Fun measuring crossovers....

[sfogg]

Al,

"Q is always the limiting factor. Q as I think of it is "Quality factor" of individual components."

Different type of Q I think. There are no individual components in this filter... it is done all digital.

In parametric EQs 'Q' refers to the width of the filter... sort of like

the bandwidth that it covers. The EQs in this box are limited to a Q of

10.. which is pretty narrow but not excessively so. A higher Q would

let the rolloff be sharper still.

Shawn

[Al Klappenberger]

Shawn,

That Q specification is probably still the same thing. In an L-C resonant circuit you can measure the Q by measuring the bandwidth of the peak it generates. I think it's the same thing. I just think of it as X/R becasue that's the way my analysis program handles it.

Al K.

[sfogg]

Had a little time to tweak the response of the initial filter. This is

still done using 3 parametric EQs and a LR 8th order crossover. I

basically moved the LR crossover in a little closer to the target

crossover point to get this response and tweaked the parametric a

little to smooth out the blend more. This one has about 5dB more

attenuation on the initial slope and the arc drops by about the same

amount too. There is still a touch of ripple in the blend between the

two crossovers but it is pretty small.

[sfogg]

High pass

[sfogg]

Low pass...

[Al Klappenberger]

Shawn,

That response looks very good now. You also have 30 dB ultimate rejection plus the sharper skirt.

On think I noticed about showing the sum of the two channels is that the SD analyzer do not include phase in the A+B sumation. I found that I had to use my oscilloscope on A+B and look at the output of the Y axis to actually see the sum. That way you can actually see that reversing one channel will casue a big dip. The SD375 wouldn't show it. BTW: The channels of the ES600 network are 180 deg out of phase at the crossover.

Al K.

[sfogg]

Al,

I'll try checking phase too but I think the delay (DSP latency) through

the crossover is throwing off the phase measurement in the SD380 in

transfer mode. That is one extra thing I can try with this crossover...

time alignming the drivers to see if it makes an audible difference

with the steep slope crossover.

In my sum mode I'm just tying the two outputs together and input it into channel B on the SD380.

Shawn

[Al Klappenberger]

Shawn,

I think reversing the polarity of one channel will just casue the sum to change to a big dip. I don't beleive you will hear a difference. The ear is deaf to that sort of thing.

I don't advise just connecting the to channels together to test the sum. You need to terminate each channel into the proper impedance and sum up the voltage across each load. I found that the SD analyzer doesn't account for the phase so you need to do the suming using the A+B iuputs of a dual trace oscilloacope. Look at the vetrtical (Y) axis output of the scope with the SD380. You will see a big notch at the crossover until you switch the invert switch on one of the scope input channels.

Al K.

[PrestonTom]

Re: "Q" or quality factor

I am confident that both Al and Shawn know the definition, but I did not want others to get confused.

The "Q" is a relative measure of the bandwidth of a filter (or other devices) . The measure refers to the ratio of the center frequency divided by the bandwidth. Where bandwidth is defined by the half power points (3dB down points) . So if the device lets frequencies 10% (above and below) around either 100 Hz or a 1000 Hz (or bands of 20 Hz and 200 Hz) then the Q in both cases is the same: 5. So it is a measure of bandwidth relative to the center frequency. The bigger the number, the sharper the tuning.

Good luck,

-Tom