This is mostly a consequence of THD being expressed as a percentage. A fixed voltage error around cut-off will be a higher percentage of a 4V signal than a 20V signal. The voltage error is inherent to class AB. Its severity depends on the magnitude.
Ed
Well https://linearaudio.nl/sites/linear...0 to 100uF caps and 100 Hz measurements_0.pdf
this research clearly shoows that there IS distortion in electrolytics!
this research clearly shoows that there IS distortion in electrolytics!
There can be distortion in electrolytic caps. The usual rule of thumb is to keep the AC voltage drop across the electrolytic to less than about 60mV at the lowest frequency of interest. That may require a rather high value cap.
The other thing about polar caps (electrolytic and tantalum) is that they can produce 1/f noise as a function of their leakage current.
The other thing about polar caps (electrolytic and tantalum) is that they can produce 1/f noise as a function of their leakage current.
Hi esl 63,
We are stuck with electrolytic capacitors. With some values there is no other practical way. So size the part properly, or correct the circuit so that you don't need large values if possible. As Mark pointed out, and is well known and repeated in so many places, distortion occurs after you develop a signal voltage across the capacitor beyond whatever voltage you choose to reference. Of course, balance that distortion against the distortion in the entire system at those frequencies.
Hi Mark,
Hi EdGr,
Yes, a percentage of the total signal. That's one reason why the industry measures distortion at 1 watt into 8 ohms as a standard. Not many people will listen to an average level of 10 watts even. It's way too loud!
We are stuck with electrolytic capacitors. With some values there is no other practical way. So size the part properly, or correct the circuit so that you don't need large values if possible. As Mark pointed out, and is well known and repeated in so many places, distortion occurs after you develop a signal voltage across the capacitor beyond whatever voltage you choose to reference. Of course, balance that distortion against the distortion in the entire system at those frequencies.
Hi Mark,
In any circuit that matters, you would never use an electrolytic or especially a tantalum capacitor where leakage current might be a problem. Ever. Then you have a specialized case that hopefully merits the cost of the alternative capacitor types in that value. Even Polypropylene has to be aged and soaked at the circuit potential before you can use them to measure very low noise. Physics can be such a PITA! lol!
Hi EdGr,
Yes, a percentage of the total signal. That's one reason why the industry measures distortion at 1 watt into 8 ohms as a standard. Not many people will listen to an average level of 10 watts even. It's way too loud!
I have aged polymer caps in an environment accellerator (High temp high humidity) and I made some Arrenius calculations of lifetime in products.
Just recently i tested 2,2uF 320V AC X-Caps with a bias voltage in 90 deg C and 90%RH test chamber.
ESR went up 100% after only 200 hours. This is comparable to less than 2 year in Malaysia.
At 1000 Hours 2,2uF became 60nF... and ESR went up to >1kOhms at 1200 Hours it was 20nF 42kOhms representing less then 10 years in Malaysia climate (outdoors).
So my recommendation is, to increase voltage spec! And use metal foil caps, and choose a polymer that is less hygroscopic.
Run your stuff cool, and dry.
And turn the equipment off when you are not using it for a longer time.
Just recently i tested 2,2uF 320V AC X-Caps with a bias voltage in 90 deg C and 90%RH test chamber.
ESR went up 100% after only 200 hours. This is comparable to less than 2 year in Malaysia.
At 1000 Hours 2,2uF became 60nF... and ESR went up to >1kOhms at 1200 Hours it was 20nF 42kOhms representing less then 10 years in Malaysia climate (outdoors).
So my recommendation is, to increase voltage spec! And use metal foil caps, and choose a polymer that is less hygroscopic.
Run your stuff cool, and dry.
And turn the equipment off when you are not using it for a longer time.
I had a quick scan through all of the posts and I couldn't find a reference to Cyril Bateman's "Capacitor Sounds" series of articles in Electronics World starting in July 2002. It is pretty comprehensive and he made some pretty sophisticated instruments to measure distortion caused by a wide range of capacitors. Apologies if someone has already mentioned it.