Not polarized electrolytic capacitor in power supply (LM317 LM337)

[ygg-it]

Hello, I have just read some web articles that say that non-polarized electrolytic capacitor sounds like film capacitor (by Cyril Bateman) and also this

So, why don't put them in power supply??

Then I read some articles that say that low value capacitors make LM317/LM337 to sound better (by Eric Juaneda)

Well I make both the tries.
I lowered the cap from 6600uf to a not-polarized 235 uF !!

Terrific!! The sound improved dramatically !!!

The sound changes so much that different brand of same value capacitors change tonality !! For example ELNA makes sound like a valve, but i preferred old blue Philips Cap (K7 serie ?). I did not hear any noise change even!

Capacitors run cold, so I think I didn't make mistakes...
But I never seen in the web a circuit with not polarized elecrolytic in power supply...Why?
The other doubt that I have (but I didn't try) is if I have to parallel each capacitor with 47k-100k resistors

Any comment is welcome ...

[lineup]

It is recommended in datasheet of LM317/337
that to put 10 uF from adjust pin to GND. To lower noise.
Try this instead of 10 nF

About sound.
A good amplifier should not change sound
because of a small change of capacitor at regulator input.

I am quite sure, that if you did not know what capacitor was installed,
you would not be able to tell what sounds best.

Then if you have success in a valid test,
where you do not know what you should hear
I will change my opinion.

Regards

[ygg-it]

You are right.
But the discussion here is about not-polaryzed capacitors in power supply, since I never saw a circuit with them.

[Steerpike]

6600uf to a not-polarized 235 uF

But you've changd TWO things at the same time, both the 'polarisation' issue of the capacitor, and its value - quite substantially.
What about trying a 230uF POLARISED cap instead?

230uF is a very low value of smoothing capacitor for a power supply - there may be substantial ripple on the DC lines; could that have a pleasant sound?

Non-polarised electrolytic capacitors generally don't like having a DC voltage imposed upon them - it shortens their life. That's why they aren't used.

[Conrad Hoffman]

I've read a variety of things about NP electrolytic caps, some of which are contradictory and some of which suggest they are often misapplied. Not sure what to believe, but measurement-wise they aren't that great and I'll always use a polarised cap if at all possible. I think NPs sound inferior to polarised electrolytics, where the latter can be used.

The most weird thing I've read is that they shouldn't be used where the polarity across the cap reverses, like a speaker crossover. They should be biased one way or the other. Obviously they work for speaker crossovers and cap companies know that's where people use them, so the recommendation makes little sense to me. It was, however, from a cap companies literature, though I don't remember which one. I suppose if the cap is constructed using two polarised units back-to-back, it makes some sense.

I can't think of any reason that prevents use as a DC filter cap, except they are less space efficient and don't perform quite as well as a good low esr polarised cap.

[ygg-it]

>But you've changed TWO things at the same time, both the 'polarisation' issue of the capacitor, and its value - quite substantially.
What about trying a 230uF POLARISED cap instead?

You are right, unfortunately I tested this unique solution. I knew thet lower cap sounds better and I knew that Bi-polar electrolytics sound like film capacitor. So I changed once a time with both solution. I realise that the good test result can be the effect of one or the second or both the solution...

> 230uF is a very low value of smoothing capacitor for a power supply - there may be substantial ripple on the DC lines; could that have a pleasant sound?

Please read Eric Juaneda articles. And also I don't hear any hum change in my headphone. This test confirm Eric theory.

> Non-polarised electrolytic capacitors generally don't like having a DC voltage imposed upon them - it shortens their life. That's why they aren't used.

Thank you. I didn't know

> I can't think of any reason that prevents use as a DC filter cap, except they are less space efficient ...

This confort me, thank you for your reply

> and don't perform quite as well as a good low esr polarised cap.

This post says the opposite ...