Maybe the attached will give something to chew on. Regarding current noise, you know what it is. Its noise generated by the flow of charge in matter.
Those documents deal with 1/f noise. What does 1/f noise have to do with your unsupported claim that Panasonic FM has bright sound?
You should use 22.000uF at least if you power an amp. Your blocker is OK for a CD player or small preamp. Also your cirquit is not the best, look at the one in the eBay offer.
Maybe a little progress. If we can get past the "capacitors can't make noise" nonsense, at least that's a start.
I know what noise is. In a DC blocker in which amps are flowing, 1/f noise is not a problem.
These 4 of 4700uF are the same as one 4700uF bipolar. The impedance at 50Hz is about 0.7ohm + ESR from the capacitor. So at least 1ohm. This is not good, I will take 5 to 10 times larger capacitors and bypass them with motor run 300uF or more. This will drastically reduce the negative impact on the sound. And I recently made a Quasimodo JIG to do a proper snubber on the secondaries.
I don't think 1/f noise is a problem either. I just grabbed that to show that caps do produce noise. And IME they can exhibit current noise at higher than 1/f frequencies.
In the time domain it can look like this:
Again, IME it depends on the particular caps and how much current you run through them.
In addition to noise, electrolytic caps already have other well known properties such as ESR, and ESL which may be nonlinear with frequency. They may also be microphonic and or have other physical properties not usually considered, such as electrostriction, triboelectric effects, piezoelectric effects, etc. If they have an audible effect in a DC blocker, then one has only to look at known physical effects to find the cause. Nothing new is needed.
In the time domain it can look like this:
Again, IME it depends on the particular caps and how much current you run through them.
In addition to noise, electrolytic caps already have other well known properties such as ESR, and ESL which may be nonlinear with frequency. They may also be microphonic and or have other physical properties not usually considered, such as electrostriction, triboelectric effects, piezoelectric effects, etc. If they have an audible effect in a DC blocker, then one has only to look at known physical effects to find the cause. Nothing new is needed.
Mark, those noise currents are 10^-7 ie 0.1uA.
How does that even feature in a a circuit where 2-3A might be flowing?
How does that even feature in a a circuit where 2-3A might be flowing?
Nobody is publishing research to exactly explain why particular caps in particular AC power systems, in particular DC blockers, powering particular power amps can have audible effects. Just like pretty much nobody is publishing research on cable sound, where particular cables connect particular equipment in a particular system enviornment. And when they do its usually ignored. There is nothing new here. Its always people assume idealized components at first. Then when non-idealities are pointed out, there is denial they can do anything that matters. Eventually, sometimes people figure out such things do matter.
You first denied caps have current noise, now you are at the stage where it can't do anything that matters under any conditions. Nor can any of the other non-idealities in electrolytic caps.
Bottom line, I stand by my claim that sometimes DC blockers can adversely affect sound. Same for ground loops, same for radiated EMI/RFI, same for a lot of things that can be hard to measure and that may be different under different test conditions. Just because that stuff may be different at my house than your house, and just because your ears are different than mine, doesn't mean that what you hear is the only true reality. You can deny anything and everything you personally never heard for yourself, but doesn't mean I should accept your opinion, not that you need to accept mine.
You first denied caps have current noise, now you are at the stage where it can't do anything that matters under any conditions. Nor can any of the other non-idealities in electrolytic caps.
Bottom line, I stand by my claim that sometimes DC blockers can adversely affect sound. Same for ground loops, same for radiated EMI/RFI, same for a lot of things that can be hard to measure and that may be different under different test conditions. Just because that stuff may be different at my house than your house, and just because your ears are different than mine, doesn't mean that what you hear is the only true reality. You can deny anything and everything you personally never heard for yourself, but doesn't mean I should accept your opinion, not that you need to accept mine.
Last edited:
Fair enough, but you could have arranged them so each cap pair totalled 9400uF, not 4700uF.
The mains is to be seen as an electrical cesspit, and it actually is most of the time. If a device is incapable of dealing with that, what difference a µV or sub-µV noise generated could make?
A decent power supply should be able to handle a 50 to 400Hz input, and ideally it should also work from white noise having the right average and rms values.
If something is designed in such a way that a µV or two can derail it, it is just crap and should be replaced or redesigned.
That's often the case with audiophile products costing a fortune...
A decent power supply should be able to handle a 50 to 400Hz input, and ideally it should also work from white noise having the right average and rms values.
If something is designed in such a way that a µV or two can derail it, it is just crap and should be replaced or redesigned.
That's often the case with audiophile products costing a fortune...
Mark, current noise in caps might be important if you’re dealing with uV or very low mV signals. It is of zero consequence in a circuit where currents are 7 orders of magnitude higher which is precisely why I asked ‘what is current noise?’
Ground loops and EMI are easy to measure and fix.
You’ve made claims about DC blockers. Please provide proper evidence to support them. That is all you have to do.
I have said there can be adverse audible effects. That is based on what I believe to be reliable listening tests. Ultimately what matters for audio is what is audible. One can measure any number of things and then argue about numbers. If one wants to rely on numbers, we may as well make it SINAD and leave it at that since its so easy to measure. Because that's were this is going.
Prove to me bursts of oscillation in your output stage are audible. I want to see formal DBT for each and every different waveform and frequency. That's all you have to do.
Its like this: when you hear something and or you believe something then its real. But how do I know you aren't imagining things? How do I know a scope waveform is audible? What I know is I can rely on SINAD because if the numbers are good enough then it guarantees audible transparency (this is me thinking like Bonsai or any EE who wants numbers he can believe in; if its hard to measure and or takes a lot of R&D to figure out how to measure, then it isn't real).
You know what some EE types say? They say a distortion residual isn't a measurement because it doesn't give a figure of merit number. Those guys want to see numbers. Numbers they can believe in. Not just pictures, not anything but numbers. Numbers from DBT studies are at least numbers, even if the EE can dismiss the study saying one study doesn't prove anything (one is a small number of studies, and small numbers are suspect). Let's see some independent replication studies with some more numbers, then maybe I will believe it.
Bottom line, all you're going to get is what I have given so far. If something more comes up later I will be happy to let you know.
Its like this: when you hear something and or you believe something then its real. But how do I know you aren't imagining things? How do I know a scope waveform is audible? What I know is I can rely on SINAD because if the numbers are good enough then it guarantees audible transparency (this is me thinking like Bonsai or any EE who wants numbers he can believe in; if its hard to measure and or takes a lot of R&D to figure out how to measure, then it isn't real).
You know what some EE types say? They say a distortion residual isn't a measurement because it doesn't give a figure of merit number. Those guys want to see numbers. Numbers they can believe in. Not just pictures, not anything but numbers. Numbers from DBT studies are at least numbers, even if the EE can dismiss the study saying one study doesn't prove anything (one is a small number of studies, and small numbers are suspect). Let's see some independent replication studies with some more numbers, then maybe I will believe it.
Bottom line, all you're going to get is what I have given so far. If something more comes up later I will be happy to let you know.
Last edited:
If there’s one thing that is audible in an amp, it’s oscillation. You of all people I would have thought would see that.
The whole idea about loop compensation is to ensure stability and you can readily test for that and make the claim ‘this amplifier is stable into this type of load [qoute load]’. No need for a DBT.
We are 5 or 6 posts into this exchange and you have still not offered a shred of evidence to back your claim that DC blockers affect the sound Other than they do what they are supposed to do, and block mains DC and stop transformer noise.
The whole idea about loop compensation is to ensure stability and you can readily test for that and make the claim ‘this amplifier is stable into this type of load [qoute load]’. No need for a DBT.
We are 5 or 6 posts into this exchange and you have still not offered a shred of evidence to back your claim that DC blockers affect the sound Other than they do what they are supposed to do, and block mains DC and stop transformer noise.
I never made any DC-blocker listening test, but my experience with conventional amplifier power supplies tells me, if anything is under dimensioned, it will be audible at some point. The better the amp, the more important the supply. The DC-blocker with the 4.700uF caps, in that configuration, should be audible.
Otherwise the whole thing about low inner resistance of power supplies would be pointless. Also, throtteling an amp, just to silence a transformer, seems a questionable thing, because you won't hear it any more when the music starts. After all, it is about music reproduction or did I miss a thing?
Otherwise the whole thing about low inner resistance of power supplies would be pointless. Also, throtteling an amp, just to silence a transformer, seems a questionable thing, because you won't hear it any more when the music starts. After all, it is about music reproduction or did I miss a thing?