Coupling caps blues
More than 25 years ago, I had an interesting conversation with two audio designers from two different "audiophile quality" companies. One of the topics was coupling caps. The guy from "brand A" told their philosophy was to use a coupling cap on the preamp output an none on the poweramp input. And the guy from "brand B" said they had the opposite approach. So a "brand A" preamp would work fine with a "brand B" poweramp, but not vice versa. How silly!
Although I have not seen so many detailed schematics on "audiophile quality" consumer audio, I think that use, misuse and no use of coupling capacitors is a major concern. For instance, I have seen MC inputs without caps that draws large bias current through the pickup. Is that a good idea? Permanent magnetization is not a good thing.
So I ask you guys: If I design, say a line level output, MUST it always be DC-free? And if I design a line level input, can I assume that the source is always DC-free?
And if "mid-fi" equipment is involved, where leaky electrolytic caps often are used, what should I expect of DC?
Next post: Is DC around zero ideal for coupling caps?
More than 25 years ago, I had an interesting conversation with two audio designers from two different "audiophile quality" companies. One of the topics was coupling caps. The guy from "brand A" told their philosophy was to use a coupling cap on the preamp output an none on the poweramp input. And the guy from "brand B" said they had the opposite approach. So a "brand A" preamp would work fine with a "brand B" poweramp, but not vice versa. How silly!
Although I have not seen so many detailed schematics on "audiophile quality" consumer audio, I think that use, misuse and no use of coupling capacitors is a major concern. For instance, I have seen MC inputs without caps that draws large bias current through the pickup. Is that a good idea? Permanent magnetization is not a good thing.
So I ask you guys: If I design, say a line level output, MUST it always be DC-free? And if I design a line level input, can I assume that the source is always DC-free?
And if "mid-fi" equipment is involved, where leaky electrolytic caps often are used, what should I expect of DC?
Next post: Is DC around zero ideal for coupling caps?
Last edited:
More coupling cap blues
All amplifiers are bandpass filters. Some audio designers like to extend LF down to DC. So their amps are lowpass filters. But what is the advantage of a flat frequency response down to DC? What is the audio equivalent of DC? It is the static air pressure! Is that of importance in audio reproduction? However, frequency response down to DC eliminates the coupling caps and introduces DC problems in stead. Nice....
So most amps use coupling caps. This may be on inputs, interstage and/or outputs. FET-based input stages draw neglible bias current from it's surroundings, so DC-coupling is tempting. But what if the source have a DC component? Audio designers may be tempted to use servos to control DC on outputs. Nice, but what if the input carries DC. And what happend when you switch between inputs....
Now, assuming that we use coupling caps, what is the ideal cap? And what are the ideal working condition for that particular cap? Since coupling in audio usually implies low impedance, coupling caps may reach both tens and thousends of microfarads. Economy tells us to use electrolytics for these sizes, mayby in parallel with some good stuff. But electrolytics are polarized. These caps may be more linear with an applied DC. Non-polarized electrolytic caps are just caps you can not linearize with Dc
Does this make a preamp with single supply superior to a symmetrical design?
And what about leakage currents in electrolytic caps? Where is this taking us DC-wise?
All amplifiers are bandpass filters. Some audio designers like to extend LF down to DC. So their amps are lowpass filters. But what is the advantage of a flat frequency response down to DC? What is the audio equivalent of DC? It is the static air pressure! Is that of importance in audio reproduction? However, frequency response down to DC eliminates the coupling caps and introduces DC problems in stead. Nice....
So most amps use coupling caps. This may be on inputs, interstage and/or outputs. FET-based input stages draw neglible bias current from it's surroundings, so DC-coupling is tempting. But what if the source have a DC component? Audio designers may be tempted to use servos to control DC on outputs. Nice, but what if the input carries DC. And what happend when you switch between inputs....
Now, assuming that we use coupling caps, what is the ideal cap? And what are the ideal working condition for that particular cap? Since coupling in audio usually implies low impedance, coupling caps may reach both tens and thousends of microfarads. Economy tells us to use electrolytics for these sizes, mayby in parallel with some good stuff. But electrolytics are polarized. These caps may be more linear with an applied DC. Non-polarized electrolytic caps are just caps you can not linearize with Dc
Does this make a preamp with single supply superior to a symmetrical design?And what about leakage currents in electrolytic caps? Where is this taking us DC-wise?
Last edited:
a 10uf cap used in a circuit
where 1uf will do the job
(LP wise)
will exhibit 1/10 of the suggested
distortion
............
a 10uf cap used in a circuit
where 1uf will do the job
(LP wise)
will exhibit 1/10 of the suggested
distortion
............
What about the added inductance, ESR, microphony, etc, are these too insignificant to worry about?
Personally if a 1uf will do the job then that's what I use. As long as you meet your requirements why increase size and cost? What am I missing?
Last edited:
OT Hi Ed I did a benchmark with my new laptop 158mS to average 1000 4k FFT's and 4.48S for 1000 64k FFt's. So why does an AP take an hour to do these measurements?
Tomorrow, I will be on the way to Las Vegas, for CES and 'The Show', so I will stop commenting soon, but I hope that some, looking in, have learned a few things.
For example: Usually, the best cap, is no cap. This is because ALL caps are flawed in some way. Some distort, others are expensive, still others that don't distort, are VERY large and make compact construction difficult.
While there is relatively little wrong with NPO or COG caps, they are not perfect, and they may show their spots with special testing such as for dielectric absorption. They are all relatively small in value and get VERY EXPENSIVE if you try for the larger, precise values. They are not so small either.
Most other ceramics are inferior to some degree to COG, and are the typical ceramic caps that are available, and used in audio equipment, especially as output coupling caps. They can be almost OK to really lousy, so watch out and note what you use. They are NOT miracle devices, just barely adequate and cheap.
Polystyrene and polypropylene caps can be very good, but NOT all brands are great. Some are poorly made and have high Q mechanical resonances. Avoid them, if you can.
There are other types such as PPS, Mylar, Teflon, and Polycarbonate. Each type might do the job in a specific location, but except for Teflon, they are inferior to polypropylene and polystyrene.
Remember, it is not just harmonic distortion, that is significant. It MIGHT be pretty low, but dielectric absorption will still be there in Mylar, for example, as well as inductance, mechanical resonance, and vibration sensitivity. This might well change the sound of the audio equipment you are making.
Good luck, you are going to need it! ;-)
For example: Usually, the best cap, is no cap. This is because ALL caps are flawed in some way. Some distort, others are expensive, still others that don't distort, are VERY large and make compact construction difficult.
While there is relatively little wrong with NPO or COG caps, they are not perfect, and they may show their spots with special testing such as for dielectric absorption. They are all relatively small in value and get VERY EXPENSIVE if you try for the larger, precise values. They are not so small either.
Most other ceramics are inferior to some degree to COG, and are the typical ceramic caps that are available, and used in audio equipment, especially as output coupling caps. They can be almost OK to really lousy, so watch out and note what you use. They are NOT miracle devices, just barely adequate and cheap.
Polystyrene and polypropylene caps can be very good, but NOT all brands are great. Some are poorly made and have high Q mechanical resonances. Avoid them, if you can.
There are other types such as PPS, Mylar, Teflon, and Polycarbonate. Each type might do the job in a specific location, but except for Teflon, they are inferior to polypropylene and polystyrene.
Remember, it is not just harmonic distortion, that is significant. It MIGHT be pretty low, but dielectric absorption will still be there in Mylar, for example, as well as inductance, mechanical resonance, and vibration sensitivity. This might well change the sound of the audio equipment you are making.
Good luck, you are going to need it! ;-)
I don't think ESL and ESR in a coupling cap are an issue. They may alter the HP curve but that would only be significant in pathological cases, if ever.
jan didden
Actually Jan, some caps are very inductive. Can you imagine why and how they are built? Keeping open to all the permutations of different caps can give a more optimum placement of the cap type. For example, a relatively highly inductive cap, MIGHT make a lousy power supply decoupling cap. Conversely, a very LOW ESR cap with long leads might make a regulated power supply oscillate. I have found some polypropylene caps that had several inches of INTERNAL inductance do to internal termination. These caps were made by a 'reputable' manufacturer, but a REL of the same value did NOT have this added inductance. Personally, I don't there is a coupling cap made on this Earth that is truly without compromise, sonically.
Last edited:
All true, and well known, but not my point in the post you appear to reference.
I wrote about cap ESR in power supplies in 1994 for the first time.
jan didden
Can you build two identical preamplifiers, one with and one without caps in the signal circuit and demonstrate their differences in performance by electrical measurements such as THD, FR, IM, to a degree sufficient to be audible? If you can't, are we back to what people like? Are we down to mysterious gremlins that can't be traced even by the most sensitive electronic measurement equipment that is the hallmark of audiophile neurosis and the niche industry that caters to it? Can they be distinguished by AB blind testing? Where's the proof that any of this matters and isn't just a bunch of the usual audiophile hooey?
And what exactly is your mission here? Consumer protection, warning us about the High End industry?
I find it rather presumptuous to ask for proof. If you don't believe it, let it be.
Buy a $50 CD Player, build a LM3885 Amp and thats all you need.
It will be hard to distinguish this equipment in a DBT from a $10'000.- one.
But then, I'm asking, why posting in a DIY Forum?
Tino
Well.
It could be because it is way more expensive, the instrument is 5 years old and the design even older. And I may overstate the wait a bit, but it seems like an hour, it is at least 20 minutes.
The big advantage is that it is a well respected instrument and in calibration.
Also I sometimes use it at it's maximum which I recall as 4096 x 32K.
It does have a higher speed with a lower resolution A/D.
So when doing a setup I use smaller samples that only take a few seconds, then do my data runs while doing something else, like inventory!
But of course what you really wanted to say was "I got the best toy on the block!"
ES
Last edited:
Nope, just my hobby horse. A decent soundcard and free software can do just about any audio measurement you would ever need.
Nope, just my hobby horse. A decent soundcard and free software can do just about any audio measurement you would ever need.
Agreed! And the biggest advantage is... if you blow up the input by mistake it doesn't set you back real money.
- Status
- Not open for further replies.