I would like to see the actual circuit first. 🙂
A circuit diagram is important to determine where this capacitor is hypothetically going to be placed, and if it’s part of the filter, a RTF time constant network, or if it’s even needed at all to prevent the passage of DC, the position can have some importance regarding the value, type and capacitance.
A circuit diagram is important to determine where this capacitor is hypothetically going to be placed, and if it’s part of the filter, a RTF time constant network, or if it’s even needed at all to prevent the passage of DC, the position can have some importance regarding the value, type and capacitance.
Last edited:
If you talking about dc blocking, I would use a BIPOLAR electrolytic capacitor, they are suppose to have less distortion than any polar electrolytic capacitor.
Nichicon mentions ES series as their “Bi-polarized “nichicon MUSE” acoustic series”.
But other brands have their bipolar caps, Panasonic, Elna and others.
Film caps are of course good but the largest capacitance for a miniature cap ( with lead spacing 5mm) is 10uf, someting like wima mks2 10uf(I heard about wima mks2 22uf but I never seem them).
Nichicon mentions ES series as their “Bi-polarized “nichicon MUSE” acoustic series”.
But other brands have their bipolar caps, Panasonic, Elna and others.
Film caps are of course good but the largest capacitance for a miniature cap ( with lead spacing 5mm) is 10uf, someting like wima mks2 10uf(I heard about wima mks2 22uf but I never seem them).
If the nominal voltage across the capacitor is zero, bipolar are a reasonable choice, but if there's an explicit voltage across the cap use unipolar electrolytics with the correct polarity!
If the cap value is large enough the distortion will be minimal as only a tiny fraction of the signal voltage appears across the cap.
Nichicon UES are by far the best non polar electrolytic caps for the signal path. They sound very close to most smaller MKP film caps. You'll be hard pressed to hear a difference in any application that calls for a compact sized non polar cap. The next best ones are ERO brand, but they're harder to find.
| What I have found to be the best audio path capacitors: | |||
| ≤0.01uF---Wima FKP 2 or 3 film/foil polypropylene (5.5x7.2mm max) | no attenuation of the audio signal and therefore provide absolutely precisely accurate pulse reproduction. | ||
| 0.01 to 0.1uF---Wima MKP 10 double-sided metalized polypropylene (6x13mm max) | 4 film layers gives 5-10 times greater pulse rise time, much better pulse behavior, and 30-50% lower dissipation factor than versions metallized on one side only. | ||
| 0.1 to 1.0uF---Wima MKP2 single-sided polypropylene (8.5x7.2mm max) | smallest possible box size at a given capacity. | ||
| ≥1 uF ---Nichicon UES bipolar (5 to 16 mm dia.) | MUCH less distortion than polarized electrolytics |
And, in light of that information, I have revised my list:
|
I've tried the Nichicon UES in various coupling and feedback stages. They are almost as good as an average MKP in terms of distortion given the voltage drop doesn't exceed a few hundred mV and the DC offset is close to zero. Once you have appreciable DC offset present, a polarized cap is a better choice. This allows the cap to work in a more linear range. The main concern is to use a cap which has non-magnetic leads. All UES series caps have copper leads as opposed to most other electrolytics which have steel leads. The parasitic inductance can be a problem when using magnetic lead caps in feedback loops with any significant bandwidth.
Inductance due to ferromagnetic materials in leads is non-linear, and can be an issue at very low impedances / high currents, such as components in speaker leads, unlikely to matter with low power signals though as we are talking µ-henries v. k-ohms.
Inductance of any form can be an issue if it introduces a pole or zero low enough in frequency to affect circuit operation. Large wound capacitors (& wirewound resistors) can have significant inductance just from the geometry, though its possible to wind in a way that cancels circulating currents and is low inductance. Sometimes the datasheet will have ESL given, or self-resonant frequency.
Inductance of any form can be an issue if it introduces a pole or zero low enough in frequency to affect circuit operation. Large wound capacitors (& wirewound resistors) can have significant inductance just from the geometry, though its possible to wind in a way that cancels circulating currents and is low inductance. Sometimes the datasheet will have ESL given, or self-resonant frequency.