Using cheap polar electrolytics in Xovers

[Puggie]

I know I need non-polar caps for a Xover but this is just for a parallel notch filter and needs to be 860uF (or there abouts).

Can I get away with a pair of double value polar electrolytics wired in series and opposite phase? IIRC this is how a non polar electrolytic is constructed.

Does it mater if I join the +ves or -Ves together,

Sound quality is not all that important as this is a parallel filter to remove the resonant impedance peak on a small wide range driver, so I can make the high pass filter behave properly. I have stacks of 1000 and 2200 polar caps so I'm hoping to parallel series 4X 1000uF caps of just a pair of 2200uF in series to make a 1000 non polar for this filter, its a bit high but in the right ball park.

I know its not ideal but is it possible? the speaker will only receive about 15Watts.

Thanks.

[singa]

2X2200uF will give 1100uF in series and must be connected back to back.I normally choose the negatives as common.Will 1100uF be to far from 860uF?

[Puggie]

Hi

thanks for the response, yes, 2200s will give me 1100uF, I'm hoping it will be close enough, I'll see tonight.

I have some smaller value caps as well, do they need to be matched pairs? could I use say a 2200uF and 1000uF in series. would that work ok with them back to back and negatives connected? I see no reason why the total capacitance would not be the same as series connecting any capacitors normally (i.e not back to back).

[singa]

Connecting same value caps is easier to calculate by dividing the value by the number of caps.That is 2200 divide by 2.If caps are diferent value the calculation is a bit different, becoming 1 divided by capacitance +1/capacitace of different value. If you have a capacitance meter it's straight forward. The caps need to be connected back to back or else it will not function as non-polar.

[Elvee]

Hi Puggie,

Although I don't belong to the "golden ears" brigade, I 'd still advise you strongly against using electrolytics in such a way:

First there is the accuracy problem. A typical tolerance spec for an electrolytic looks like +50%/-10%, and it can be much worse.
You might think of picking the right value by measurement/selection, but it won't work well: an unbiased capacitor will see its capacitance increase over time.

Then there is the loss factor problem: your notch will be nowhere as deep as expected, and may even be completely absent.

And finally, there is the linearity problem: under large signal condition (ie capacitor subjected to significant reactive power), such a cap is non linear and will generate distorsion. Even if it isn't directly in the audio path, its harmonic currents will find their way into the load.

I'd recommend you find another way of solving your problem.

Cheers, LV

[kevinkr]

I'll second Elvee's comments, they jibe with my experience even with smaller back to back polarized el caps. Might be ok to test concept for the very short term if the esr and esl components in the caps reactance are not excessive.

[lndm]

Originally posted by Puggie
...as this is a parallel filter to remove the resonant impedance peak...

Is this three components in series, and the string placed in parallel with the driver? If so, I wonder whether the resistance wouldn't swamp the caps flaws.

[Puggie]

Originally posted by lndm


Is this three components in series, and the string placed in parallel with the driver? If so, I wonder whether the resistance wouldn't swamp the caps flaws.

It is, I'm looking at 866uF, 4Ohm and 4.3mH all in series then that whole package in parallel with the driver. this is to remove the resonant peak due to Fs at about 90Hz.

I don't really understand how this would be any worse than using an off the shelf non-polar electrolytic which is my next solution.

[lndm]

In this case, I feel that the series resistance and inductance of the cap would combine with the series resistor and inductor and make a somewhat insignificant difference. Given the large value of capacitance needed in this sort of application, I have used electrolytics here myself. YMMV.

There are not many locations in a crossover where I would have said this about electrolytics. The comments above by others is usually how I feel.

[Puggie]

well I generally do not touch eletrolytics in a capacitor unless its over 250uF, but in this case it is significantly higher, its a parallel circuit and the drivers are cheap so I'm inclined to try a cheaper option. Hence me asking about electros.