Hi diyiggy,
Possibly, I do not know at this time. They are normally used for power supply decoupling and directly replace tantalum types.
Hi Salas,
Yes, an HP 4195A along with the impedance test fixture. It is a real pain to use to be honest. I think I will test these caps in an existing circuit and compare distortion with only this part replaced. That way you have real world comparisons in the circuit people will tend to use them in. The cathode bypass application will wait until I get something in suitable to test with. The other application is a very sensitive application in the feedback network of many solid state amplifier and signal circuits.
Possibly, I do not know at this time. They are normally used for power supply decoupling and directly replace tantalum types.
Hi Salas,
Yes, an HP 4195A along with the impedance test fixture. It is a real pain to use to be honest. I think I will test these caps in an existing circuit and compare distortion with only this part replaced. That way you have real world comparisons in the circuit people will tend to use them in. The cathode bypass application will wait until I get something in suitable to test with. The other application is a very sensitive application in the feedback network of many solid state amplifier and signal circuits.
Yep, thanks...I know they have less esr and better ripple and longer life (due to the reduced esr?) , are used in smps but do not see them much so I asked. I had once a good experience with one on them A-PSA serie for a dac chip but I really wonder if usefull but smps in our hobby ? Maybe lythics will disseaper as industry just do not care of audio....
No, electrolytic caps are superior in some ways. They are here to stay.
I have tested Polymer caps for various things. ESR isn't a concern, they simply have much better HF response. But, with higher leakage current they are completely unsuitable for some applications. Read some datasheets and note what is said from the manufacturer. There is no mention for audio applications, but that doesn't mean they wouldn't be suitable.
ESR values only matter for power supply storage applications and bypass. In audio, ESR values are completely pointless. If you compare the ESR value to the circuit impedance, you will see what I mean. What matters is anything that is non-linear. D-A is the critical value.
I have tested Polymer caps for various things. ESR isn't a concern, they simply have much better HF response. But, with higher leakage current they are completely unsuitable for some applications. Read some datasheets and note what is said from the manufacturer. There is no mention for audio applications, but that doesn't mean they wouldn't be suitable.
ESR values only matter for power supply storage applications and bypass. In audio, ESR values are completely pointless. If you compare the ESR value to the circuit impedance, you will see what I mean. What matters is anything that is non-linear. D-A is the critical value.
See page 26:
https://www.analog.com/media/en/technical-documentation/application-notes/an47fa.pdf
This is for high freq stuff. Audio, hopefully, won't be as critical.
https://www.analog.com/media/en/technical-documentation/application-notes/an47fa.pdf
This is for high freq stuff. Audio, hopefully, won't be as critical.
Hi Conrad,
Read that a long time ago, really valuable hints in there. I don't recall much on capacitor non-linearity. Obviously time to give it another read or at least skim.
I have seen unfortunate pairing cause problems. Normally you need an inductive electrolytic capacitor coupled with a nice ceramic with high Q to do that. Adding a higher value film type sorts that out.
High speed techniques are important in sub-sections in audio equipment when the individual parts can misbehave at high frequencies. Normally well beyond the pass band of the entire circuit if they act up. Normally it takes very poor layout to get these things to take off.
Read that a long time ago, really valuable hints in there. I don't recall much on capacitor non-linearity. Obviously time to give it another read or at least skim.
I have seen unfortunate pairing cause problems. Normally you need an inductive electrolytic capacitor coupled with a nice ceramic with high Q to do that. Adding a higher value film type sorts that out.
High speed techniques are important in sub-sections in audio equipment when the individual parts can misbehave at high frequencies. Normally well beyond the pass band of the entire circuit if they act up. Normally it takes very poor layout to get these things to take off.
Last edited:
Roughly for our purposes, DA represents energy lost in the dielectric, and it is going to be non-linear. That means it will add distortion to the signal. How much of an effect depends on the circuit itself and how much signal is across the part. For a coupling capacitor, there is rarely any substantial signal across it. What does that mean? I'm not going to tell you, think on this. Some of you will actually realise the entire picture - like a light bulb going off.
I've never been sure about DA affecting audio. It's a huge deal with low frequency integrators. I've found opinions that it doesn't matter at all, and opinions just the opposite. At higher frequencies, in terms of measurement, I don't know how to separate it from DF. Regardless, there's a simple test one can do to shed some light on this stuff, but it requires a high gain scope preamp or plug-in. Basically, you feed a square wave through a resistor and the cap under test. Get the frequency and ratios right and the cap will charge and discharge more or less linearly, giving you a triangle wave on the scope. Some caps will give you a nice triangle (typically low DF ones) and some will give you a wildly distorted waveform (typically tantalums). It may or may not matter for audio, particularly if there's no substantial signal across the cap. Just fun stuff to do at the bench when fixing things is too difficult.
Conrad, you did the measurements showing third harmonic effects of polyester caps. Looking at vintage guitar amplifiers and effects pedals you see polyester caps everywhere. Yes, it was the state of the art of that era but it is an interesting argument in favor of the audibility of capacitor dielectric/construction.
It's easy to be led astray with this stuff but there's a lot of anecdotal reports about polyster (Mylar) caps being less than great. I have to concur with that but remember what @anatech says in 2129- it's critical to the whole discussion. Application and circuits matter.
That third harmonic stuff I did turns out to reflect DF more than anything. With the better caps (polypropylene and such) the discrepancies were due to test oscillator distortion! It's actually nice that fairly simple and traditional cap measurements predict performance quite well. Now if I could just get people to think in terms of DF, rather than the fad of esr, things would be easier.
That third harmonic stuff I did turns out to reflect DF more than anything. With the better caps (polypropylene and such) the discrepancies were due to test oscillator distortion! It's actually nice that fairly simple and traditional cap measurements predict performance quite well. Now if I could just get people to think in terms of DF, rather than the fad of esr, things would be easier.
Thanks for that!
If the signal level is small, I’m not sure how the DA affects it because, presumably, it is a property of the dielectric. The Pease article mentions something about how fast the signals are and how that affects distortion.
And also vintage amps like the Leak's, the Quad's, and so forth. And carbon composition resistors.
And if you dare replace them with contemporary parts, the collectors will burn you: the sound is shrill and hard, they say. Yet another proof, or rather reverse proof of audibility, even if only 10% of these guys are trustworthy.
A guitar amp is a totally different thing. With that you are creating a sound and distortion is desired. We here are mostly reproducing sound and there shouldn't be any added distortion.
Whenever I had talked to older engineers that designed the classics, they used the best parts they had, and today they wouldn't dream of using the old parts in deference to the new ones. A vintage amp sounds the same, but far better rebuilt with quality components we have today. They still sound like whatever they are, just a lot cleaner and quieter. To those who argue that they want the equipment to sound like it was intended, the intention was to have as low a distortion as possible. The folks who designed the stuff to start with themselves prefer the new components. They don't even know what was intended for crying out loud!
Hi Conrad,
Measurements have confirmed lower distortion using low DA parts in the right locations. Otherwise I would not have said anything.
Whenever I had talked to older engineers that designed the classics, they used the best parts they had, and today they wouldn't dream of using the old parts in deference to the new ones. A vintage amp sounds the same, but far better rebuilt with quality components we have today. They still sound like whatever they are, just a lot cleaner and quieter. To those who argue that they want the equipment to sound like it was intended, the intention was to have as low a distortion as possible. The folks who designed the stuff to start with themselves prefer the new components. They don't even know what was intended for crying out loud!
Hi Conrad,
Measurements have confirmed lower distortion using low DA parts in the right locations. Otherwise I would not have said anything.
Yes, I left the polyester caps in both my Fender Champ and my Pignose G60 guitar amps
I beg to differ: for me, "cheapest" would be closer to the truth, at least through the '50's. Leo Fender's story is well documented, and, IIRC, Patrick Turner (RIP) did much ranting regarding Leak's bean counting approach. There were only a few exceptions, with the like of Marantz and McIntosh.
I was originally trained at Marantz, and was authorized warranty for McIntosh. These among other well respected manufacturers, so my mindset comes from that direction.
Some parts would not survive the mass production process. They couldn't use them from a practical standpoint.
Some parts would not survive the mass production process. They couldn't use them from a practical standpoint.
Or the cost of polyester caps.
And depending what "vintage". The 1950s Fenders were paper/oil and then various plastics. Even the earliest "Orange Drops" were not the guts we know now (also rarely used in factory production because cost).
The first victim isn't an "orange drop" for starters. Those should be polypropylene. He didn't identify the plastic dielectrics.
The "paper" types are often leaky. They were known to be leaky in the '70's, so nothing new there. That You Tube clip wasn't very informative at all. It doesn't take much to open up a capacitor and unroll it, and a video was made from this? Really?
The "paper" types are often leaky. They were known to be leaky in the '70's, so nothing new there. That You Tube clip wasn't very informative at all. It doesn't take much to open up a capacitor and unroll it, and a video was made from this? Really?