Yes, however IMHO they were both mistaken about DA not being able to materially affect sound (PSS distortion at audio frequencies is IME much, much less common). IMHO and IME DA is more complex than the model that was commonly used. May also require consideration of the threshold of audibility of group delay in a particular circuit, as well as signal-correlated noise.
As far as linear waveform distortion of asymmetric signals due to DA, that was shown long ago and wrongly dismissed on the basis that humans can't ever hear phase.
Regarding signal-correlated noise, mostly we only heard about resistor current noise, and not much about its perception in humans. Intermodulation of noise with signal in semiconductor junctions was also known, but usually there was some PSS evidence on an FFT. If it looked too much like noise then it might be missed. With the advent of SD dacs we now have a lot more ways to generate signal correlated noise problems.
Don't think anyone ever studied signal correlated noise as a possible perceptual issue related to DA disequilibrium noise possibly with changing voltage (for want of better terminology; say, electrons migrating a bit in and out of dielectric traps as voltage changes). Wouldn't be surprised if there can be some of that going on.
Last edited:
Are the metalized polypropylene caps considered to have a direction or polarity? I’ve seen videos explaining that some capacity have a polarity in terms of less interference even if they function each way. Is this relevant to polypropylene film caps and is it a good idea to test with a oscilloscope?
Dear Matt,
you can get an answer to that if you read the appropriate chapters of H&H. The reason that's a good idea is that it's written by a professor of physics. He doesn't care about hi-fi and has no opinion on the sound quality of one versus another. When he doesn't know the answer, he says "We don't understand the mechanism for how...." rather than make a stand on a topic he can't back up with a sensible explanation.
kind regards
Marek
you can get an answer to that if you read the appropriate chapters of H&H. The reason that's a good idea is that it's written by a professor of physics. He doesn't care about hi-fi and has no opinion on the sound quality of one versus another. When he doesn't know the answer, he says "We don't understand the mechanism for how...." rather than make a stand on a topic he can't back up with a sensible explanation.
kind regards
Marek
A film capacitor is a tightly wound-up coil of two metalized film layers. At the outer end, one layer acts as shield of the layer beneath. So the outer layer is most likely to pick up stray electric AC field. It is wise to connect the outer foil (and the connected leg) to a low impedance point, to ground if possible.
However a capacitor acts as a shunt of AC voltage between its legs, depending on the frequency. The above is for low frequency, where the capacitive reactance is high, and the shunt effect does not work. For coupling capacitors it is all the same, IMHO.
While a interesting article with useful testing it only focuses on one capacitor aspect under conditions that are encountered only in some cases of audio use.
It is helpful to understand how capacitor impairments affect the audio path in actual use and not to generalize test results to "best in all cases"
The test would be very useful in evaluation of a capacitor for a zoble network on a transformer primary where exposure to a large AC signal is the normal condition. DF is also important in this application.
Also a useful test for a capacitor used to provide high frequency feedback network compensation where again the capacitor operates with a significant AC voltage across it in normal operation.
These tests would be important in a capacitor used to provide high frequency roll off in the driver gain stage where again the capacitor operates with the majority of the AC signal across the capacitor.
The test is of less value for coupling capacitors that in normal operation have little AC voltage across them and so any capacitor non-linearity will have only a small contribution to the total signal passing through the capacitor. This is the application that gets the most focus.
The sensitivity of any amplifier design to coupling capacitor nonliterary is strongly affected by the ratio of the capacitor reactance to the total resistive portion of the circuit reactance meaning any advantage will be strongly affect by the amplifier design.
Coupling capacitors should also posses fairly low self inductance for good high frequency operation especially in amplifiers using global feedback, something that I am not sure the very large wound capacitors sold for coupling capacitor upgrades will posses.
As coupling capacitor physical size increases so does stray capacitance increase causing more coupling to other circuits increasing possible frequency response errors and amplifier stability issues. Finally long term leakage current is so important in coupling capacitors connected to the grids of power tubes in preventing amplifier "melt downs". Leakage current seems rarely specified in "audiophile " capacitors.
Lastly for de-coupling capacitor's linearity is often of low importance compared to total value, ESR, self inductance and resonant frequency.
Sure a more linear capacitor will not hurt however if there is any useful change in amplifier performance (or harm) is very dependent on the design and application at hand.
Not to dissuade anyone from attempts to update their equipment however results will improve if the context of the application of the capacitor is carefully considered in the capacitor's selection.
This sums up capacitor selection very succinctly and explains the source of so much of the controversy.
Much of the audible differences are the result of cost bias. The more it costs the better it is perceived to sound. It’s more money, it must be better! This is where double blind tests are useful. Also, the best testing sap may not be the one that sounds best to you, the only criteria that ultimately matters. I build things to enjoy, not quote the specs on. This happened after I built an amazing amp with spec and parts that were the best money could buy. It was technically superior in every way. The sound sucked! It was lifeless and sterile to me! Bottom line, make stuff to enjoy not to impress anyone but you! Ifit sounds good it is good. Just my opinion, others will disagree and that’s fine.
Hey guys, thank you all for your help I enjoyed learning from you all. I ended up getting some solen sm caps. Just finished finding and marking the outside foil on each and will install the new caps soon.
Hi Metalmatt,
Below is a link with an extensive article concerning different types of capacitors. It compares different
insulating materials, as the DA of those materials are quite different. HD is only one parameter, as conductor
material (foil and leads), conductor material thickness, termination techniques vary and affect sonic accuracy.
However, the article does not recommend any particular brand as:
1. It was written in 1980, long before many brands in existence today.
2. They did not perform scientific, specialized listening tests to determine capacitor accuracy in "absolute terms".
Hope this link helps.
https://milbert.com/Files/articles/Picking_Capacitors_1.pdf
pos
Below is a link with an extensive article concerning different types of capacitors. It compares different
insulating materials, as the DA of those materials are quite different. HD is only one parameter, as conductor
material (foil and leads), conductor material thickness, termination techniques vary and affect sonic accuracy.
However, the article does not recommend any particular brand as:
1. It was written in 1980, long before many brands in existence today.
2. They did not perform scientific, specialized listening tests to determine capacitor accuracy in "absolute terms".
Hope this link helps.
https://milbert.com/Files/articles/Picking_Capacitors_1.pdf
pos
Here is part 2 of "Picking Capacitors"
https://milbert.com/Files/articles/Picking_Capacitors_2.pdf
Capacitors, a chemist's view
https://audience-av.com/oem-diy-parts/axo_description/a_chemistview-html/
As one can read, there is more than DA or HD.
Hope this helps.
pos
https://milbert.com/Files/articles/Picking_Capacitors_2.pdf
Capacitors, a chemist's view
https://audience-av.com/oem-diy-parts/axo_description/a_chemistview-html/
As one can read, there is more than DA or HD.
Hope this helps.
pos
what is the opinion of use of DC-LINK, EPCOS etc. MKP/CBB in power supplies, for tube amplifiers?
In the past (1980) I found that MKP 5mF motor cap // to 50uF electrolytic did wonders.
but nowadays the big blue MKP/CBB boxes pitch 27, 37 mm and larger wash the market. example: B32778G0406K000, 40µF foil.
ARE THEY GOOD?
In the past (1980) I found that MKP 5mF motor cap // to 50uF electrolytic did wonders.
but nowadays the big blue MKP/CBB boxes pitch 27, 37 mm and larger wash the market. example: B32778G0406K000, 40µF foil.
ARE THEY GOOD?
Then, if the cap is used as reservoir - after a capacity multiplier - will that not have a good sound? I thought the low impoedance of a source say 5 Ω would overcome those problems.
Otherwise a small cap ex. Mylar/Polyester say 10µ is better in the same situation - .
The new generation MKP is designed for use in automotive situations, after a SMPS, so focus on high frequency and highest absorption.
Otherwise a small cap ex. Mylar/Polyester say 10µ is better in the same situation - .
The new generation MKP is designed for use in automotive situations, after a SMPS, so focus on high frequency and highest absorption.
Don't know. Haven't tried them for use in a multiplier. Usually those can use an electrolytic bypassed with a small wima cap (maybe .01uf) cap at the multiplier node. Input and output caps for stability can usually just be electrolytic. Sound may depend on which brand and model electrolytic to some extent. Also, when new they may need some time to settle in. They can make some 1/f noise until initial leakage currents settle down to more of a long term value. Depends on the specific chemistry.
That's interesting, something that could be expected. Win something. Lose something.
Translated into use, that would imply that the input from the rectifier bridge should preferably be 'on' all the time, so there is no burst in on/off up to 400V (like 300B) or 800V (like 211), beceause then the fragile MKP can become unstable, might even deteriorate (but self-heal). And while the level settles it would be prone to small electric charge waves and ripples. In the capacitor.
- In the past I used 4.700µF electrolytics cans as reservoir.
So my thought is that this MKP stuff must be an improvement.
Translated into use, that would imply that the input from the rectifier bridge should preferably be 'on' all the time, so there is no burst in on/off up to 400V (like 300B) or 800V (like 211), beceause then the fragile MKP can become unstable, might even deteriorate (but self-heal). And while the level settles it would be prone to small electric charge waves and ripples. In the capacitor.
My intent is simple: a CCS of 1mA charges a high value resistors (of 400K@300B; or 800K@211) and there is a small 100nF cap. This works very well; I have that running for years. Then this goes via 10K to 4µF Mylar and that to the gate of a high voltage Mosfet. It is very stable. The output us another 4µF that goes to MKV 40µF via a 20Ω resistor. So I have cells.
Now I was thinking to supercharge this with large MKP reservoirs (40µF to even 110µF) at the ouput too.- In the past I used 4.700µF electrolytics cans as reservoir.
So my thought is that this MKP stuff must be an improvement.
Attachments
If you suspect a turn-on surge may overload the film cap voltage rating, probably you need a better cap and or soft start circuit to limit the surge.
in sim my start up is > 300mSec, that isn't too bad. it is the settling that worries me.
these caps look cheap but the cost is easy $50 in total so I like to learn about the hidden 'qualities'.
If indeed 'vague' in presenting a stage, that is very scaring.
these caps look cheap but the cost is easy $50 in total so I like to learn about the hidden 'qualities'.
If indeed 'vague' in presenting a stage, that is very scaring.
With capacitors less is more. Swapping to a lower value of coupling capacitor of the same type and adjusting the following resistor to ground results in a clearer sound.
For example if the coupling capacitor you are looking at is 0.22uF followed by a 470k resistor to ground the -1dB figure is 13.85 Hz.
You can now play around with maths as the formula is fixed and the following are the -1dB figures;
0.022uF into 470k = 138.5 Hz
0.47uF into 470k = 6.48 Hz
0.22uF into 220k = 29.6 Hz
0.047uF into 220k = 138.5 Hz
0.1uF into 220k = 65.1 Hz
If you raise the resistor value the roll off is at a lower frequency for the same coupling capacitor value and if you raise the capacitor value the roll off is at lower frequency for the same value resistor to ground.
Since less is more ore simply if you halve the capacitor value you then double the following resistor value to ground for the same roll off curve.
From the above examples we have two for 138.5 Hz, showing how you can swap over the values of resistance and capacitance to obtain the same result.
Caveat emptor. The global feedback and associated output transformer need the same time constants or one at a higher frequency to avoid oscillation.
The other issue many overlook is the frequency response of your loudspeakers. You wouldn't need a -1dB figure of 6.48 Hz, because it's never going to get recorded, even if it existed and your speakers might well struggle to be -1dB even at 50Hz.
Selecting a smaller value capacitor sounds better, gives you speaker an easier time, so it also sounds clearer, a more stable amplifier with less transformer steel losses and therefore less output transformer distortion, as steel losses or increased transformer flux are adding 3rd harmonic distortion even with M6 or Nickel Iron.
The other caveat emptor is if your valve / tube amplifier is grid biased, ie it has a negative voltage applied to the input grid rather than using cathode bias the following resistor should be unaltered. The Dynaco ST70 being such an example with 0.1uF into 270k being -1dB at 53 Hz.
On Capacitor sound Polystyrene are audibly better for RIAA than even Polypropylene.
One last point. Resistors add distortion given away by there ppm figure, metal films are best used to ground though, as they harden the sound if used as anode resistors. 2 watt carbon film are a good upgrade for anode resistors as ppm is related to wattage, so if the ppm is the same but the wattage is greater it follows the distortion is reduced.
I hope some of that helps
For example if the coupling capacitor you are looking at is 0.22uF followed by a 470k resistor to ground the -1dB figure is 13.85 Hz.
You can now play around with maths as the formula is fixed and the following are the -1dB figures;
0.022uF into 470k = 138.5 Hz
0.47uF into 470k = 6.48 Hz
0.22uF into 220k = 29.6 Hz
0.047uF into 220k = 138.5 Hz
0.1uF into 220k = 65.1 Hz
If you raise the resistor value the roll off is at a lower frequency for the same coupling capacitor value and if you raise the capacitor value the roll off is at lower frequency for the same value resistor to ground.
Since less is more ore simply if you halve the capacitor value you then double the following resistor value to ground for the same roll off curve.
From the above examples we have two for 138.5 Hz, showing how you can swap over the values of resistance and capacitance to obtain the same result.
Caveat emptor. The global feedback and associated output transformer need the same time constants or one at a higher frequency to avoid oscillation.
The other issue many overlook is the frequency response of your loudspeakers. You wouldn't need a -1dB figure of 6.48 Hz, because it's never going to get recorded, even if it existed and your speakers might well struggle to be -1dB even at 50Hz.
Selecting a smaller value capacitor sounds better, gives you speaker an easier time, so it also sounds clearer, a more stable amplifier with less transformer steel losses and therefore less output transformer distortion, as steel losses or increased transformer flux are adding 3rd harmonic distortion even with M6 or Nickel Iron.
The other caveat emptor is if your valve / tube amplifier is grid biased, ie it has a negative voltage applied to the input grid rather than using cathode bias the following resistor should be unaltered. The Dynaco ST70 being such an example with 0.1uF into 270k being -1dB at 53 Hz.
On Capacitor sound Polystyrene are audibly better for RIAA than even Polypropylene.
One last point. Resistors add distortion given away by there ppm figure, metal films are best used to ground though, as they harden the sound if used as anode resistors. 2 watt carbon film are a good upgrade for anode resistors as ppm is related to wattage, so if the ppm is the same but the wattage is greater it follows the distortion is reduced.
I hope some of that helps
Keep in mind that tubes have a rated maximum value for the grid resistor, and that maximum value was often the value chosen (sometimes several times that value in guitar amps!).
And also consider that the percentage effects of capacitor non-linearity and dielectric absorption decrease with the AC voltage across the capacitor (near zero for a coupling capacitor). For tone control and crossover capacitors, there may be a considerable portion of the rated AC voltage across the capacitor and such effects may be measurable. Some high-K ceramic dielectrics like Z5U have major changes in value with temperature and applied DC voltage; Excluding those, no one has proven any measurable difference in coupling capacitor performance within the audible range with different dielectrics.
And also consider that the percentage effects of capacitor non-linearity and dielectric absorption decrease with the AC voltage across the capacitor (near zero for a coupling capacitor). For tone control and crossover capacitors, there may be a considerable portion of the rated AC voltage across the capacitor and such effects may be measurable. Some high-K ceramic dielectrics like Z5U have major changes in value with temperature and applied DC voltage; Excluding those, no one has proven any measurable difference in coupling capacitor performance within the audible range with different dielectrics.