I think this is an interesting topic to explore, namely, what is the useful cap value range for most audio circuitry? What is the most practical, optimum, and cost effective?
SY has stated that HIS practical range is .01-.033uF. What if we were all limited to these values? What would we gain, what would we lose?
Well, of course, this is what got me into trouble with ceramic caps, 40 years ago. Please remember that I was replacing something like a .047uF cap with a 2.2 uf 'monolithic' CERAMIC CAP and got disappointing results, in my Fisher KM60 FM tuner.
Why was I doing this, was I crazy? No, just naive. (more later)
SY has stated that HIS practical range is .01-.033uF. What if we were all limited to these values? What would we gain, what would we lose?
Well, of course, this is what got me into trouble with ceramic caps, 40 years ago. Please remember that I was replacing something like a .047uF cap with a 2.2 uf 'monolithic' CERAMIC CAP and got disappointing results, in my Fisher KM60 FM tuner.
Why was I doing this, was I crazy? No, just naive. (more later)
Last edited:
Pooge, I generalize, when I advise to avoid ceramics, unless absolutely necessary, BUT I use them in my designs in less critical areas than the audio path. I tend to avoid COG, because their inherent temp stability is not too important in audio equipment I use caps for, their larger size, and ultimately, their cost. Most every DIY'er doesn't know the difference between COG and another type of ceramic cap. Even I have trouble, at first glance. Therefore, I warn against using something that LOOKS like a monolithic ceramic for general audio coupling.
I do not think so. In case i use a coupling cap (very rarely), it is 1uF and more, and loaded by 100k or more.
The first thing we should address is low frequency response. In the early days, for example the 1950's, OUTPUT coupling caps were small, and input impedances high. For example 500,000 ohms might be the input vol pot value of most tube equipment of the day, AND they did not have sub-wolfers generally available!
Now, using SY's maximum cap value, .033uF and 500,000 ohms, what do we get? I calculate about 10 Hz, should be good enough for the old days, except, of course if you have even more caps rolling off the bass.
NOW, what would happen if you plugged this preamp directly into a Parasound jfet input, power amp, just pick any one we normally make, or have made over the last 20 years. Well, the input impedance is USUALLY 50K ohms or 10 times less! What happens now? All of a sudden, your prize tube preamp is 3dB down at 100Hz! Is this OK, or is SY perhaps off by a factor of 10? (more later)
Now, using SY's maximum cap value, .033uF and 500,000 ohms, what do we get? I calculate about 10 Hz, should be good enough for the old days, except, of course if you have even more caps rolling off the bass.
NOW, what would happen if you plugged this preamp directly into a Parasound jfet input, power amp, just pick any one we normally make, or have made over the last 20 years. Well, the input impedance is USUALLY 50K ohms or 10 times less! What happens now? All of a sudden, your prize tube preamp is 3dB down at 100Hz! Is this OK, or is SY perhaps off by a factor of 10? (more later)
Last edited:
BIGOT!
But I don't recall anyone advocating them for coupling duties. They are too small in value, and I think even SY could figure this out.
No, you are. There's no reason to use a 500k gate or grid resistor. 1M works for nearly any small signal tube; in my Heretical, I use 10M in the RC coupling (with a 0.01uF cap), which is fine for just about any FET, or for a small signal tube operating with a servo or CCS. Some engineering needs to be done in low noise stages, but nothing impossible.
Monolithic ceramics are the HIGH VALUE ONES, you know .15 - 2.5uF. These are the worst example of capacitors I have ever measured. Most here will ONLY use caps for power supply decoupling or interstage or output coupling. EQ is not necessary for digital, is it? Now, EQ could be done EITHER with film or COG ceramic caps. SY seems to think that film caps are inferior in some way. Prove it, SY. Compare a REL RT polystyrene against your reference standard. I dare you! Guess what Charles Hansen and I use?
Sy,
for a Moderator that is poor posting.
JC did not accuse you of being wrong by a factor of ten.
He showed why 500k would give F-3dB@~10Hz.
You then confirm using 1M to 10M would be acceptable, (which would give F-3dB ~1Hz to 0.1Hz).
A simple explanation of your choices and why you made them instead of
for a Moderator that is poor posting.
JC did not accuse you of being wrong by a factor of ten.
He showed why 500k would give F-3dB@~10Hz.
You then confirm using 1M to 10M would be acceptable, (which would give F-3dB ~1Hz to 0.1Hz).
A simple explanation of your choices and why you made them instead of
would have been a better example to set for us Members.
Well, yes, that's exactly what he did. And that's OK, it gave me a chance to point out why that's an unnecessary restriction.
Why would I want to prove something I never said and don't generally believe?
You said they were unreliable and microphonic. Some may be, but why not stick to an audio designer approved brand like REL?
You included REL into your 'Audiophile cap' description, because REL is used by most major hi end manufacturers of hi end equipment. What do you think Parasound uses in my designs? Ayre? Audio Research (much of the time) Vendetta, Constellation Audio, etc., etc.? Still, where and what value is your line stage volume control. Do you buffer it with a jfet follower? If not, why did you sneak that 2.2uf OUTPUT COUPLING CAP into your phono design? Let's find an APPROPRIATE ceramic cap for that position, shall we?
COG, no way!
COG, no way!
Nope, I didn't even mention them.
If you actually read my articles, you will see that the volume controls are anywhere from 25k to 100k, with no input coupling cap. Why would I need one? Competent sources don't have significant offset. And input transformers don't pass DC.
The output cap from my linestage is... a piece of wire. Direct coupled. That buffer design can be incorporated into my phono stage. To save you the grief of actually having to read my prose on the phono stage in toto, I'll quote from that article:
Moving on, if possible. One of the REAL PROBLEMS of using smallish (.033uf or so) coupling caps on an input or interstage is NOISE. NOT from the cap, but from the resistor that loads it. (more later)
To paraphrase Samuel Johnson, I have given you an answer, I cannot give you an understanding.
Well, SY, IF I CAN'T UNDERSTAND, who will? Yes, I took advantage of your poor selection of phrases, because I KNEW that you used a larger value coupling cap than you stated (for the output of your phono stage), but you should be less condemning of my design practices, and more carefully define your own.
John, I never "condemned" your design practices. You need to read better and with a lesser sense of paranoia. I do condemn dismissing things out of hand that are valid engineering solutions avoided out of prejudice.
- Status
- Not open for further replies.