You're never going to find a C0G big enough to use as a coupling cap. If you do, it's either a) not a C0G in which case you get what you deserve, or b) it's a UHF/microwave amplifier.
john curl said:
That's like saying that most MOSFETs are optimized so much for switching that they are no good for audio, therefore they are no good. You need to be aware of the properties of each capacitor and choose accordingly. Specifically:
1) Even the poor quality ceramics have their uses, e.g. supply decoupling.
2) High quality ceramics don't have the same problems, so you can use them in more critical locations. I gave the example of a Miller compensation capacitor precisely because it only needs a small value which will be easy to find and won't cost very much; it's not like it's going to break the bank to get a 47pF capacitor, especially compared to the popular silver mica.
For supply decoupling, low inductance and high dielectric absorption are desirable properties. Yes, you want it lossy at high frequency. Otherwise, inductances tend to ring.
For Miller comp, the jury is still out on C0G's as far as I'm concerned. For most of what I build, they can't handle the voltage. 500 volt micas are more common than 500 volt class 1 ceramics. 50 volters will get nonlinear when hit with 200+ volt swings. Micas also have high ESR. Miller comps often benefit from a few ohms in series - in can produce a more desirable open loop roll off.
For Miller comp, the jury is still out on C0G's as far as I'm concerned. For most of what I build, they can't handle the voltage. 500 volt micas are more common than 500 volt class 1 ceramics. 50 volters will get nonlinear when hit with 200+ volt swings. Micas also have high ESR. Miller comps often benefit from a few ohms in series - in can produce a more desirable open loop roll off.
john curl said:
What is this based upon? The 30-year-old data that didn't distinguish, if I recall, Class 1 C0Gs from run-of-the-mill ceramic discs? No one is even suggesting "MOST" ceramics are good choices. And neither did Bateman, who clearly showed their distortion.
You offered to us no nothings in another thread that you measured or listened to what you thought *might* have been a C0G, but you couldn't be sure since you couldn't recognize the physical difference etween them. Well us no nothings should buy that one without question, sir.
You condemn Bateman without any refuting evidence, just because he MAY have a vested interest in the outcome. Jeez, Curl. Show some professionalism. Maybe it is you that has a vested interest against someone showing some objective results in the last 30 years. And you expect us to take your word without question???
Your approach to communication is nonsense. You have not shown any evidence for negating Bateman's tests. You argue for measurements when they are on your side. Then you argue against measurements when you decide, or when someone DOES some measurements. You have not even said you compared C0Gs against lesser ceramics, much less tested them.
If there is something wrong with Bateman's tests, then for God's sake be professional about it instead of using your usual petty, nonsensical attacks.
wg_ski said:
Kemet has lots of 200V C0G types available. In higher voltage ranges there are the Vishay/Sprague 561R series which are 1kV, but physically rather large. Both these types are available at Mouser. There are also 500V Xicon C0G available, but I'm not sure if I trust them.
I would think the 50 Volt caps would get a bit nonlinear for 200 Volt swings too 🙂.
OK, use ceramics at your own risk. You can fine read Bateman and learn something important too. Just make sure that you read the article with an OBJECTIVE slant, rather than the way Bateman sometimes puts forward. Still, where he measures distortion, distortion exists.
I had plenty of 10 nF 2 KV, now I have a half of Ziploc... Time to start using something different.
that is a very nice argument
thanks everybody for the input
can somebody place a conclusion of all this ????
thanks
the OP
thanks everybody for the input
can somebody place a conclusion of all this ????
thanks
the OP
Re: that is a very nice argument
There's very little of audio science to be found, as in audible science, not measured or simulated. Almost everything is anecdotal.
I think you have to define what you're going to use the caps for. Caps in the audio pathway, handling audio frequencies, like coupling caps and the caps sinking/sourcing the speaker return currents will for sure affect the sound.sakis said:
There's very little of audio science to be found, as in audible science, not measured or simulated. Almost everything is anecdotal.
Re: that is a very nice argument
Conclusions? What conclusions? There is only one conclusion:
The uncertainty 😀 😀
sakis said:
Conclusions? What conclusions? There is only one conclusion:
The uncertainty 😀 😀
DIY cap test
Amazing! Don't need no Bateman oscillator. (See picture) Take the left and right ouputs of a good sound card and program a frequency sweep with the outputs 90 degrees apart and the same amplitude and find the null point. Then sweep the amplitude of one channel a small amount to get the deepest null.
I used Audition/Cooledit. You get the exact R/C time constant and the residual third (the thirds from the soundcard are only 12dB nulled so you can expect -115dB-ish resolution). I used a Vishay .005% 10k resistor. Also a great way to sort RIAA values without a fancy meter.
Amazing! Don't need no Bateman oscillator. (See picture) Take the left and right ouputs of a good sound card and program a frequency sweep with the outputs 90 degrees apart and the same amplitude and find the null point. Then sweep the amplitude of one channel a small amount to get the deepest null.
I used Audition/Cooledit. You get the exact R/C time constant and the residual third (the thirds from the soundcard are only 12dB nulled so you can expect -115dB-ish resolution). I used a Vishay .005% 10k resistor. Also a great way to sort RIAA values without a fancy meter.
Attachments
syn08 said:
Now don't pick on Johnny that way. You'll give him a complex.
Look at it as Teacher's way of warming up our subhuman brains in order to be blessed with a useful tidbit of Teacher's knowledge in another couple dozen posts or so.
sakis,
Chuck your 10.000 ceramic capacitors into the nearest wastepaper basket. (Or sell them to Mr Evil).
Generalizations are worthless so no, there is no conclusion unless talking about a specific capacitor part number from a specific manufacturer.
I have to comment on the "tap" test though. Certainly hi k ceramic caps are microphonic as they use similar ceramics to piezoelectric actuators. C0G caps don't have this problem to any worrying extent, but cables do. Unless you get special non-microphonic coax (expensive and hard to find) you'll find most cables are microphonic- just whack a microphone cable and listen. Maybe we should stop using cables?
My compliments on the statement a few posts back about bypass caps and losses. Most people don't appreciate this. D=1/Q, so very low D caps have high Q and will tend to ring on fast edges if the circuit has any inductance (all of mine do, and I'd bet all of yours do as well). You need a controlled dissipation factor (D) to prevent this, so hi k is way better than low k for bypassing duty. That's also why tantalums are good for bypassing- they have a moderate and consistent D, though I'd never use them where they could affect an audio signal. Always think before putting a high Q (low D) capacitor in a very low impedance circuit.
I have to comment on the "tap" test though. Certainly hi k ceramic caps are microphonic as they use similar ceramics to piezoelectric actuators. C0G caps don't have this problem to any worrying extent, but cables do. Unless you get special non-microphonic coax (expensive and hard to find) you'll find most cables are microphonic- just whack a microphone cable and listen. Maybe we should stop using cables?
My compliments on the statement a few posts back about bypass caps and losses. Most people don't appreciate this. D=1/Q, so very low D caps have high Q and will tend to ring on fast edges if the circuit has any inductance (all of mine do, and I'd bet all of yours do as well). You need a controlled dissipation factor (D) to prevent this, so hi k is way better than low k for bypassing duty. That's also why tantalums are good for bypassing- they have a moderate and consistent D, though I'd never use them where they could affect an audio signal. Always think before putting a high Q (low D) capacitor in a very low impedance circuit.
- Status
- Not open for further replies.