I don't see how a 190pf cap could materially affect the SQ of the crossover. I've been designing and building for years and my experience has been that generally you need a cap of significant percentage to improve the SQ. I have tried using a 0.22uf cap in parallel with a 7 uf cap and didn't really notice. Some publications say it needs to approach 50% of the total value.
There are several reasons. Also those "rules of thumb" one often sees on forums are almost always inaccurate.
1. The venue might be, maybe one. No obvious echos.
2. The best, most accurate/natural sounding capacitors have been extinct for some years due to improper "testing" and "comparisons", individual and article capacitor comparisons alike, and then posting their results on audio forums. So what you have are 2nd rate caps these days.
3. Not the best designs, which includes values, ratings, biasing etc. A few equations won't design a
preamplifier or amplifier. It takes a whole lot more than that.
4. No one has developed or seems to use any specialized listening tests to determine if their resultant product is actually accurate across the audio frequency range. Claims of sounding good, "to die for"
really means little,,, and compared to what.
That 150pf bypass cap affects the ultrasonic frequencies, thus rise time (attack time). The ultra sonic response affects everything down to the bass in a top notch system when masking is minimal to non existent.
By the way, the crossover capacitor is over 530,000 times larger than the 150pf cap. That information tells us how sensitive my speakers are, the lack of masking is incredible. Another way of putting it, there is virtually no sonic signature.
There are several changes that we can correlate.
A. Rotating my test speakers by 0,5mm is perceptible, tonal balance, sound stage etc.
B. One millionth of an ohm resistance change with a variable speaker control, tonal balance perception, sound stage etc.
C. the number of 5 foot speaker wires in parallel per speaker leg. My last adjustment was 10 parallel wires in one leg and 9 parallel wires in the other leg.
D. ~100pf capacitor in parallel with ~100uf capacitor. I am using ~ values due to proprietary reasons.
Update, after some tweaking with an inductor (about 1" more wire was added to a turn). Will continue to test over the
entire audio band.
The result in my lab setting is a system with virtually no sonic signature, and total natural accuracy to the source.
The remaining limitation is the recording due to inferior electronics.
cheers
steve
