decoupling capacitors

Oct 21, 2007 at 9:36 AM Thread Starter Post #1 of 49

villekille

Head-Fier
Joined
Nov 22, 2006
Posts
85
Likes
10
Is there some reason for not using tantalum caps for decoupling purposes? For example, in PPA V2 there are electrolytics and some film caps for opamp decoupling but why not use tantalums, they are smaller and cheaper.
 
Oct 21, 2007 at 10:06 AM Post #2 of 49
because they are not as linear, as accurate as film caps despite their pretty decently low inductance.

It's an audiophile thingy when linearity is concerned.

I personally prefer Electrolytics --> Tants --> Film caps.

Heck one should even consider ceramics for power supply decoupling
biggrin.gif
Boils down to what kind of "sound" you want.
 
Oct 21, 2007 at 2:56 PM Post #3 of 49
IIRC, MUSE caps as well as BlackGates are Tantalums. But tantalum caps are only marginally better than normal electrolytics. Films are the real way to go, but sometimes a circuit calls for a value in microfarads which is too large and expensive to be practical for your particular piece of equipment. In any case, make sure the cap you use is non-polarized. (or bipolar, whatever you want to call it)
 
Oct 21, 2007 at 4:14 PM Post #4 of 49
Quote:

Originally Posted by Logistics /img/forum/go_quote.gif
IIRC, MUSE caps as well as BlackGates are Tantalums.


IIAC (if i am correct) you are wrong. both of the above are electrolytic types. Both do use "very unique" electrolytic "mixes" but they are electrolytic types none-the-less.
Quote:

In any case, make sure the cap you use is non-polarized. (or bipolar, whatever you want to call it)


why use a non-polar cap in a PSU application, or one where the cap will ALWAYS have a voltage differential facing the same way? i have always been at a loss to explain why people use bipolar caps on the outputs of tube & hybrid tube amps. i think at some point people equated a non-polar electrolytic with a film cap because both are non-polar, too bad one is an electrolytic and the other a film. i suppose you could argue that you like the sound of THAT specific cap in that location, but it is totally un-necessary as far as the statement "always do it this way" is concerned.

as far as the specific question of why the PPA uses film caps for decoupling the PSU rather than tantalum, i dont really know. tantalum caps although non-linear do EXCEPTIONALLY well at high frequency. please dont make me laugh and say that the 1uf film caps are there to supply power to an SS amp at low frequency...
 
Oct 21, 2007 at 4:43 PM Post #5 of 49
Are we talking about bypassing or decoupling here? For bypassing, when you have fairly large values of capacitance, 10-100 uF, film or non-polar caps can be exhorbitantly expensive, and may not even run up to those values. Try looking up 100uF film caps at Mouser or digikey, and tell us what you find! As for decoupling op amps or other smt chips, where low values of capacitance (and impedance) are needed, I don't really understand why you would use film caps either. The best caps for decoupling have low impedance, and those are surface mount C0G or X5R, preferably 0805 or smaller placed as close as possible to the supply pins - not bulky through-hole film caps. Yeah, they may look better - but they don't serve the purpose as well.
 
Oct 21, 2007 at 7:42 PM Post #7 of 49
Oct 21, 2007 at 8:55 PM Post #8 of 49
Quote:

Originally Posted by ezkcdude /img/forum/go_quote.gif
Are we talking about bypassing or decoupling here? For bypassing, when you have fairly large values of capacitance, 10-100 uF, film or non-polar caps can be exhorbitantly expensive, and may not even run up to those values.


At the first place, why you need film cap with such a HUGE value?
smily_headphones1.gif
Why don't you have "Elec.Cap+FilmCap" as combined effort to make the bypassing/decoupling effect better. Speaking for filmCap in "Elec.Cap+FilmCap" configuration, I use PPS film as the best choice.
 
Oct 22, 2007 at 12:23 AM Post #9 of 49
Quote:

Originally Posted by nikongod /img/forum/go_quote.gif
IIAC (if i am correct) you are wrong. both of the above are electrolytic types. Both do use "very unique" electrolytic "mixes" but they are electrolytic types none-the-less.


I checked the Nichicon website, and apparently they make Muse in both normal and tantalum. I'll make a mental note.

Quote:

why use a non-polar cap in a PSU application, or one where the cap will ALWAYS have a voltage differential facing the same way? i have always been at a loss to explain why people use bipolar caps on the outputs of tube & hybrid tube amps. i think at some point people equated a non-polar electrolytic with a film cap because both are non-polar, too bad one is an electrolytic and the other a film. i suppose you could argue that you like the sound of THAT specific cap in that location, but it is totally un-necessary as far as the statement "always do it this way" is concerned.


I thought he was talking about using decoupling caps in the signal path. From what I've always read, there are A/C characteristics in an audio signal, and polarized electrolytics are said to cause more distortion. While it may not be necessary, it can't see how it would hurt.

Quote:

as far as the specific question of why the PPA uses film caps for decoupling the PSU rather than tantalum, i dont really know. tantalum caps although non-linear do EXCEPTIONALLY well at high frequency. please dont make me laugh and say that the 1uf film caps are there to supply power to an SS amp at low frequency...


There's an interesting snippet in the Jung/Marsh article about picking capacitors, and they compare tantalums, ceramics, electrolytics and films. They claim to hear audible distortion when using ceramic capacitors even in the PSU stage of an amplifier.

Edit: PartsExpress sells metallized polypropylene up to 200uF.
 
Oct 22, 2007 at 8:43 AM Post #10 of 49
Quote:

Originally Posted by Logistics /img/forum/go_quote.gif


I thought he was talking about using decoupling caps in the signal path.



Decoupling caps aren't in the signal path; caps in the signal path are coupling caps!
 
Oct 22, 2007 at 10:17 AM Post #11 of 49
Quote:

Originally Posted by Jambo /img/forum/go_quote.gif
Decoupling caps aren't in the signal path; caps in the signal path are coupling caps!


If used from rails to ground and if not having an active ground, from what I understand they're somewhat in the signal path, sinking current. But I'm not really sure. I don't find it easy to understand how grounding/sinking/sourcing works.
 
Oct 22, 2007 at 11:26 AM Post #12 of 49
When we're talking about the small value capacitors on a power supply rail (either to ground or to the opposite rail, used to augment the bigger electrolytic rail filter caps), "bypassing" = "decoupling". The two terms are used interchangeably. The purpose of such a cap is to keep the rail impedance low at high frequencies.

That is not to be confused with signal "coupling" caps, usually found at the input or output of an amplifier, or sometimes in between stages of an amp, or in the feedback loop. The function of signal coupling caps is usually to block DC voltage from passing through. Such coupling caps could also be bypassed with another smaller, higher performance cap.

Confusing? Yes, but that's the way it is.
 
Oct 22, 2007 at 2:28 PM Post #13 of 49
Quote:

Originally Posted by NelsonVandal /img/forum/go_quote.gif
If used from rails to ground and if not having an active ground, from what I understand they're somewhat in the signal path, sinking current. But I'm not really sure. I don't find it easy to understand how grounding/sinking/sourcing works.


Well, you do have a point, but if you think of it that way everything is in the signal path and it becomes impossible to distinguish between things. Amb's description is good, I guess decoupling all comes down to parasitics and the fact that there is no such thing as a perfect power supply.
 
Oct 22, 2007 at 3:41 PM Post #14 of 49
Quote:

Originally Posted by Logistics /img/forum/go_quote.gif
There's an interesting snippet in the Jung/Marsh article about picking capacitors, and they compare tantalums, ceramics, electrolytics and films. They claim to hear audible distortion when using ceramic capacitors even in the PSU stage of an amplifier.


This is sth I often heard from people around me saying. Caps even in PSU stage are also doing sth to the audio output of the amp.

Have anyone here verified this yet?
 
Oct 22, 2007 at 4:07 PM Post #15 of 49
^
Because there aint no such thing as a perfect power supply(capacitor).

Example:

Tantalums = works great at high frequencies BUT Terrible stability and accuracy.

Ceramics = even better at higher frequencies BUT REALLY terrible linearity and stability.

Electrolytics = Well....only good thing about them is the huge ass capacitance.

Film caps = self explanatory, doesnt work quite as well as tants/ceramics at high frequencies BUT does have a good linearity , accuracy etc.

THERE'S A CATCH IN EVERYTHING!! ARGH!!

So in Electronics, a compromise to achieve balance is needed to optimize the circuit.
 

Users who are viewing this thread

Back
Top