XO Capacitors in Parallel to Get the Value

skrol

This is a continuation of a discussion in another thread ( http://www.polkaudio.com/forums/showthread.php?t=73017&page=2 ) that I thought might be beneficial to the DIY community.

In my upgrading of the crossovers in my RTA11TLs, I found that I made an apparent mistake with my parallel capacitors to get an exact value. The needed value was 16uF but the closest I could find was a 15uF. Knowing that capacitance in parallel is additive, I used the 15uF and a 1uF in parallel. Later I was told that that this was not a good idea and that I should either stick with the 15uF or use two 8uF in parallel. I recently removed the 1uF and there was an obvious improvement in sound.


I can't deny that the unequal, parallel capacitor configuration in series with my tweeter was detrimental to sound quality. When I removed the 1uF from one speaker, there seemed to be a phase error as I moved between the speakers. When the 1uF was removed from both, the sound was more natural in the upper mid-range (male voices) and open. My question is, Why?

PSpice analysis shows no difference but it may be the generic capacitor models. All the text books and on-line sources I have found so far say that this is a legitimate approach regardless of if they are of equal value or not. Even Polk used a 3.4uF Mylar and a 12uF electrolytic in parallel to make the 16uF (cost saving measure?).

One response was:

F1nut wrote: »

A smaller cap charges and discharges faster than a larger cap. In essence, the 1uF is acting as a bypass cap.

This may be true but I am thick in the head sometimes and am trying to understand how.

Help me think this through. The voltage across elements in a parallel circuit are equal. During the discharge, if the 1uF starts to discharge faster than the 15uF, wouldn't the 15uF immediately prop up the voltage of the 1uF and wouldn't they have to discharge at equal rates? After all, they are in parallel and their voltages must be equal.

Current is another story and may be the root of it all. The current through each parallel element can be, and most likely is, different. The ESR is very likely different for the 1uF vs the 15uF. I believe this is the theory behind the use of "Bypass" capacitors in crossovers. This is fitting with F1nut's answer. I am surprised that it makes a difference with these values at audio frequencies.

Other thoughts?
Stan

Joe08867

I found this on the web. I was looking for a way to explain it but this was better.

Here is the answer:
Caution: If the capacitors are different, the voltage will divide itself such that smaller capacitors hog more of the voltage! This is because they all get the same charging current, and voltage is inversely proportional to capacitance.

http://www.coilgun.info/theorycapacitors/capacitors2.htm

skrol

Thanks, though what you say is true for series connected capacitors. In parallel, the voltage has to be equal but the current would be different. I'll have to look into it more but I have a feeling that it is current phase related.
Stan

Joe08867

OK, let me go at this in a different way.

Think of the caps as cans of water. The larger cap holds more water and takes longer to empty into a given drain. The smaller cap holds less so takes less time. Which means the smaller cap is loading and unloading at a different rate than the large cap.

So with two caps of different volume you will have a caps that will react differently to a given frequency.

skrol

I see what you are saying but in this case, the analogy doesn't really hold. You are assuming that one bucket can not fill the other as it would with parallel capacitors.

This is one of the most fundamental rules in electronics. Parallel elements in a circuit will have the same voltage across them. All of the (+) terminals are electrically common and all the (-) terminals are electrically common. One (capacitor A) can not become more empty.

Hypothetically, if capacitor A could be at a lower level, the level would be equalized by the other (capacitor . When it does this, the current through capacitor B would momentarily increase (emptying at a faster rate) until the levels are equal.


More theories:
ESR: I have not seen ESR specs for the Claritycap capacitors but it may be that it is quite different for the 15uF vs the 1uF. If this is the case then we are no longer looking at simple parallel components but parallel series networks. The voltage across the caps would still be the same but there could be and even greater difference in the current flows.

Impedance: The reactactive component of the impedance of each capacitors is different at different frequencies. The impedance is not the same as looking at 2 static value resistors in parallel. It is two parallel components whose impedance varies independently form each other with frequency.

Resistors:
1/((1/R1)+(1/R2))

Capacitors:
Xc1 = 1/(2*pi*f*c1)
Xc2 = 1/(2*pi*f*c2)
The Xc1 and Xc2 will vary differently as the frequency changes.

Xc = 1/((1/Xc1)+(1/Xc2))

I'll have to ponder the impedance a bit more in regards to how the total Xc changes for the parallel caps vs the Xc of a single 16uF cap.
Stan

skrol

According to PSpice analysis, when comparing the 15uF+1uF parallel caps vs the single 16uF cap, there is no difference in the voltage phase and magnitude across or the current phase and magnitude through the tweeter. I'll ponder and crunch more on this.
Stan