What's the better input coupling capacitor? 160v polypropylene vs 250v polyester
I'm upgrading the capacitors in my two tda7297 amps and my supplier only has the following two capacitors in stock that I think will be suitable as input coupling caps for my application:
The stock input caps on one amp are small no-name chinese PET caps rated at .22uf 63v. The other amp has larger no-name .22uf 250v (PP I think) caps.
I'm upgrading the input capacitors because I'm also upgrading the very poor power capacitors and might as well have all the caps be made from reputable manufacturers. The reason I would like to replace them with slightly high capacitance ratings is to lower the bass frequency roll-off and subsequently phase distortion(at 50kohm). After reading nearly every thread I could find on polypropylene vs polyester, the general consensus is that polypropylene is the superior dielectric material given that all else is equal. But what if it's not?
Any input would be appreciated. Thanks.
I'm upgrading the capacitors in my two tda7297 amps and my supplier only has the following two capacitors in stock that I think will be suitable as input coupling caps for my application:
- Kemet .27uf 160vdc metalized polypropylene capacitor (PP) 5% 105c
- Epcos .33uf 250vdc metalized polyester(PET) 5% 125c
The stock input caps on one amp are small no-name chinese PET caps rated at .22uf 63v. The other amp has larger no-name .22uf 250v (PP I think) caps.
I'm upgrading the input capacitors because I'm also upgrading the very poor power capacitors and might as well have all the caps be made from reputable manufacturers. The reason I would like to replace them with slightly high capacitance ratings is to lower the bass frequency roll-off and subsequently phase distortion(at 50kohm). After reading nearly every thread I could find on polypropylene vs polyester, the general consensus is that polypropylene is the superior dielectric material given that all else is equal. But what if it's not?
- Is a lower voltage rated polypropylene capacitor still better than a higher voltage rated polyester capacitor with a thicker dielectric?
- Also, how do Kemet and Epcos compare as manufacturers?
- Are there any other significantly better input caps I should look for in this price range and size? (from a different supplier)
Any input would be appreciated. Thanks.
Last edited:
22uF NP elco!
Input coupling caps (as opposed to filter caps) should be large enough to not see significant signal frequency voltage, and as long as you make them large enough there is nothing wrong with an elco in this position (as long as you can stand the leakage possibly slightly moving the bias point).
Film can be fine, as can NP0/C0G ceramic, but in both cases values tend to be small so you end up with significant signal voltage across the cap (In this sense it makes it a filter), make the cap huge and the signal voltage across the cap goes away taking the distortion with it.
If you really cannot face using an elco here, I would suggest investigating some of the newer C0G/NP0 ceramics that can be had up to a microfarad (AVX as I recall), brilliant parts.
Regards, Dan.
Input coupling caps (as opposed to filter caps) should be large enough to not see significant signal frequency voltage, and as long as you make them large enough there is nothing wrong with an elco in this position (as long as you can stand the leakage possibly slightly moving the bias point).
Film can be fine, as can NP0/C0G ceramic, but in both cases values tend to be small so you end up with significant signal voltage across the cap (In this sense it makes it a filter), make the cap huge and the signal voltage across the cap goes away taking the distortion with it.
If you really cannot face using an elco here, I would suggest investigating some of the newer C0G/NP0 ceramics that can be had up to a microfarad (AVX as I recall), brilliant parts.
Regards, Dan.
I was considering using 2(per channel) 1uf bipolar nichicon muse caps in reverse orientation like in the salis build guide, but opted to use small film caps thinking(erroneously?) the quality would be better.
Doesn't a larger capacitor effect high frequency response and/or other parameters negatively?
Also, how is the aforementioned signal frequency voltage across the cap related to the voltage rating of a capacitor, if at all?
Doesn't a larger capacitor effect high frequency response and/or other parameters negatively?
Also, how is the aforementioned signal frequency voltage across the cap related to the voltage rating of a capacitor, if at all?
First get the low frequency pole a decade of so down on the audio, the phase shift will make more difference then the distortion from even the shonkiest of cheap China special elco, then and only then worry about the details of the cap.
The DC rating must be safely greater then any bias voltage you are blocking (obviously), but if you make the cap large enough the signal voltage across it (even at full level) should be in the order of 10mV or so, which anything will cope with.
Have a look at the ESR vs frequency curves for modern elcos sometime, these are not your grandfathers caps, often they quote figures at 100KHz (Intended use is switching power supplies), at the few tens to few hundreds of uF level it is not an issue (I have measured it).
If 1uF will do you then you should seriously look at the high value C0G/NP0 parts, about as blameless as anything out there.
Regards, Dan.
The DC rating must be safely greater then any bias voltage you are blocking (obviously), but if you make the cap large enough the signal voltage across it (even at full level) should be in the order of 10mV or so, which anything will cope with.
Have a look at the ESR vs frequency curves for modern elcos sometime, these are not your grandfathers caps, often they quote figures at 100KHz (Intended use is switching power supplies), at the few tens to few hundreds of uF level it is not an issue (I have measured it).
If 1uF will do you then you should seriously look at the high value C0G/NP0 parts, about as blameless as anything out there.
Regards, Dan.
- Status
- Not open for further replies.