Hello everyone,
At the input of an amplifier, we put a capacitor of a few microFarads to block the possible continuous component coming from the preamp.
What type of capacitor gives the best results?
What is your experience in the field?
Thank you for your feedback.
At the input of an amplifier, we put a capacitor of a few microFarads to block the possible continuous component coming from the preamp.
What type of capacitor gives the best results?
- a chemical type Silmic II, BlackGate, Nichicon Muse,
- a MKP (Polypropylene)
- a MKS (Polyester PET)
- a MKT (Polyester)
What is your experience in the field?
Thank you for your feedback.
MKT is a good choice.
The value is dependent of the input impedance. Often a resistor to the ground.
The value is dependent of the input impedance. Often a resistor to the ground.
I would personally use a polypropylene if there is adequate space, but any decent non-polar or BP electrolytic could be a close second choice. Sometimes polarized types are fine, but you have to anticipate what reverse voltage conditions you expect it to be subjected to. If you feel spendy, Audio Note Kaisei is the new Black Gate.
I would make an effort to avoid a coupling cap, the easiest is perhaps a DC detection circuit that shorts the input.
If there is a fixed DC voltage, per example due to biasing, then a BG N cap would be my choice. Good films are physically huge and quite inappropriate at this position and bad films really suck.
If there is a fixed DC voltage, per example due to biasing, then a BG N cap would be my choice. Good films are physically huge and quite inappropriate at this position and bad films really suck.
The late great Cyril Bateman in his measurement of capacitor distortion found that non-polar electrolytics, or polar electrolytics back to back had harmonic distortion down to the noise level.
If you decide to use a single polar electrolytic, it needs to be a higher value than you might imagine. Douglas Self in his book Small Signal Audio Design derives a rule of thumb that to prevent low frequency distortion at quite surprising level (0.05% to 0.1% typical), you need to ensure that there is less than 80mV AC across the capacitor at the lowest audio frequency (say 10Hz).
Polyester capacitors - again from Small - have significant mid band distortion, with the level dependent even on capacitors from the same production batch.
So for low harmonic distortion, polypropylene if you have room. Next smallest is back to back polar electrolytics. Smallest is a single non-polar electrolytic. capacitor.
For some light bedtime reading Bateman's landmark capacitor articles are here https://linearaudio.net/cyril-batemans-capacitor-sound-articles
Craig
If you decide to use a single polar electrolytic, it needs to be a higher value than you might imagine. Douglas Self in his book Small Signal Audio Design derives a rule of thumb that to prevent low frequency distortion at quite surprising level (0.05% to 0.1% typical), you need to ensure that there is less than 80mV AC across the capacitor at the lowest audio frequency (say 10Hz).
Polyester capacitors - again from Small - have significant mid band distortion, with the level dependent even on capacitors from the same production batch.
So for low harmonic distortion, polypropylene if you have room. Next smallest is back to back polar electrolytics. Smallest is a single non-polar electrolytic. capacitor.
For some light bedtime reading Bateman's landmark capacitor articles are here https://linearaudio.net/cyril-batemans-capacitor-sound-articles
Craig
Thank you for the Bateman's articles, very interesting
I have a lot of place.
I will use polypropylene capacitor.
Here is a part of the datasheet of PHB polypropylene capacitor from NICHICON.
"Main applications
Switching capacitor for industrial and motor speed controls, high frequency electronic ballasts, switching mode power supplies, resonant circuits, induction heaters, high-end audio applications"
But it is a big capacitor. I will use a 10µF value. 18 x 33 x 32 mm size.
I have a lot of place.
I will use polypropylene capacitor.
Here is a part of the datasheet of PHB polypropylene capacitor from NICHICON.
"Main applications
Switching capacitor for industrial and motor speed controls, high frequency electronic ballasts, switching mode power supplies, resonant circuits, induction heaters, high-end audio applications"
But it is a big capacitor. I will use a 10µF value. 18 x 33 x 32 mm size.
Large body film caps can be problematic for audio. They can sound very clear and transparent, yet they may tend to lose low level sound stage information. Often a good electrolytic, say, maybe 50uf, with a small .01uf MKP bypass can work well, although the electrolytic may need some time to settle into its best sound.
So you would not recommend, say, Jupiter or Miflex for input coupling caps ... since they are physically large? Yet they are widely used in boutique (ie. expensive) amps.
What do you think of Rubycon 25MU225MB23225 for coupling duties? They're tiny! 😱
The 25MU225MB23225 is a surface mount part. It has an acrylic dielectric too - acrylic like the lid of a record deck. Very very sensitive to heat when soldering. And I suspect that it has high dielectric loss, although the datasheet is silent on that.
Last edited:
See my above post regarding electrolytic distortion at low frequency. You cannot recommend a part value without knowing the signal level, the impedance of the circuit, and the lowest frequency in the audio source.
Craig
Film caps in the same pitch as the PCB anytime. You can't go wrong with Wima MKS2 but give them a day to break in. Black Gate caps are long gone unfortunately.
Please no extension wires (creating antennas) to D cell sized caps with beautiful chinese audio names in golden lettering and precious metals for the film/foil. The PCB is just copper, remember?. That you put the device gently on a table and still hear caps come loose from the board 🙂
If you want BG N series 4.7 µF 50V ... I have 2 new/unused ones left.
Please no extension wires (creating antennas) to D cell sized caps with beautiful chinese audio names in golden lettering and precious metals for the film/foil. The PCB is just copper, remember?. That you put the device gently on a table and still hear caps come loose from the board 🙂
If you want BG N series 4.7 µF 50V ... I have 2 new/unused ones left.
Last edited:
My personal opinions only:
You can try whatever you want. I would only suggest to listen for best sound stage and not compromise on that (such as for some seemingly improved clarity/transparency; that can be a trap). If there are audible problems then you need to find a way to fix them without hurting sound stage. Again, electrolytics can take some time for the sound to settle in.
MU caps, IME, do not sound good for analog audio coupling. They are good for some decoupling, and or sometimes for voltage regulator input and output caps. The good thing about them is they are more linear than ceramic or tantalum while still being lossy enough not to ring like NPO/C0G can.
MKS coupling caps can be good for obscuring what's wrong with poor recordings, poor dacs, etc., but they make the best recordings sound worse. It only takes one MKS coupling cap somewhere in a system to make for a more "forgiving" sound. But the trick works by adding a little distortion. Its not something I would do unless maybe there was a switch to allow use of a better cap for better recordings.
EDIT: Again, I would suggest to use your ears and decide for yourself. I have done the experiments and you probably should too. If you don't have a good sound stage now to use for reference then you might need to work on that first.
Last edited:
What a BS. The least trouble giving and normally priced and available fine sounding film cap (practically never needing replacement) is not advised. “Obscuring what’s wrong with poor recordings and poor DACs”, pfff…
Also blaming all MKS caps is like blaming all americans for all that goes wrong in the world.
Also blaming all MKS caps is like blaming all americans for all that goes wrong in the world.
Last edited:
MKS is polyester. Doug Self warns about polyester cap distortion. I do too. Also, please see post #6 above.
Last edited:
Yeah it is that simple. “Names” & papers & horror stories. Nothing used by industry or normal people is OK. All polyester caps are boo boo. Even (insert famous name here) that sells a book says so and that is important. Capacitor factories and their lowly educated designers don’t know jack about what dielectric is OK. They just do something. They should be glad that unaware simple souls buy their mediocre stuff.
Thanks everyone for your comments
I have MUSE capacitors and I ordered polypropylenes.
I will compare these two types of capacitors.
I have MUSE capacitors and I ordered polypropylenes.
I will compare these two types of capacitors.
I agree completely with you. Wima polyester caps are VERY good, better than any average bipolar capacitor. But they need burn in time , about one day where they transform completely from sounding very colored in the midband and bass the first hour, to a very transparent sound with fine details. Many Diy:er or constructors seems to listen directly and loose their way making wrong conclusions.
Douglas Self has done many good things, but recommending 5532 is not one of them, and not a believer of burn in makes him less believable in my opinion.