Non polarized Caps

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
My father said that years ago he saw a circuit, to make a equivalent circuit to a non polarized capacitator using normal polarized ones but he doesent remember where or how it was made..

This could bring my costs down by alot...

Anyone here ever saw this somewhere? and if it exists, how good is it?
 
Ilianh said:
My father said that years ago he saw a circuit, to make a equivalent circuit to a non polarized capacitator using normal polarized ones but he doesent remember where or how it was made..

This could bring my costs down by alot...

Anyone here ever saw this somewhere? and if it exists, how good is it?

Hi,
All you need to do to make this work is to connect two of the same value electrolytics (or any other polarised cap, for that matter) "back to back".

I.E. connect the two negative leads together, or connect the two positive leads together, and then use the two remaining free leads as the 'new' combined cap's leads.

Of course, to achieve the same ultimate value of cap in a case like this, you need to double the value of the two individual caps you use, as, by connecting them in series, it halves the resulting value.

If you are sufficiently keen to get the best out electronics (like me!) you may wish to experiment with 'biasing' the junction between the caps (which is otherwise not connected to anything) by applying a suitable DC voltage to this junction.

Almost all, if not all, polarised caps appear to perform their function better when they have a distinct 'bias', or DC voltage differential across their two leads.

If you have connected the caps two positive leads together, you need to apply a (relatively) positive DC voltage to the junction, and similarly, If you have connected the two negatives of the caps together, apply a negative DC voltage here.

Different voltages can make a difference to the sonic effects, I have found, but you must stay within the overall voltage rating of the caps used, of course.

Take care when you are assessing the limit here, as it will be the difference in voltage between the 'bias' point and the 'worst case' of the other two cap leads.

For example, if the cap is used for say DC blocking in a signal path, where there is a positive voltage at the input of 20volts, but only + 10 volts at the outlet of the 'combined' cap, you would connect the two negatives together in this case, and bias the junction with a negative voltage (or a lower positive voltage than 10v), so that there is the correct polarity of 'bias' across both caps.
The maximum voltage you can set the 'bias' at would be determined by the worst case differential, so that if you were using say 63 volt caps, you could go up to say -40volts as a negative bias between the two caps (difference between +20v & -40v = 60v).

However, for safety, it would be usual to allow say a 20/30% margin with all such caps anyway, to stay well within their ratings, and any under all circumstances.

The quality and sonic attributes of the caps used in a case like this will determine how good the resulting sound is, but, if, anything, I feel that using caps in this way does very slightly improve their sonics, especially when they are suitably biased.

However, don't expect to "make a silk purse out of a sows ear", as if the caps you use are lousy sounding in the first place, it is unreasonable to expect this configuration to do a great deal to improve matters!

I hope this helps.

Regards,
 
Back to back

Of course, to achieve the same ultimate value of cap in a case like this, you need to double the value of the two individual caps you use
But if the reversed biased one isn't acting as a cap, why does the value halve?
I'm "picking" here, because I want to know the answer.:)

I was taught the same thing. But I don't know if it that simple:

When an electrolytic is reverse biased with any significant voltage, a lot of current flows, giving the appearance of a short circuit.

Consider a large AC signal across the composite cap:
Each cap will charge on the appropriate half cycle, whilst the other is almost a short circuit on the other cycle.
Well if the other cap is "short circuit", then the value isn't halved at all.

On the other hand, electrolytics, reverse biased with a small voltage, behave properly.
In this case the value will be halved because of the series connection.

So, I think that back- to - back series electrolytics' value, depends on the size of the AC waveform.
But, for DC, the value is not halved.

Who wants to use a component like that?:xeye:
 
Thanks for the information guys.

One thing I was wondering, If i integrate the crossover inside the speaker, I will not be able to bias the junction with a dc voltage.

How do the caps simply conected in series with the 2 same phase pins conected perform?
 
Ilianh,

Please note that I was not recommending or dismissing the use of this cap configuration in my post. I have insufficient experience of their use in this application.
What I was trying to do, was to stimulate a technical debate on a peculiarity I percieve with these components.

I hope there wil be other contributions.

Cheers,
 
Dhaen.

Hi,

Maybe I didn't make myself clear but I would never recommend
"reverse biasing" any electrolytic cap, as this is asking for trouble!

What I intended to indicate was, that no matter what the standing DC voltages are at each lead of the 'combined cap' arrangement, each individual cap in the pair should be 'biased' properly as far as *polarity* is concerned. (i.e. That the + lead is always more positive than the - lead).

In most cases like this, there will be probably no voltage at the 'output' side anyway, but if if there is some voltage here, and one does bias the middlle connection, both conditions (i.e. each individual cap's bias) must be considered.

In my earlier example, if one merely considered the 'input cap' with +20v at that point, one could try biasing the middle connection with say +15 volts, as there would still be a differential of 5volts (of the correct polarity) across that particular cap. (input of that cap at +20v, and output at +15v, which is fine).

However, it would not be satisfactory from the point of view of the 'output cap' as this would then see 5 volts (reverse polarity) across it, since the middle point (the - lead of the individual cap) would be sitting at +15 volts, but the other lead would see +10 volts. (this is a reverse polarity of 5 volts, and will damage the cap).

Sorry if I didn't explain it clearly before, and whenever any two caps are connected in series the resulting value is halved when they are of the same value.

Regards,
 
Hi Bob,

I was only commenting on the practice of connecting 2 caps + to +, and then using the negative leads as a "non-polarized" capacitor.
I've done this myself, and while I was uncertain of the combined value - it worked.
As far as I know, non-polarized electrolytics are actually this, but fitted into one can.
What I was trying to get across, was my feeling that the behaviour of such a cap is not straighforward.

In the case of polarizing voltages being applied (suitable circuit taken for granted), I believe the behaviour would become more sensible.

What I didn't make clear at first, was that I was talking "caps" generally, rather than staying "on-topic".

My "on-topic" response is that I'd use a naturally non-polar, non-electrolytic cap if one existed in the value required.

Cheers,
 
Hi Ilianh & dhaen,

Please remember that in my earlier post I answered the question which was originally asked, i.e. basically, how was this suggested non-polar arrangement made from ordinary (polarised) caps and how good was it?

This I believe I did, satisfactorily and comprehensively, together with appropriate caveats over voltage ratings and values etc., but on subsequent further query, I felt that I had not perhaps explained the application, and polarisation issues as well as needed.

To avoid any further possible misunderstandings, I should also say that I would not recommend the use of this dual cap arrangement (which I outlined above) in preference to a 'conventional' non-polar cap, although this was not part of the original query.

I was not, of course, aware that the intended arrangement was to be used in a speaker when I replied, and there are suitable non-polar caps readily available for this application.

The cost of non-polar caps is not generally significantly greater than polarised caps which would do the same job, and there would not be any real benefits to using my suggested dual arrangement in such a case.

However, as I also said originally, it would still be possible to do what I suggested *without* any 'biasing' between the two connected caps, and use this arrangement in a speaker, if, for, example some development work was being carried out whilst designing a new Xover, and suitable non-polars were not to hand.

Also, as mentioned before, it is unlikely that one can *significantly* improve the sonic performance of any polar electrolytic by this means, so my recommendation (now I am aware of the intended application) is to stay with caps which are designed for the job.

If the dual cap arrangement is to be used for any reason, my experiences are that thare are possible gains to be had by biasing the junction of the dual caps, as mentioned, and there appears to be a "slight improvement in their sonics" if they are biased properly.

Of course, the matter is irrelevant when dealing with non-electrolytic caps, as they are non-polar, anyway.

I have not come across anyone before, who wished to try such an arrangement in speakers, so I (incorrectly) assumed that the query was over caps which were intended to be used in a circuit in say an amp, or preamp etc.

There can be occasions where a larger value non-polar cap could be required in a circuit (than the usually available maximum value of just a few hundred microfarads for conventional non-polars), and, in such a case, my above suggestion would be how that could be achieved. Indeed, due to a lack of any explanation to the contrary, I thought this was why the query had arisen.


There is no DC "value" when related to a capacitor, so I don't understand the comment "for DC, the value is not halved" but there is a maximum voltage rating, which relates to the maximum voltage which can be applied across any capacitor's leads. In my earlier post, and in any conventional description of a capacitor, the "value" relates to the amount of capacitance (i.e. in microfarads etc), and this is always halved in the 'series' illustration I gave.

This maximum rating can be different for DC voltages, as opposed to AC, for any particular capacitor (you will often see the two different ratings marked on the same cap), and if a cap has both DC and AC applied to it, it is wise to add both voltages together to determine the total voltage which the cap will need to stand.

No caps *value* is "dependant on the size of the AC waveform" as has been suggested, unless, perhaps it is breaking down and completely failing due to over voltage, which, of course, should never be allowed to happen!

Also, when dealing with caps in series, the calculation for the 'combined' value is the reciprocal of the additions of the reciprocals of the two caps concerned. Regrettably I cannot show this well with my keyboard but it is :1/c =1/ca + 1/cb etc. for any number of caps in series.
Therefore, as I said, with any arrangement of two identical value caps in series, the resulting total capacitance is always one half of one of the caps.

With parallel combinations, you merely add all of the capacitances together, irrespective of their individual values which can all be different.

I hope that this is now clear, and I'm sorry, but there really is little more that I can add to this.

Regards,
 
Any ideas on this one?
 

Attachments

  • npcap.jpg
    npcap.jpg
    2.7 KB · Views: 346
diyAudio Senior Member
Joined 2002
1 ELCO + 1 ELCO=1 BLACKGATE?

Hi,

This is really something not very useful nowadays since non polarized caps are readily available and likely cost less than a comparable pair of polar electrolytes anyway.

It's however good to know this since, as Bob pointed out, this is how most (?) non-polar caps are still made to this day.

If you want to use this in a X-over I would rather recommend filmcaps even for impedance correcting Zobels.
Sounds so much better.

Just my:2c: .

Ciao,;)
 
well the thing is a 200 uF NP cap is around 10$, a polarized 400uF one is around 50 cents (even less) here.

for a 5 speakers system where you use second or third order Xo's... that might save you a nice 80$ or so, wich btw is alot of money...
considering I could make a whole subwoofer with that money.
 
diyAudio Senior Member
Joined 2002
MORE OF IT

Hi,

Not that I want to discourage you from running experiments but IME it does not yield good results in a X-over.

Also I'm surprised at the cost ratio between the two of 1:10 which may hold true for a 200 mF but for smaller values?

Re:
Any ideas on this one?

In a X-over these diodes aren't going to do anything useful since no DC is present.

Cheers,;)
 
In a X-over these diodes aren't going to do anything useful since no DC is present.

I use them here as half wave rectifiers... diodes are not only used where dc is present

Also I'm surprised at the cost ratio between the two of 1:10 which may hold true for a 200 mF but for smaller values?

basicly the ratio is about the same,
On a 10 uF one, wich is aorund 3$ for a NP, its aorund 5 cents for a polarized one..

Btw,
which may hold true for a 200 mF but for smaller values?
how low would you make a cut with 200mF?
200mF would make a nice 10hz highpass filter ;)
but ouch on the price of a NP 200 mF cap..
 
diyAudio Senior Member
Joined 2002
TRICKS OF THE TRADE.

Hi,

A bit tongue in cheek:

I use them here as half wave rectifiers... diodes are not only used where dc is present

Don't forget diodes aren't linear in behaviour.

how low would you make a cut with 200mF?

I look at what I need for a complete project and than try to figure out a way to buy n *n value and // those values where appropriate so I get a volume discount.

It may not be applicable to your project though.

Ciao,;)
 
Ilianh said:
well the thing is a 200 uF NP cap is around 10$, a polarized 400uF one is around 50 cents (even less) here.

for a 5 speakers system where you use second or third order Xo's... that might save you a nice 80$ or so, wich btw is alot of money...
considering I could make a whole subwoofer with that money.

Hi,

There seems to be a huge and quite disproportionate difference in the price of these different types of caps, in your country.

Especially (as Dhaen & Frank have suggested) as the non-polar caps are (mostly) simply two polarised ones in the same can.

I say mostly, since I am not sure without looking it up whether this is precisely how the very special non-polar Black Gates are made.

I can get 200uf - 50 Volt non-polars for £1.10 each (considerably less than $2 US) in the UK, especially designed for xovers, but I would never use electrolytics in a speaker anyway.

If you wish to end up with a speaker that sounds even half decent, I would steer clear of any polarised electrolytics which merely cost a matter of a few cents, as they are most unlikely to give you a good sound, unless they are being sold at very much below cost!

You have been warned!!!!

Regards,
 
The one and only
Joined 2001
Paid Member
Re: Re: Non polarized Caps

Bobken said:
Almost all, if not all, polarised caps appear to perform their function better when they have a distinct 'bias', or DC voltage differential across their two leads.

If you have connected the caps two positive leads together, you need to apply a (relatively) positive DC voltage to the junction, and similarly, If you have connected the two negatives of the caps together, apply a negative DC voltage here.

I don't necessarily agree with the thesis that bias improves
electrolytics, but for what it's worth, you can achieve a bias
by placing diodes in parallel with each cap, creating a bias
rather quickly when the amplifier drives the back-to-back
caps. You perhaps would argue that this might create some
distortion while the caps are biasing up, but playing a little
head banger music while warming your system up should
take care of that.

;)
 
cap bias

It seems that Nelson is right again ;) - in an article recently published in british Electronics World magazine (already mentioned elswhere) distorsion figures for electrolytic caps were much better without bias. High quality film caps exhibited same performance regardless of bias. Back to back polarized electrolytics were worse than non-polar lytics, and very good results were achieved with non-polar lytics in unusuall (for this type of cap) back to back configuration.
Anyway, I wouldn't use electrolytics of any type for xover.

Vuki
 
Non Polar caps

I find that Solen do carry a reasonable wide range of film caps up to about 300uf. Their prices are "fairly decent" in hifi terms costing about $20 to $30 for a 240uf 250V polyprop. The sonic improvement over electros I feel is quite justified.
I've also tried "motor starter" caps(by Plessey) which are non polar caps rated at 110V ac and found their sound quite good too!

ckt
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.