Will the harmonic performance of a capacitor degrade if there is a DC voltage over it in addition to the AC signal component?
I am designing a passive lowpass filter for a DAC I-V converter. I get the simplest layout by allowing a DC bias voltage over the capacitor. The DC is much below the voltage rating of the capacitor.
Should I rather go for 0V DC in addition to the audio signal?
Regards,
Børge
I am designing a passive lowpass filter for a DAC I-V converter. I get the simplest layout by allowing a DC bias voltage over the capacitor. The DC is much below the voltage rating of the capacitor.
Should I rather go for 0V DC in addition to the audio signal?
Regards,
Børge
sreten.
janneman said:
That a whole lot of pages to flip through. I was hoping for something like a tutorial or recomendation on what capacitor to use for what application. My case now is to find a cap in the 1nF to 1uF range where some DC may be superimposed on the signal.
Mr Bateman, if you're reading this, you should publish a "Capacitor Pocket Reference".
--
Børge
Hi Børge, DC levels may make a difference with some kinds of capacitors, even non-polarised ones. With some dielectrics, especially some ceramic types, the dielectric may "saturate", and also have hysteresis. However, C0G/NP0 capacitors are considered to be quite good. For simple DC bypassing, an aluminium electrolytic is usually good enough.
Remember that you will achieve the best performance by keeping the design reasonable. By this I mean that there is no point going to extreme lengths trying to solve one problem, because in the process you may create 10 other problems.
As a thought-experiment consider the use of a capacitor with an air-dielectric. They say it's the ultimate dielectric, but who would ever consider that mechanical vibrations may be introduced by the varying charge stored across the capacitor? Because the dielectric is not a solid, there is nothing apart from the stiffness of the plates to reduce vibrations, and this could be a source of distortion and resonances in the output signal. Anything that is put between the plates to reduce vibrations, will obviously introduce the problems that other dielectrics have.
In critical applications there are all sorts of other things to consider too. Look up "dielectric absorption" for example. If the capacitor is radial or axial, the parasitic inductance could also be an issue.
CM
Remember that you will achieve the best performance by keeping the design reasonable. By this I mean that there is no point going to extreme lengths trying to solve one problem, because in the process you may create 10 other problems.
As a thought-experiment consider the use of a capacitor with an air-dielectric. They say it's the ultimate dielectric, but who would ever consider that mechanical vibrations may be introduced by the varying charge stored across the capacitor? Because the dielectric is not a solid, there is nothing apart from the stiffness of the plates to reduce vibrations, and this could be a source of distortion and resonances in the output signal. Anything that is put between the plates to reduce vibrations, will obviously introduce the problems that other dielectrics have.
In critical applications there are all sorts of other things to consider too. Look up "dielectric absorption" for example. If the capacitor is radial or axial, the parasitic inductance could also be an issue.
CM
That and even worse, the inverse may happen!! This is not imaginary, after all, an electrostatic speaker and capacitor microphone are nothing else.
Far worse is the piezo electric properties of ceramic capacitors. This is a very real thing. You can safely forget it with plastic film and NP0, but anything else has it. It starts at X7R, and is a pain with Z5U. And the smaller sizes like 0805 and certainly 0603 and 0402 have a real big problem. I do speak from experience here. Try troubleshooting a circuit that fails if someone bangs the table...this could be traced to a badly chosen ceramic capacitor at the decoupling pin of the reference voltage of the A/D. Swapped for a tantalum and solved. Banging the table injected charge into the reference and made the samplevalues jump all over the place.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.