Esoteric PLL crossover idea

I've made some pretty sophisticated PLL filters by sacrificing insertion loss... so if you're not afraid to throw away 10-20dB. As it happens the filter shown above is only down 1dB.

1k to 50k ohms gives you a couple of orders of magnitude to work with.
Allen,

What will be your (real-world) source for that circuit?? Something with 1 milliohm output resistance?? 🙂

Dave.
 
So far I've only found the result of my calculations, rather than the calculations themselves. This circuit was for someone who wanted to mix a headphone signal to mono and to have an approximately second-order Linkwitz-Riley crossover while only using resistors and capacitors.

mr18-001.jpg

See also https://www.diyaudio.com/community/...db-line-level-calculation.320216/post-5375368
 
Last edited:
I found the actual calculation in the attic.

What it boils down to, is that when you cascade two first-order RC low-pass filters with equal time constants, the first with a resistance R1 and the second with a resistance R2, you get a second-order roll-off with a quality factor

Q = 1/(2 + R1/R2)

while ideally, you would need Q = 1/2 for a Linkwitz-Riley filter.

For a given ratio R1/R2, you can improve the quality factor a little bit by making

R1C1 = sqrt(1 + R1/R2) * (1/(2 pi fc))

R2C2 = (1/sqrt(1 + R1/R2)) * (1/(2 pi fc))

where fc is the cross-over frequency. The same holds for the high pass.

The resulting Q is

Q = 1/(2 sqrt(1 + R1/R2))

which is only marginally better than the value with equal time constants when R1 << R2.

I intended to also look at the flatness of the sum of the high-pass and low-pass output signals when you allow unequal values of fc, but didn't manage to get a result out of that.

ja14-003.jpg