Well, you only have to look at the measurements, you don't have to read the conclusions derived from them.
This is the same tired objectivist / subjectivist debate in a nutshell.
You can accuse Sy, Scott, etc. of bias or contempt if you like, but I know which side of this debate would reconsider their conclusion if presented with proper evidence.
This is the same tired objectivist / subjectivist debate in a nutshell.
You can accuse Sy, Scott, etc. of bias or contempt if you like, but I know which side of this debate would reconsider their conclusion if presented with proper evidence.
I will guess you mean Scott, because he seems the type that would be willing to change his mind in the light of new and compelling evidence. Am I right?
All we need is proper evidence that it is possible for at least some people to hear measured low level distortion associated with some capacitors.
Unfortunately, hearing and listening skills research is hard and expensive. So far, nobody wants to chip in on the cost. Earl Geddes says no one really cares. Well, maybe only a few people in audio forums it would seem. But, even they don't care enough to pay.
In addition, unfortunately we don't have a timeline on a couple of features that need to be added to blind aural skills testing software: looping and unsighted one-button switching. Again, nobody cares enough, as Geddes might say.
Lastly, I would say that when it comes to understanding humans and other biologicals, the so-called objectivists haven't always gotten it right the first time around, or the second.
Last edited:
Servo's are usually tuned to become effective below 1 Hz or something, and typically have 2nd order filtering properties AFAIK. The bleed through of a servo at low frequencies (~60dB @ 30 Hz, 72dB @ 60 Hz) therefore is higher than the aberrations that would be introduced by a coupling cap . Do the math, or in case of emergency, measure. Now, this bleed through does not per se introduce distortions, but it will cause phase shift.
It would be interesting if you can measure this.
IME according to cable direction the nature of the subjective distortions change, not the subjective distortions magnitude.
For replay neither direction is optimal, for guitar cable expect two different sounds, one direction subjectively more acceptable/more fun than the other.
Last night I ran some after show experiments on an electric guitar totally changing the sound for the better running unplugged, will wait for reports on how it performs on stage.
Dan.
Okay. What does it sound like? More or less objectionable than the coupling capacitor alternatives, assuming either one might be audible?
Also, I wonder how hard a servo usually has to work in a well designed amp with substantial heatsinks? Does LF phase audibly slew around much?
Last edited:
You want neutrality?Markw4 said:
Is this a neutral statement? People look for evidence of something when they already believe it to be true, despite lack of evidence. If there already was evidence, then they would not need to look; if they did not believe it, then they would not be motivated to look.Markw4 said:
That about sums up my whole life philosophy. I like to spend my limited time looking at things that are there
I believe that, as long as servos are in the feedback path, they are in the signal path. Whatever the return point of the signal that will correct the offset.
Using caps to extract DC from the output signal: the evils of those caps, if those evils exists, will always be in action.
The thing is that, as I am unable to hear them, I don't worry about evils hidden in film caps.
I use a 10µF as the first stage of one of my 18dB/oct servos. But, as it is followed by a 12dB/oct filter, that reject any frequencies higher than some 0.5 Hz, I do not really care about distortions it can produce ;-)
(BTW: no visible phase shift at 10Hz.)
Attachments
Last edited:
I don’t argue the audibility of capacitors, not because of my beliefs but lack of background or test equipment.
However I’m deeply confused as to why such talented folks are so highly invested in this issue which to me seems utterly trivial compared to other issues... rooms and loudspeakers for example.
It’s like focusing ones energy on prosecuting jaywalking.... waste of a law degree.
If all caps, including the best possible offerings from smaller boutique companies, were as inexpensive as a Vishay or Wima PP, I don’t think this conversation would continue to drag on.
Ditto if no forum contributors manufactured, sold products with or otherwise profited from these capacitors.
After all, you can’t call something snake oil if it doesn’t have a price tag.
Then it’s just called a home remedy.
Fully agree, but also look what happens at high frequencies. In an ideal servo, there is NO high frequency bleedtrough to speak of. For instance, if your cut-off is 1Hz, and the servo has (normally) a 6dB/oct slope, by the time you reach 256Hz, you're 8 octaves up so for that 256Hz the bleedtrough is -48dB. At 16kHz, its -78dB.
Now, this assumes that the servo opamp has enough loop gain, and is linearly enough, so that the -72dB bleedthrough at 16kHz causes just a minute change in frequency response. BUT if that opamp does not have enough linear loop gain, the bleedthrough can be a distorted version of the original signal and be much larger than -72dB. So this servo can easily cause more hf distortion than the whole rest of the amp!
Counter-intuitively, even if the servo has its main task at VLF, use a good opamp with high loop gain and low distortion at the hf part of the audio spectrum.
The upshot is also that a servo can easily distort orders of magnitude more than a good coupling cap. Also that a super-duper servo cap can be completely undone by an unfortunate choice for the servo opamp.
Jan
Last edited:
Unfair and unproductive enough.
I don’t focus on biases.
You know it’s close to futile to expect truly non-biased people exist (*).
I am only interested in those electrical measurements shown by Ed and SY and why one shows distortion and the other not. (whom I both thank, they moved a few steps away from the armchair talk)
George
(*) besides, biases is one of the motivations in moving our cart forward.
Please mind looking at the black and white horses in Plato’s Chariot allegory.
So are the things and can’t change, the horses have to be in front of the charioteer for the cart to move on.
Funny. Would not our ears be the best equipment to judge the audibility of something ?
Incorrect, on my point of view. And revealing the way of thinking of a part of the contributors of this forum.
Test equipment, mathematics and understanding tell us what has happened, They have no idea of what we can hear or not.
In fact, they have no idea at all ;-)
George,
As has been mentioned you must have a voltage drop across the capacitor to have the capacitor introduce distortion. My feel is that when the impedance of the capacitor matches the load resistance the distortion is at a maximum.
For example if I take a nonlinear resistor of one ohm that distorts any voltage across it by 10% and place it in series with 100,000 ohms that would put it down to -120 dB.
Same point has been made with capacitors, if you use them set the corner frequency way below the lowest frequency signal.
Then there is a level issue. Electrolytic capacitors have a large delta C with voltage. Manufacturer's used to also list a minimum voltage for them. That is so they do not age and drop value.
I will get around to showing an image of a 2.2 uF electrolytic capacitor, so you can judge for yourself.
As has been mentioned you must have a voltage drop across the capacitor to have the capacitor introduce distortion. My feel is that when the impedance of the capacitor matches the load resistance the distortion is at a maximum.
For example if I take a nonlinear resistor of one ohm that distorts any voltage across it by 10% and place it in series with 100,000 ohms that would put it down to -120 dB.
Same point has been made with capacitors, if you use them set the corner frequency way below the lowest frequency signal.
Then there is a level issue. Electrolytic capacitors have a large delta C with voltage. Manufacturer's used to also list a minimum voltage for them. That is so they do not age and drop value.
I will get around to showing an image of a 2.2 uF electrolytic capacitor, so you can judge for yourself.
- Status
- Not open for further replies.