Again, again, again...design help (ACD)

[lorenmjones]

I've been continuing to use my Peerless 831735 / Seas H1212 MTM as a learning experience for me regarding the rudiments of crossover design. I initially was doing an iterative method using my DSP and putting in filters, remeasure, tweak etc. I got kind of OK results doing that but mostly by brute force use of parametrics to force a flat on axis response. Imaging is poor with the several designs I've tried that way.

I have now put the data in ACD and started to do some design efforts there and I'm hoping someone can look at my ACD files and make suggestions. As I did with the iterative method I can add parametric filters until the response is flat but I'm looking for insight from people on how to decide which type of filters (symmetric or asymmetric electrical slopes) at which frequencies will give a good result without relying too much on extra EQ.

I have tried lots of different options. Below are graphs of what I have come up with in ACD.

Here is my mag and phase plots from ACD using a 1900 hz LR4, tweeter wired reverse polarity, 3 parametrics on the tweeter to get it's response even close to flat. Broad cut at 7k (-4db with Q of 1 iirc), Cut at 2100 of -5 db with a Q of 2 or 3, and a small narrow boost at 2800. The woofer just has a notch at 3700 for the breakup and there is a little boost at 390 hz for a dip there. The BSC is ballpark, ignore everything below 200 hz due to gating.

Is the design shown below reasonable or should I be able to come up with something more elegant. I'd be happy to email my ACD excel files to anyone that wants to take a peek at them.

Thanks much to all

[lorenmjones]

Re: Again, again, again...design help (ACD)

These are the raw responses. The tweeter seems to have some significant diffraction problems that are what I am finding I need to add all the parametric EQ to. Any ideas? Also, in basic terms what am I looking at on the phase plots to determine phase tracking through the crossover?

[[email protected]]

Re: Again, again, again...design help (ACD)

You want the phase to track together as much as possible above a below the crossover point. This will most certainly effect the imaging. What is ACD?

[lorenmjones]

Re: Again, again, again...design help (ACD)

It's like PCD but for active crossovers. It's at Charlie Laub's site. So if my crossover is at 1900 hz I'd like the tweet and woofer phase both to be at the same place (say arbitrarily 30 deg or whatever) AND have the slope of the phase for the two drivers be similar through that region?

You want the phase to track together as much as possible above a below the crossover point. This will most certainly effect the imaging. What is ACD?

[charlielaub]

Re: Again, again, again...design help (ACD)

It's like PCD but for active crossovers. It's at Charlie Laub's site. So if my crossover is at 1900 hz I'd like the tweet and woofer phase both to be at the same place (say arbitrarily 30 deg or whatever) AND have the slope of the phase for the two drivers be similar through that region?

I suggest that you try using the TWO-DRIVER PHASE TRACKING extension. This is very helpful in identifying phase tracking between drivers. It plots the relative phase angle versus frequency for any two drivers (See below). Follow the instructions and then post the results here.

I will follow this thread and help you.

-Charlie

[lorenmjones]

Re: Again, again, again...design help (ACD)

Thanks Charlie,

I will check that out once I get home and post results.

I suggest that you try using the TWO-DRIVER PHASE TRACKING extension. This is very helpful in identifying phase tracking between drivers. It plots the relative phase angle versus frequency for any two drivers (See below). Follow the instructions and then post the results here.

I will follow this thread and help you.

-Charlie

[ATTACH=CONFIG]46408[/ATTACH]

[lorenmjones]

Re: Again, again, again...design help (ACD)

OK so here is the phase information from the extension. At crossover they aren't tracking too badly (I guess), but I don't really know what bad really is. They get way out of phase less than an octave away however. How bad does this look?