The most linear electrolytics (lowest DF)

[linuxguru]

This thread is to discuss findings of electrolytic capacitors which have the lowest measured dissipation factor (but not necessarily low ESR). This generally translates to audibly better-sounding capacitors when used either in the audio signal path or as bypasses.

So far, I have found the following:

Brand/Series Tan(delta)
------------------- --------------
Rubycon Black Gate PK 0.4%
Elna Cerafine (Black/Gold) 0.5%
Rubycon YXA 0.3%
Rubycon YXG 0.5%
Elna Silmic II 0.6%

In comparison, a Wima MKS2 film cap has a typical tan(delta) of around 0.1%. Most aqueous and low-ESR electrolytics are no better than about 1%, though there may be exceptions. The aim of this thread is to find out which ones have comparatively low tan(delta). I measure tan(delta) using a neat little AtMega-based component tester that's similar to the ones discussed at this thread:

http://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/

Some of these display tan(delta) and some don't, depending on the sophistication of the measurement firmware - look around on EBay for the ones with a green ZIF socket for the DUT.

[budm]

Re: The most linear electrolytics (lowest DF)

Df is the calculated number based on the ESR measured value at known frequency and Xc at known frequency, you do not really measure the Df direct, the test equipment takes those data and calculate the numbers for you.

[linuxguru]

Re: The most linear electrolytics (lowest DF)

Df is the calculated number based on the ESR measured value at known frequency and Xc at known frequency, you do not really measure the Df direct, the test equipment takes those data and calculate the numbers for you.

Thanks, I'll read it carefully. The meter has a microcontroller that takes a few seconds to take some readings, and computes and displays C, ESR and something called Vloss in %. That Vloss shows values that are consistent with the published tan(delta) figures in the datasheets.

Also found another extremely low Vloss capacitor - a discontinued/obsolete Panasonic SMD series called V-HA: 10uF/100V shows 0.3%. Apart from Rubycons (Black Gates and a few others) and Elna (Cerafines and a few others), I have found very few electrolytics so far that go below 1%, leave alone 0.5%.

[mariushm]

Re: The most linear electrolytics (lowest DF)

You should really use a proper lcr meter for such measurements. Hate to sound like that greek guy but if you're serious about it, the cheap small meters are probably not accurate enough.

Otherwise... my opinion is that sometimes it's really subjective. As an example, I've had quite a few guys say Panasonic FC capacitors "sounded" better than Silmic capacitors to them.

[budm]

Re: The most linear electrolytics (lowest DF)

Let your ears be your final judgement to what you like, it is very suggestive as mariushm pointed out. It is more than just the cap, you have to look at it as the whole system from the source all the way to the speakers/or headphone, the room you are in, so many factors. You can have two amplifiers with exact same spec hook up to the same set of speakers, and they will not sound the same. Just listen to what sounds good to you. Because some one spend $1000 on a pair of speaker cables, they must be good.

[linuxguru]

Re: The most linear electrolytics (lowest DF)

Df is the calculated number based on the ESR measured value at known frequency and Xc at known frequency, you do not really measure the Df direct, the test equipment takes those data and calculate the numbers for you.

Neither of those papers seems to talk much about Dielectric Absorbtion, aka Soakage, which is a non-linear phenomenon with hysteresis. The first paper seems to be talking about an idealised R + Xc *linear * capacitor model - but in reality there's an additional non-linear component to the loss in a capacitor, due to soakage. Some of that energy re-appears as a small residual voltage at the terminals of a capacitor when measured with a high-Z voltmeter, even after the capacitor has been fully discharged by shorting out its terminals for a while.

[rievax_60]

Re: The most linear electrolytics (lowest DF)

I stopped worrying about this long ago.
A properly sized coupling capacitor should have so little signal voltage drop across it that any non-linearity of the capacitor will result in little distortion to the overall signal.

[budm]

Re: The most linear electrolytics (lowest DF)

If you worry so much about non-linearity of the caps, how about the semiconductors or even tubes, transformers, speakers, etc.