[yejun]

For coupling capacitor which type is best?
In wima's website, MKS(metallized polyester) is used for coupling.
http://www.wima.com/applicguide.htm

 

[Squalish]

This might help.
For coupling purposes, non-metallized polypropylene, or Teflon(not practical) seems to be the best choice, but metallized polypropylene beats metallized polyester easily, and is available in the 5uf to 50uf range.

 

[yejun]

Quote:

The metalized polyester unit is far better than either electrolytic, measuring less than 0.15 percent. This may be quite good for polyester types, as typical specification data available do not always show comparably low figures. The metalized polypropylene unit is extremely good in terms of DA with a measured figure which compares quite well with the manufacturer's data.

no idea, but why wima site use mks as coupling?

 

[Tomo]

Hey,

I would stay well away from polyester for signal coupling. I use them for other purposes. I don't care what WIMA says.

Polypropylenes are actually most suitable for audio. You will realize that it works very nicely with audio frequencies. Also, unlike Teflon caps, it is far easier to obtain. (I don't know any manufactures of Teflon caps that I trust, anyway. All kinda obsecure.)

Metalized means that metal layer is bonded onto the dielectric. This makes the capacitor smaller in size, but the metal layer is so thin that it has some resistance. Foil Capacitors have real metal foils. They tend to have higher current capabilities. Like Squalish says, they are better than metalized counterpart. However, be warned dielectric failure in Foil caps will be permanent and should be avoided if the cap is DC-Blocking. Metalized caps on the other hand can recover from that failure. (So better suited in PSUs)

Tomo

 

[morsel]

Here are some numbers for dissipation factor (DF), smaller is better:

Polyester (PETP) .5%
Polycarbonate .05%
Polyphenylene sulfide (PPS) .05%
Polystyrene .05-.02%
Polypropylene .025-.01%
Teflon (PTFE) .025-.01%

These numbers were culled from over 100 documents on the web.

 

[peranders]

Quote:

Originally posted by yejun For coupling capacitor which type is best? In wima's website, MKS(metallized polyester) is used for coupling. http://www.wima.com/applicguide.htm

Normal small polyester like Ero 1816 aren't so bad but I feel better if I use a polyprop. In my latest SMD headphone amp I have used PPS, polyphenylene sulfide.

Dielectric absorbtion is rather unimportant parameter in AC applications like audio. It VERY important in high precision dual slope ADC and similar DC applications.

My opinion is that coupling caps are a rather uncritical part as long as "small" signals goes through it, especially if you have no or little DC in the signal but in a tube amp or power amp I can imagine that it's more important to choose the right quality.

I wonder when dissipation factor start to get noticable. Is 0.5% "much"?

Morsel, "cull" was a new word for me. Never heard it before.

 

[aos]

WIMA probably mentions MKS for coupling because, guess what, their customers are primarily big companies where cost is the most important thing. And polyesther is what commercial gear would use (if not using electrolytic which they usually do).

I also use PPS in my portable DAC, in the filter. I still use Black Gate NX for signal coupling. I've tried Auricap - which is large metallized polypropylene cap - but I didn't like it. For some reason BG NX sounded better. Perhaps Auricap was designed for speaker crossovers so when you pass low voltage / current signal though it, it doesn't sound as transparent. Still the best cap is no cap - but often you don't get to make that choice.

 

[jamont]

Quote:

Originally posted by morsel Here are some numbers for dissipation factor (DF), smaller is better: Polyester (PETP) .5% Polycarbonate .05% Polyphenylene sulfide (PPS) .05% Polystyrene .05-.02% Polypropylene .025-.01% Teflon (PTFE) .025-.01% These numbers were culled from over 100 documents on the web.

I'm in awe of your diligence in tracking down these numbers! Can you say any more about exactly what this measures?

 

[morsel]

DF is the loss of a capacitor. Check out these classic documents for more info:

Picking Capacitors
The "Sound" of Capacitors

I disagree with Peranders regarding DF. I think there is a strong correlation between low DF and better sounding coupling caps. This is based upon the personal preference of dozens of people, who by and large seem to agree that certain caps llike polypropylene film and foil, which have low DF, sound better than higher DF caps like polyester.

Peranders, I like words, too. My latest acquisition is "undrentide", which means "late morning" in Middle English:

Sir Orfeo

 

[Tomo]

Oh, I don't know.

Quite a few people exclusively use paper-in-oil caps. They have higher dissipation. Some people even use Vitamin Q's, now that stuff really change your sonics. Apparently, they like the "change."

Don't get all caught up in DF. I say Polyester is bad, because when I was using 225P (Orange Drop, Polyester) I got this ringing in the high end. You would think polystylene are good but I think it is too stale.

I gotta try out BG NX/N when I get a hand on them.

T

 

[peranders]

Quote:

Originally posted by jamont I'm in awe of your diligence in tracking down these numbers! Can you say any more about exactly what this measures?

A small part of the cap is "R" and the rest is "C". You loose some signal beacuse of this "R". If it's linear, it really doesn't matter but I suspect that it's not like everything else in a cap. Check for example a horror type like XR7 or SiBaTit ceramics. Sensitive to temp, voltage, frequency, everything!

 

[aos]

I heard paper in oil also changes the sound...

By the way I tried N and didn't like it. It seems that the size/voltage of the cap also matter, and I'm not talking about corner frequency. I had <10uF 25V series N and 6.3V 47uF or so series NX and I prefered NX over N easily. They were about the same physical size. So I am guessing either voltage or capacity or both matter - too small a cap and it won't sound as good. You gotta experiment in every situation...

 

[Tomo]

Aos -

It is true, I have experimented with Paper-in-oil myself. It is kinda nice. It would even nicer, if the foil is more conductive stuffs. ie Jensen or VTV.

But just so you know, you are "crafting" sound here. Depending on amplifiers, caps will have different effects. I wouldn't give judgements using it on single amplifier.

PerAnders -

Don't be so horrified. I like Ceramics where they are more suitable. Besides, who would use ceramics for coupling anyways?

T

 

[jeffreyj]

Quote:

Originally posted by peranders .... Dielectric absorbtion is rather unimportant parameter in AC applications like audio. It VERY important in high precision dual slope ADC and similar DC applications. ... I wonder when dissipation factor start to get noticable. Is 0.5% "much"?

Regarding DA: In circuits in which the average voltage is zero (true AC), I'll agree with the statement that dielectric absorption is not relevant (go silver mica!). But if there is a DC bias present, then DA becomes one of the worst errors because it will essentially supply an input signal when there ought to be none (that is, the absorbed charge fights the true input signal as it goes toward the opposite polarity).

Regarding DF: the absolute value of DF is not so important, within reason. It is the change in DF with frequency, voltage, temperature, whatever that matters. Paper and oil capacitors have high DF, but it is very stable and so it is usually not offensive. Z5U ceramics and electrolytics have the the worst change in DF over any other varied parameter, and consequently are disasters when put into the signal path!

Polyester's DF is also erratic - not quite so bad as high-permittivity ceramics, but still up there - and it is because of this that it is a poor choice for signal coupling. Makes a great bypass capacitor, though.

edited to add the following response to Morsel's comments:

Those PPS and PP dielectrics not only have a low DF to begin with, but very little delta DF w/r/t pretty much any other parameter changed! This is why I take plastic over paper, and not just at the grocery store

 

[jeffreyj]

Quote:

Originally posted by Tomo ... PerAnders - Don't be so horrified. I like Ceramics where they are more suitable. Besides, who would use ceramics for coupling anyways? T

Note that NPO dielectric performs very well with respect to changes in termperature, frequency, voltage, etc., and therefore is an excellent choice for use in the signal path where low capacitance values are needed! Indeed, it may be the only choice when values under a few hundred pF are specified (across the feedback resistor, say, to prevent high-speed amplifiers from oscillating).

X7R, Z5U, Y5P, X5R, Semiconductor, etc. dielectrics are so different from NPO that the only thing in common is construction method; other than that it is misleading to lump NPO with them.