Electronics > Beginners
Ceramic caps in audio revisted
(1/2) > >>
John B:
I'm aware of some reasons why you shouldn't be using ceramic caps in an audio path: capacitance change with DC bias, or piezoelectric/microphonic effects.

Normally I would stick to film capacitors, but for the output stage of some preamps, I would like at least a 10uF cap in order to drive 10k inputs at low frequency. I have 2uF film caps in the parts bin, but they're physically large. So I experimented with some cheap ML caps, similar to these:

https://www.aliexpress.com/item/100pcs-lot-Multilayer-ceramic-capacitor-10uF-106-50V-106k/32455168081.html

You could easily fit >60 uF in the same footprint as the film caps, so they are worth a second look. In the intended application, I never use half rail biased stages, so there will never be a constant DC bias across the caps (except for inherent DC offsets in the op amps).

I also checked them for monophonic characteristics, but they didn't exhibit anything more than film caps do. This involved a good hit with a plastic rod to register a signal on the scope, but it was the same with film caps, and totally excessive compared with anything they would be subjected to during use.

They do exhibit audible whines when subjected to rapid changes in DC bias, such as charge pump circuits, but these are basically 0-12V square waves applied to the caps several kHz. In the intended applications, the caps would only be subjected to 10's of mV max across them.

Obviously Ill just have to build something up to do further testing, as there's too much external noise on a breadboard. But beyond that, why else aren't they recommended? Has conventional wisdom not kept pace with the advent of large capacitance ceramic caps?
edavid:
The voltage dependence can lead to higher distortion than using electrolytics.  You should definitely measure that.
wraper:
Those are the worst trash, Z5U or Y5V. Not even X7R. Use nonpolar electrolytic instead. If you want better performance at high frequencies, attach film or C0G ceramic cap in parallel.
floobydust:
High capacitance MLCC are very non-linear. The only low distortion ceramic capacitor dielectric is COG/NPO. So only small values available.

If you want to experiment with microphonics, use the gravity wave detector circuit:
https://www.eevblog.com/forum/projects/gravity-wave-detector-radio-electronics-april-1986/msg1103545/#msg1103545
John B:
OK they may be not great for their main purpose, but I can't really measure the issue for AC coupling of audio. I tried measuring the harmonics with an audio interface spectrum analyser, first without caps in series, then with. I couldn't measure any meaningful difference in the first 5 harmonics. This was about 40uF series capacitance into a 10k load. I couldn't measure a voltage drop or phase shift on the scope (within reason, obviously if you take it down to <50Hz you'll see it), and I couldn't measure any harmonic distortion attributable to the caps. I appreciate the responses, but to reiterate, there's no DC bias across the caps, and in circuit the voltage across the caps is in the mV range.
Navigation
Message Index
Next page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod