I respect your apparent authority here. Are the tolerances cumulative if one uses caps combined to equal a given value? I would presume they would be, but then again I am asking you to be certain.
I am neither an authority nor an engineer, so I consider my contribution here to be provisional. In other words, if one of AK's authoritative professionals or consummate hobbyists contradicts what I say with science at hand, and because this is a 'sticky', I will either edit or delete things unless otherwise directed by the moderator.:ntwrthy:
Tolerances will be arithmetic in a configuration of parallel capacitors. In other words, they will be both additive and subtractive. Electrolytic capacitors can be + 50% to minus 20% of the nominal value. Other types such as polypropylene, polyester, Mylar range from 1% to 10%. Remember that on top of this, electrolytic types are subject to much greater drift due to ageing and instability than film types.
Now, the following is my opinion culled from more hours spent trawling the knowledge of better minds and education than WAF may consider sane or healthy (slow learner?:tears
: if one is going to parallel electrolytic caps, the best way to ensure success would be to use a capacitor tester to match specimens before installation. (I don’t own one myself). That being said, my approach would be to source a single cap of the specified/original value and refrain from using multiple caps in parallel.
As for film caps: failure to match caps may or may not have negative consequences. For one thing, specified tolerances are (apparently) conservative. And, variances may cancel each other. Furthermore, any residual variation from nominal value may not be detractive to performance in relative terms. There are just so many other factors involved. Obviously we want to be as precise as is practical, but WHERE the determinative limits lie is unpredictable for anything but a specific case, and even then problematic. If you use film caps at 1% (coded as F), 2% (G) or perhaps even (J) 5% you should at least have peace of mind. If you end up putting two or three film caps in parallel, and if you find evidence to support the prediction that a 5% variation from nominal spec will produce a 50 Hz shift, I'd look into the situation.
As for my projects, I just bunged-in what I had stocked for the purpose and then worried more about lead-length, good soldering and heat dissipation.
Your thoughts on using a "bypass" capacitor as JBL uses it. IE like a .15 uf cap or so used in the upper freq. part of the xover to aquire a better transient response.
IMHO, if the original design uses a bypass on the same type of capacitor as you will be using as a replacement, then use one. I have deleted bypass caps after substituting the original electrolytic primary cap with a film cap. In general, bypassing gets into a controversy of galactic proportion, and the science that might apply to specific instances becomes, at least for me, opaque. In the case of bypassing, you are free to experiment. The values are small, so parts will be inexpensive. And if you use heat clips on any leads to protect your primary cap, you can solder and snip with immunity.
