advertisement


Using polarised caps 'back to back' to create 'super cap'

John_73

John_73

pfm Member
Hi folks!

I've read in various forums that using two polarised capacitors back to back can create a higher quality non-polar cap.

Can mixing two different types of cap - but of the same value - be used back to back to mix the best virtues of each, to hopefully creater an even better quality cap? To illustrate this better was thinking of using a 100uF Elna Silmic II with a 100uF Elna Cerafine, back to back, to create a 50uF (have I got that right?) non-polar cap which may well sound better than using one type of cap on its own.

Any opinions - anybody tried this?

Cheers!
 
But in this instance the caps are back to back (+ to + or - to -) so would this still be the case then? (i.e. doubling the value?)
 
absolutely - they are in series - regardless of polarity. And I believe this is a standard means of adapting electrolytic caps for AC use...
 
In series the total capcitance will halve, but the various spurious contributions of the caps (ESR, ESL, DA, tan delta) will add. I think you'll get a cap that is worse than either...
 
valveheaduk said:
Hi folks!

I've read in various forums that using two polarised capacitors back to back can create a higher quality non-polar cap.

Can mixing two different types of cap - but of the same value - be used back to back to mix the best virtues of each, to hopefully creater an even better quality cap? To illustrate this better was thinking of using a 100uF Elna Silmic II with a 100uF Elna Cerafine, back to back, to create a 50uF (have I got that right?) non-polar cap which may well sound better than using one type of cap on its own.

Any opinions - anybody tried this?

Cheers!
Click to expand...

I think you'll find this is a Black Gate concept for use with their 'N' & 'NX' non-polarised caps. The result is what they call their "Super-E" configuration.

Regards,

Andy

PS: And no, 100uF Silmic II back-to-back with a 100uF Cerafine, gives a 200uF cap.
 
Cyril Bateman did a mamoth eight article capacitor distortion test project in Electronics World starting July 2002. He found that two standard aluminium electrolytic caps back-to-back produced quite low distortion, considerably lower than half of one elcap. He then tested two bipolar caps back-to-back and found that these produced lower distortion again, as good as polyester/mylar/PET types. He patented the combination.

I was smitten at the time and built one of his test jigs, but came to realize that the factors that Martin mentioned (and space to put the cap) are almost always more important than distortion per se.

Tants are absolutely the worst for distortion by a couple of orders of magnitude, yet Naim and many of us find no obvious problems with them.
 
It should be no great surprise that distortion drops in series combinations, simply because the AC voltage across each cap reduces in proportion to the number in series. But thd% alone is by no means a shorthand for 'good sound' imo - i think 'good numbers' are a necessary but not sufficient condition.
 
martin clark said:
It should be no great surprise that distortion drops in series combinations, simply because the AC voltage across each cap reduces in proportion to the number in series. But thd% alone is by no means a shorthand for 'good sound' imo - i think 'good numbers' are a necessary but not sufficient condition.
Click to expand...

Not sure what you're getting at here (with talk of series caps)? "Back to back" IMO means using 2 non-polarised caps (in particular BG 'N' or 'NX') in parallel but with the '+' and '-' leads reversed. That is the BG "Super-E" configuration, any way. :)

Regards,

Andy
 
The original poster did say 100 + 100uf caps to give 50uf. That of course has to be in series and is useful sometimes.

In parallel you could get lower combined ESR etc, but not sure how you'd obtain the correct value?

Anyway my (minor) experience is that I like one component to do a job, especially when it comes to signal paths, but I can see how a blackgater would like to put more caps in their circuits.
 
I think valveheaduk was talking about this:

+CAP--CAP+

Imagine two peoples\cars\etc "Back to back" and You will undertand that this is not the same as paralleling something :)
 
andyr said:
PS: And no, 100uF Silmic II back-to-back with a 100uF Cerafine, gives a 200uF cap.
Click to expand...

Sorry Andy but NO - putting 2 100uF caps back-to-back (i.e., in series) will result in a 50uF cap - if you don't believe me just put any 2 caps back to back (in this case 2 old 10uF electrolytics) and try measuring the result...

... two in parallel would give you 200uF.
 
Stream said:
I think valveheaduk was talking about this:

+CAP--CAP+

Imagine two peoples\cars\etc "Back to back" and You will undertand that this is not the same as paralleling something :)
Click to expand...


Thats exactly what I meant - sorry for any confusion.

Will give it a try once I get my Quad IIs rebuilt - this will be for cathode bypass duty. Will also try a single cap on its own.

I do like the sound of the Black Gate N, but being as these are nearly phased out been looking into alternatives and thought a cerafine + silmic combined might offer a nice alternative. Besides which (as in the BG N) anything that takes over 300+ hrs to burn in, and THEN needs to be kept on constantly (to avoid doing it all over again) isn't suitable for practical Hi Fidelity IMHO. 300+ hours of naff/variable sound is 300 hours of listening you'll never be able to get back again afterall ;)
 
neiljadman said:
Sorry Andy but NO - putting 2 100uF caps back-to-back (i.e., in series) will result in a 50uF cap - if you don't believe me just put any 2 caps back to back (in this case 2 old 10uF electrolytics) and try measuring the result...

... two in parallel would give you 200uF.
Click to expand...

Hehe - I obviously have a different concept of what back-to-back means! :D

To use the people analogy mentioned earlier in this thread ... if two people are standing b-t-b and holding hands then one person's right hand is holding the other's left hand. Translated to caps ... the 2 cap's each have a long lead connected to a short lead (ie. they're in the same hole in the PCB).

If 2 people were holding hands and standing in line then, yes, the right hand of one person would be clasping the left hand of the other ... but each person's other hand would not be holding anything.

Regards,

Andy
 
hacker said:
This is why we use the terminology series and parallel ;)
Click to expand...

Of course ... but what would you call BG's "Super E" configuration for their non-polar caps? (IE. where 2 caps are placed in parallel ... but with the '+' & '-' leads opposite.) :eek:

Regards,

Andy
 
hacker said:
Antiparallel.
Click to expand...

That's an interesting wiki definition, hacker. However, a logician might be tempted to debate it! :p

The wiki entry says: "... Two LEDs can be paired this way, so that each protects the other from reverse voltage."

Is this true for polarised electrolytics? I can understand how you can quite safely reverse the leads of a non-polar cap but (having had a recent unfortunate experience with a pair of electrolytics! :( ) I would've thought that one should never ever connect standard electrolytics the wrong way round?

Regards,

Andy
 
Yeah, I agree - it sounds like it'd be a bad idea to do that with polarized caps. I think one of the more experienced bodgers needs to fill in the blanks for us ;)

Cheers,
Carl
 
This quote comes from the following discussion on using back-to-back electrolytics to form a non-polarised capacitor: http://yarchive.net/electr/electrolytic_caps.html.

This works, but you do not need to double the capacitance.

When an electrolytic capacitor is reverse-biased, it conducts if the
applied voltage is greater than approximately 1.5 volts. So when one
capacitor is reverse-biased, the other capacitor is charged. As the
polarity reverses, the roles reverse.

If you measure the capacitance with a very low voltage bridge, the
capacitance will measure half the capacitance of each unit. But if you
increase the voltage the capacitance measured will increase as you get over
the 1.5 volt excitation level. This diode action is why non-polar
electrolytics generate so much distortion in audio circuits.

Andy
 
You must log in or register to reply here.


advertisement


Back
Top