In a speaker crossover, is there any practical or sonic penalty that arises when two film capacitors are placed in parallel to achieve a required capacitance value?
I understand that in power supply decoupling there is concern. The effects of parallel values have to be the measured to verify the resonance of the capacitors (at RF frequencies) does not raise the impedance. It’s unclear if this has any real world effect in a speaker crossover at audio frequencies.
I understand that in power supply decoupling there is concern. The effects of parallel values have to be the measured to verify the resonance of the capacitors (at RF frequencies) does not raise the impedance. It’s unclear if this has any real world effect in a speaker crossover at audio frequencies.
When capacitors are paralleled, the lead resistance and inductance of the capacitors are in parallel.
I read that this has a sonic advantage in that it results in a capacitor with better high frequency performance.
I read that this has a sonic advantage in that it results in a capacitor with better high frequency performance.
A crossover will not see action outside the audio band which means reactive parasitics are of little concern. Even if they were it would only affect the response.
In principle, yes there would be improvement. Sonically, the effect would be microscopic (read: inaudible).
Great.
The scare articles in the audio circles that stated different values shouldn’t be paralleled in audio signal coupling had me questioning things.
The required value is about 19uF, so 18+1 seems pragmatic and leaves the option to increase the capacitance if needed.
The scare articles in the audio circles that stated different values shouldn’t be paralleled in audio signal coupling had me questioning things.
The required value is about 19uF, so 18+1 seems pragmatic and leaves the option to increase the capacitance if needed.
In some pieces of properly functioning equipment it can be a foolish thing to do and you may regret trying.
Parallel film and electrolytic can be a problem as the high value electrolytic has a fairly high series resistance, so at high frequencies the low ESR of the film type cuts in and makes a small step in the frequency response.
Peter Snell used massively parallel electrolytics in his crossovers to overcome their high ESR.
But that was "back in the day" before better caps were widely available.
But that was "back in the day" before better caps were widely available.
Hi!
Stacked film capacitors are made of several "capacitors" in parallel. Even the rolled ones, you can think that you have a continuos parallel plates, some inner some outter (with different diameters).
If a 10uF stacked film cap has 10 layers, the first 9 layers paralleled can be considered 9uF and the remaining layer 1uF.
So you have 9+1 and it's fine. In fact, in this example, you have 10 x 1uF capacitors in parallel.
And as already mentioned, if paralleled capacitors are in separated packages, you have the advantage of reducing terminals inductance, which for passive crossover doesn't matter much since (low 20kHz frequency and low 4 to 8ohm impedance load).
Does it make sense?
There are a lot of different opinions on this in regards to crossover caps. Many people love a "bypass" capacitor where they stack something like a .01uf cap on a larger value cap and they swear it makes the "soundstage" better and the "top end more airy". I generally think this is non-sense and haven't seen any real science on it. I have tried it and certainly cannot hear a difference. I'm sure the proponents of it would say I don't have trained ears and I would tell them that their confirmation bias is tricking them into "hearing" a difference.
I have seen some science on the ESR in caps being an issue, but I think most modern caps made for passive crossovers have overcome this. I think that as long as you are using proper crossover capacitors then it's totally fine to stack them and get to the desired value. I have done it and I have never seen any problems in measurements (distortion or phase) by using multiple stacked caps. I can tell you that using the correct value of capacitance is more important than if you stack caps. For example, if you use an 18uf cap where you need 19uf then you will see a difference in measurements. But I don't think you would see any difference in measurements if you used one 19uf cap vs. an 18uf and 1uf stacked in parallel.
Having the proper value and properly designed speaker with flat frequency response and drivers playing within their distortion limits will make a much larger difference than if you stack caps.
Appreciate the warning, I understand it. It can negate their ability to function as intended and can yield a dangerous situation in decoupling positions.
What we’re investigating is a 50-year old passive crossover having several paper in wax or oil capacitors that otherwise measure quite poorly in several areas, including DA and ESR of a few Ohms. I am hoping to restore original function and linearise the speaker’s transfer function by a few decibels
Very clearly described. I have some general familiarity with both construction techniques, but not at the academic physics level. That is the major reason why I need to ask here before proceeding.
I appreciate that. I was trying to give perspective to the source of the 'scare articles'. As stated earlier I don't expect these to apply to a crossover.