under 50mV, can be sustained by a solid tantalum cap, even if it is of a polar type. This is because both sides of the film are exposed to the air during fabrication, and this gives a certain amount of bi-polar protection due to the oxide film that is there due to exposure to oxygen, BEFORE forming to a higher voltage on one side. Tantalum was popular in the late 1970's, but then aluminum electrolytics improved, especially those made in Japan, and we stopped using tantalum caps. Tantalum has MORE distortion generation potential than an aluminum electrolytic, but it was more compact and easily available at the time.
This 1978 test shows that for the actual capacitance available, that inverted parallel connection could be better (lower measured distortion) than series connection or just a single cap used (as was typical at the time). It was a measured effect that I found interesting at the time. Today, I would not do it, but I think that Samuel Groner recently rediscovered this property in one of his tests.
For those of you who are not closemindedly critical of anything that Richard Marsh and I measured in the past, just LOOK at the data. You might devise your own test if you are interested, but remember, when Richard and I were researching in the 'wilderness' back in the late 70's, almost everybody used ceramic or tantalum with impunity, not realizing their serious limitations for audio use.
This 1978 test shows that for the actual capacitance available, that inverted parallel connection could be better (lower measured distortion) than series connection or just a single cap used (as was typical at the time). It was a measured effect that I found interesting at the time. Today, I would not do it, but I think that Samuel Groner recently rediscovered this property in one of his tests.
For those of you who are not closemindedly critical of anything that Richard Marsh and I measured in the past, just LOOK at the data. You might devise your own test if you are interested, but remember, when Richard and I were researching in the 'wilderness' back in the late 70's, almost everybody used ceramic or tantalum with impunity, not realizing their serious limitations for audio use.
Yeah, I never bought the "more natural" argument. But if you're going to have distortion, it's good if you can keep it low order and monotonic, because of masking. Oh yeah, there's an electrolytic output coupling cap on this too.
An externally hosted image should be here but it was not working when we last tested it.
se
Last edited:
Goooooood Morning....... DIY !!
I see everyone has been busy. having a good time, I hope. Just some random comments:
The high speed sample/hold of music waveforms compared with and without high DA cap showed the peaks lower and the later signal valley filled in. IMO that was caused by DA.
Again..... npo is not a high DA dielectric. It may be high K. But not high DA. aluminum electrolytics are very much higher in DA than all other types. So the non-linearity issue still does track with DA of the material.
THD does not tell anyone what the sonic character will be... sometime individual harmonics can be used to help characterized the distortion in listening. Only DA affects match the listening/sound characteristics of the caps and that also follows DA.
The way end terminations are made varies and the potential for bad connection are especially prevelent in the worst but cheapest methods. Probably spot welded leads are the best and most reliable. BTW --- pstyrene I have seen use spot welding because they cant take the heat of a metal plasma spray.... but distortion test during sampled runs will show it so that is rare in quality parts. That just leaves the dielectric.
In polar caps.... the 'wet' part is there to provide oxygen to the alum foil which repairs and creates the oxide insulator or dielectric. Alum oxide and a conductor back to back.
So. Where are we so far?
Actually, I'm going to let each individual claim their victory and let it go another 30 years.
THx-RNMarsh
I see everyone has been busy. having a good time, I hope. Just some random comments:
The high speed sample/hold of music waveforms compared with and without high DA cap showed the peaks lower and the later signal valley filled in. IMO that was caused by DA.
Again..... npo is not a high DA dielectric. It may be high K. But not high DA. aluminum electrolytics are very much higher in DA than all other types. So the non-linearity issue still does track with DA of the material.
THD does not tell anyone what the sonic character will be... sometime individual harmonics can be used to help characterized the distortion in listening. Only DA affects match the listening/sound characteristics of the caps and that also follows DA.
The way end terminations are made varies and the potential for bad connection are especially prevelent in the worst but cheapest methods. Probably spot welded leads are the best and most reliable. BTW --- pstyrene I have seen use spot welding because they cant take the heat of a metal plasma spray.... but distortion test during sampled runs will show it so that is rare in quality parts. That just leaves the dielectric.
In polar caps.... the 'wet' part is there to provide oxygen to the alum foil which repairs and creates the oxide insulator or dielectric. Alum oxide and a conductor back to back.
So. Where are we so far?
Actually, I'm going to let each individual claim their victory and let it go another 30 years.
THx-RNMarsh
Last edited:
I ran across an old preamp article by Mike Lampton, by day an astronomer specializing in X rays, and was amused to see tantalums, clearly being regarded as the best electrolytics available for other instrumentation. Therefore they must be the best for audio, right?
And there was a fairly small ceramic cap at the input of the MC240. When I repaired my brother's unit I replaced it with a PP, but probably ruined the value of the amplifier for connoisseurs of antiques.
But up to 100X worse than PS or PP Samuel's data does not show commensurate THD results. One industry guide lists X7R as low DA.
Sample & Hold? Moved the goal posts again, these days you won't find them in the signal path. It's almost a certainty that you will probably find as many horrifying caps in the signal path as 8-legs.
Last edited:
- Status
- Not open for further replies.