|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
67.189.165.161
I stumbled across this one a few years back and since TK brought up the "AC signal across a cap" I figured it might be worth a discussion.
source is some electronic music and the load is stepped from 10K (green) to 1meg (red) and the output is subtracted from the input to show the "effect" of the cap.
thoughts??
dave
Follow Ups:
Um.
Unless I misunderstand what you are concerned about, that just looks like the capacitor is behaving as a high-pass filter against the varying load resistor. The filter corner frequency (and associated phase shifts) depend on the product of R*C. It's not a nonlinear distortion, it's the characteristic of music being played through a filter of different characteristic. That's pretty much the reason for capacitors to exist, it's what they do for a living.
Subtracting one signal from another is an EXTREMELY sensitive test, anything different, particularly including linear filtering responses, will show dramatically.
Unless I'm missing something.
I suspect you're correct regarding the results as simply the output of a simple RC filter passing music. So the next simulation should be performed with the 10K load and vary "C" @1X, 10X & 100X and then compare the "residual" to 1XC@R = 10K, 100K and 1000K.Knowing more about the model would remove the need to speculate.
Edits: 12/04/14
.
Have Fun and Enjoy the Music
"Still Working the Problem"
...before there can be any meaningful discussion of simulation output?
i.e. Is the cap theoretically perfect or with real world "defects"? If real world, what defects were incorporated into the model and how were they simulated. etc. Seems to me that without more imformation regarding the model, discussion of its output is based on speculation and preconceived notions.
As a concept we may find some validity in the results, at least in principle. Obviously it's virtually impossible to predict a result for individual caps/brands because we don't know all of their parameters. There's the obvious lead inductance etc but it's possible that even the capacitance itself may vary slightly with frequency.
As you know, Stephie did a great job years back in comparing at least one form of distortion in caps where she showed comparative plots that resembled hysteresis curves. Well worth a read for anyone who already hasn't.
I'd be interested in the comparative fft results when adding degrees of L and R that might be expected for the cap ... bearing in mind that Spice itself can "generate" distortion in the AC voltage source if the time step settings aren't right (I know you know this as I learned it through your forum but just added for the sake of a meaningful discussion in case someone wants to play with this).
Naz
Is that the simulation shows distortion at all, and that this distortion is so depended on the load.
I would very much like to learn the mechanisms at work here,
Very interesting sim Dave.
NC
"since TK brought up the 'AC signal across a cap'"
The issue I raised was intended to address capacitors used in bypass/decoupling, such as at the cathode or power supply output. I believe there's a threshold involved regarding the amplitude of signal voltage that appears across those caps that's related to whether they're audible. Applying that concept to a coupling cap is more difficult from a measurements perspective, because neither terminal is grounded and both terminals carry signal. That aside, I don't believe all the characteristics that cause caps to color sound can be synthesized in SPICE. We may not even fully understand what those characteristics are, and I was not attempting to open a discussion on that point.
The main issues of interest to me in this regard are (first) agreement that a bypass capacitor with no AC voltage across it cannot contribute its own sound, even if AC signal current is flowing through it. This concept is fundamental to everything that might follow. I also believe it might be possible to (second issue) eventually determine the approximate threshold of audibility mentioned above for capacitors in specific circuit positions, based on the signal voltage across them, signal current through them, and perhaps signal voltage levels present elsewhere in the circuit. This would be a much different discussion than what you've initiated, so I'll wait to start it in its own thread.
--------------------------
Buy Chinese. Bury freedom.
So, how do we test this allegedly fundamental concept?
Perhaps more importantly: what would constitute proof or failure? I glean from the majority of your posts that how something sounds is of no concern.
"I glean from the majority of your posts that how something sounds is of no concern."It's of utmost concern but who decides?
If this hobby is completely subjective then all bets are off.
When someone says "it sounds good" I want to know what they mean and who they are and that they are experienced with the sound of real instruments.
I hate to even bring this up because it's controversial (only because a lot of people don't understand what Harry meant) but The Absolute Sound is the sound of acoustic instruments in a real space.
The sound of real acoustic instruments is the closest thing we have to a reference.
Without a deep personal experience with TAS, when someone says "it sounds good" they could just be saying that they like the sound.
We all have personal preferences. Before accepting a statement that "it sounds good" I want to know that the person saying that is educated in what real instruments sound like and that they are doing there best to repress their own preferences.
I have made technical improvements to my system that, at first listen, I didn't like, but the longer I listened the more I realized (and had to admit) that the piano now sounded more like a real piano, the acoustic guitar now sounded more like a real guitar, etc. What I "like" should not influence my choices as I build my system.
This brings to mind a cover of The Sensible Sound magazine, winter 1980/81
It pictures a guy sitting in the audience in a concert hall and a grand piano on stage being played by a guy in a tux. We can assume he's a professional pianist.
On the back of the seat in front of the audience member are a set of knobs, bass, treble, balance, volume. He's is sitting there adjusting them.
Does that make any sense to you at all?
Here's what I've learned from 35 years of building audio gear, the more I make my system technically correct, the more the playback sounds like the real thing.
P.S. I've been a musician since I was young and was a recording engineer for 15 years working with national acts. It's a short walk from the control room to the studio. The sound of the instrument itself is of utmost concern.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 12/04/14 12/04/14 12/04/14 12/04/14
Ggg
"I glean from the majority of your posts that how something sounds is of no concern."
Most listening "tests" reported here and on other forums are not about how something sounds. They've been conducted in such a manner as to entirely void the supposed results. They are indeed of no concern, except for the fact they foster myth and misconceptions about the value of various components and techniques.
As for the capacitors, how much signal (music) voltage on the hot side of the last filter would be a likely indicator that the cap is contributing its own character to the music? One microvolt? One hundred microvolts? Ten millivolts? I think the answer could be derived with a series of legitimate listening tests, but that would be extremely time consuming, expensive, and a lot of hard work. Surely there's a ratio where audibility would seem doubtful. For example, if the final amplifier tube outputs 200v rms at its anode, say at 1 kHz, and the last filter cap exhibits one microvolt across its terminals, can we dismiss the cap as inaudible? The ratio under those conditions between the desired signal and the undesired signal at the cap (assuming the microvolt of capacitor signal consists wholly of distortion) would be 166 dB. Does anyone believe we can hear distortion or color at -166 dB?
--------------------------
Buy Chinese. Bury freedom.
Did you specify the correct polarity of the non-polar cap? :-)
-
What parameters for the cap account for differences between types? And what parameters are loaded in your sim? And where do you find that info?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: