polypropylene capacitors

^ "It is very easy to abuse these and cause then to fail"
Wima calls them "self-healing" in their datasheet. ;)

At the input of amps in most designs , it's a 2,2 uF with a 47k resistor and a series R of 270-1k for low noise . I don't get why they keep the 47k . Doesn't it cause a lot more noise ? That's why I wanted a bigger cap , 10uF with a 10k , or even more than 10uF.
But there is no space for those big MKP's that go to 4 cm or more.
Placing those 2 "fatties" (L & R) so close together , can't be good for channel separation.
 
Semantically, for sematic's sake

Germans (or other Europeans) use "U" for AC voltages in convention for single phase systems....
But not for DC voltages.

So its quite correct to use U to refer to ACrms voltages, and V for DC Voltages (as well as L1, L2, L3 or UVW for 3 phase systems)

Self healing would be good, but well it doesnt work all that well, I have a bucket load with insulation resistance in the Ohm range...
 
RickTH said:
At the input of amps in most designs , it's a 2,2 uF with a 47k resistor and a series R of 270-1k for low noise . I don't get why they keep the 47k . Doesn't it cause a lot more noise ?
No. When a source is connected its low output impedance (if correctly designed) will short out any noise generated by the 47k.

Placing those 2 "fatties" (L & R) so close together , can't be good for channel separation.
That may be a reason why simply putting in big caps is not always a good idea. However, once again the output impedance of the source comes to rescue you.

mondogenerator said:
Self healing would be good, but well it doesnt work all that well, I have a bucket load with insulation resistance in the Ohm range...
Self healing require a sufficiently low impedance source that a short gets hot enough to do the self-healing. Unlikely for a coupling cap.


The plots shown some posts earlier probably just show the limits on AC current through the caps. Again of no relevance to most audio duties.
 
Last edited:
DF96,
I didn't know the particular mechanism causing self healing of some polymer film capacitors, so thanks :)

I have experienced self healing of capacitors before, on one or two occasions, but these were higher voltage polyester types, epoxy dipped things (and they have a tendency to fail in other ways),
I suppose the dielectric has it's own influence on that healing action, thinking aloud.

I recall hearing tales of submarine telephony repeaters using silver mica that cumulatively failed to a point they all needed replacing with self healing types. That's the sum total of my knowledge!

Mostly the 2u2 Wimas failed SC when used as coupling caps for the input of the first stage valve grid, literally the input coupling DC blocker.

Cause of failure was solder spatter/solder sucker solder 'slag' or something else shorting stripboard tracks....
 
Those tests appear to show that distortion was below the measurement threshold for all film caps, so in what sense is MKP the best - based on those tests? DA does not count because DA does not affect audio.

As stated before, we compared in our first controlled listening tests in the 80s different foil types as coupling caps got the result that the mkp (metallized polypropylene) type was preferred when compared to the mks (i.e. polyester) wima, at least in the low voltage application that we used for the comparison.
 
Max Headroom asked about the "Nakamichi" feedback circuit. a few posts back - here's the thread where I talk about it, post #19 or thereabouts...
Feedback capacitor to minimize DC offset
Thanks for the link.
I well agree that the typical FB shunt cap is mission critical to amp sound and have improved several amps by changing to BP electro or types like Silmic etc.
This circuit is a useful way to use film caps in this critical circuit position.
Have you experimented/auditioned such circuit ?.


Dan.
 
Looking around on various sites : Why is it so hard to find small (so low voltage) MKP (polypropylene) capacitors for audio ?
There are : AC/puls , DC link , motor run, power factor correction, RC snubber, RF and power , RF microwave, suppression film capacitors and... general use .
Which are for audio , in the 4,7 uF to 10 uF range ? ( for input of amps , not for speaker filters)
Sure I can find these Wima red ones and those grey Kemet, but often only for high voltages which makes them very bulky.

- What value do you need?
- Are you looking for "Audio Grade" components

Audio Grade normally means "low performance components, that is not accepted in industrial applications". Some gold painted text saying "Audio Grade" transforms these parts from ultra high cost parts :D

If I needed caps in the signal path of an amplifier, I would go for PPS.
However..... If I needed caps in the signal path, my first worry would be, that the design is not finished :rolleyes:
 
The problem with the 47k load resistor is when nothing's connected - then it is indeed a source of noise. A partial fix for this is to use a socket with a grounding switch that activates on unplugging. Alas a lot of modern sources (mp3 players, phones, etc) power-down their audio output chip when its not active, thus removing the low impedance source even though its still plugged in, and generating audible hiss at the end of playback.

One solution is to have a switchable 600 ohm load on each input that can be deployed for known low-impedance sources.

Using a lower load resistance can create problems if the source isn't a low impedance and starts to distort on what it sees as a heavy load - this is fairly rare these days.
 
- What value do you need?
- Are you looking for "Audio Grade" components

Audio Grade normally means "low performance components, that is not accepted in industrial applications". Some gold painted text saying "Audio Grade" transforms these parts from ultra high cost parts :D

If I needed caps in the signal path of an amplifier, I would go for PPS.
However..... If I needed caps in the signal path, my first worry would be, that the design is not finished :rolleyes:

Never heard about PPS until I found this forum . Mouser only has 2 stocked and they have reasonable size & price:
KEMET Polyphenylene Sulfide (PPS) Film Capacitors | Mouser

Doesn't have to be audio grade , just the best for audio.:)

If true that the 47k doesn't matter for noise , then I could go back to what nearly all amp input designs have : 2,2uF + 47k.

"If I needed caps in the signal path, my first worry would be, that the design is not finished"
I don't like caps in the signal path either but my mp3 player has a 4mV dc on it's output , so a cap on the input is a must.
I also had an opamp fail and put -12 Vdc at the input of a dc coupled preamp, luckily the CD player had a cap on it's output , so no damage.
My computer has a bridge output , no caps , so without a resistor and a blocking cap , an opamp failure would be a disaster.
 
Never heard about PPS until I found this forum . Mouser only has 2 stocked and they have reasonable size & price:
KEMET Polyphenylene Sulfide (PPS) Film Capacitors | Mouser
My computer has a bridge output , no caps , so without a resistor and a blocking cap , an opamp failure would be a disaster.
Hmmm....that PPS cap sure is expensive! The polypro is ~ the same size and much cheaper!:
Panasonic ECW-FD2W225JB 2.2µF;450v;7.5mm L.S.; 97¢
 
U is an alternate symbol for a voltage...

U in physics is potential or internal energy...

In Germany, and maybe other parts of Europe, voltage is U and not

This is hilarious and so typical of the poor terminology seen so often here on diyAudio!

***RANT MODE ON***
In proper physics (and engineering), proper physicists (and engineers) define their variables ("letters") before using them. There is no overarching convention that universally defines the meaning of specific letters in terms of physical or engineering terms / variables. If you want to do it right, tell everyone how you use the letters.
***RANT MODE OFF***
 
***RANT MODE ON***
In proper physics (and engineering), proper physicists (and engineers) define their variables ("letters") before using them. There is no overarching convention that universally defines the meaning of specific letters in terms of physical or engineering terms / variables. If you want to do it right, tell everyone how you use the letters.
***RANT MODE OFF***

Sounds like the tower of Babel to me -- is "S" to be Siemens or seconds?
 
<snip>
If true that the 47k doesn't matter for noise , then I could go back to what nearly all amp input designs have : 2,2uF + 47k.

<snip>

DF96 was correct about the 47k noise contribution (it follows from network analysis principles, which of course are based on idealizing assumption, but for audio bandwidth and conservatively designed gear are appropriate), but the 2,2 uF/47k combination isn't conservative enough. If sources and speakers allow reproduction of really low bass, then you should use ~ 10uF or raise the input impedance (if possible,when considering the specific conditions) in case of the smaller capacitor value.