Crossover Questions

[Lo-E]

So, I'm rolling my own crossover for a 15/6 project I'm working on and I have a few questions before I continue to draw this thing up.

This cab is not a nEarful. It's an old EV B115M. The cab will be re-tuned for these drivers. I'm going with a 2nd order LP and 3rd HP. The drivers in question are the Faital Pro 15PR400 and the Alphalite-6A-CBMR, but my questions are general enough that the answers should apply to just about any bass cab.

1: Should I do my calculations using the nominal 8 Ohm impedances, or the Theile-Small impedances (Re)? This is the most critical question I have.

2: In what instances, if any, would a Zobel network be worth the trouble in a bass cabinet?

3A: How critical is it to use a crossover frequency more than one octave above the mid driver's resonant frequency (fs)?

3B: If I cross over at a lower frequency and I do end up with a spike at the 6's fs, how practical is it to add a notch filter?

4: Am I just overthinking the whole durn thing?



Thanks one and all-
E

[Rockmusician]

Originally Posted by Lo-E

4: Am I just overthinking the whole durn thing?


Thanks one and all-
E

Yes. The vast majority of people will not even notice a difference. You might not either. Keep it simple. It's a lot less labor intensive and less stressful.

[Lo-E]

You're probably right, but I could still use an answer to #1. All the math sorta hinges on that. Even the simple math.

Thanks-
E

[billfitzmaurice]

Find your answer here:
The Speaker Building Bible - Techtalk Speaker Building, Audio, Video, and Electronics Customer Discussion Forum From Parts-Express.com

[Lo-E]

Outstanding!
Thanks, Bill!

[DukeLeJeune]

Originally Posted by Lo-E

1: Should I do my calculations using the nominal 8 Ohm impedances, or the Theile-Small impedances (Re)?

A better approximation is to look at the impedance curve and use the impedance at the target crossover frequency.

2: In what instances, if any, would a Zobel network be worth the trouble in a bass cabinet?

If you don't have the ability to make your own measurements and optimize your crossover accordingly, I believe that using a Zobel will give you more predictable results. Don't skimp on the power handling of the resistor in the Zobel.

3A: How critical is it to use a crossover frequency more than one octave above the mid driver's resonant frequency (fs)?

If the midrange driver has a major impedance peak at its resonant frequency, and has modest excursion capability, you run the risk of driving it into overexcursion and distortion and/or damage if the crossover doesn't adequately address that impedance peak. The reason is, a textbook filter will not perform as predicted into that peak, and the response will not be suppressed nearly as much as you would have hoped. I'm not saying this will happen, but it could, depending on the specifics. There is no easy solution just from changing parts values in your crossover; you either need to use a dedicated resonant peak filter (see Vance Dickason's "The Loudspeaker Design Cookbook"), or have some sophisticated measurement and modelling software to cook up your own filter. Having the crossover for the midrange more than an octave above its in-box resonant frequency helps, and may be sufficient; one octave isn't a magic number, but a point along a continuum, and probably a good ballpark rule of thumb if you don't have measurement and modelling equipment.

3B: If I cross over at a lower frequency and I do end up with a spike at the 6's fs, how practical is it to add a notch filter?

I've never tried addressing it with a notch filter, but that could work. However itt would be a challenge getting all the component values right without measurement equipment, as the notch filter will change the impedance that the crossover filter sees. The resonant peak filter smoothes the impedance peak so that the crossover's behavior is more textbook-like, so imo this is the more elegant approach.

4: Am I just overthinking the whole durn thing?

Imo the more you are aware of and take into account to the best of your ability, the better your results will be.

[exmaxima1]

Originally Posted by Lo-E

3B: If I cross over at a lower frequency and I do end up with a spike at the 6's fs, how practical is it to add a notch filter?

If you feel compelled to HP the mid driver close to its Fs, you should consider using a conjugate filter. It is similar to a Zobel, but instead of normalizing the rise in impedance at the upper freqs, it does it thru the impedance peak at resonance. It is essentially a Zobel with an additional inductor in series (so L, R, & C in series, wired in parallel to the mid driver)

This was a fairly common circuit years ago when passive crossovers were used used on subwoofers. It will force the HP network to work as it was intended.

[DukeLeJeune]

Originally Posted by exmaxima1

If you feel compelled to HP the mid driver close to its Fs, you should consider using a conjugate filter. It is similar to a Zobel, but instead of normalizing the rise in impedance at the upper freqs, it does it thru the impedance peak at resonance. It is essentially a Zobel with an additional inductor in series (so L, R, & C in series, wired in parallel to the mid driver)

This was a fairly common circuit years ago when passive crossovers were used used on subwoofers. It will force the HP network to work as it was intended.

"Conjugate filter" is probably the more correct term for what I was calling a "resonant peak filter".

[Lo-E]

Thanks Duke & Exmaxima1,

I did end up designing for the Z at the crossover f for the 6".

As it turns out, the z of the Faital 15" doesn't really rise considerably until well past the crossover frequency, so I'm going to forego the Zobel, at least for now.

As far as the octave rule goes, I'm crossing over at a little below 900Hz, which is a little on the close side of the 6"s 501Hz fs, and it's peak looks pretty substantial on paper.
I think I'm going to try it without any compensation and see what it sounds like. If it's really spiky around 500Hz, I'll look into the whole conjugate filter shebang. Right now the low parts count seems really appealing to me, and I'd hate to fix it before I even know if it's broken.

I was kinda kidding about overthinking the whole thing. On the one hand, I do believe that sometimes good enough really is good enough, but on the other, I've never felt that having too much knowledge was a detriment. This whole project has been an interesting learning experience. I just hope I end up with a good-sounding cabinet to show for it!

E

[billfitzmaurice]

Originally Posted by Lo-E

I'm going to forego the Zobel

E

A Zobel is a waste of parts. That comes not from me, but from Joe D'Appolito, the recognized #1 authority where crossovers are concerned. It's useful if you can't find out the Z at the crossover frequency, but with tools like WT3 there's no excuse for that for the serious builder.

As far as the Fs of the mid driver issue is concerned it's rather moot. You base the crossover point on where the woofer drops to no more than -6dB at 30 degrees off-axis, then choose the midrange driver based on its ability to run to that frequency without exceeding xmax. With tweeters it's difficult to predict excursion, so using 2xFs is a rule of thumb. OTOH Neville Theile (yes, that Theile) prefers 5th order HP that allows him to run his tweeters to just above Fs. Predicting excursion with a cone midrange, and how steep a HP slope is required for excursion/Pe protection, is a rather simple matter.

[Lo-E]

Based on this, it sounds like I'm already on a fairly safe path.

Thanks again-
E

[DukeLeJeune]

Originally Posted by billfitzmaurice

A Zobel is a waste of parts. That comes not from me, but from Joe D'Appolito, the recognized #1 authority where crossovers are concerned.

Depending on the specific situation, including what priority one places on smoothness of the system impedance curve, I might well disagree with Joe.

[Lo-E]

In my particular instance, the impedance of the 15 stays pretty darn stable throughout the f range that I'm using it in, so in my case it probably is a waste of parts.

As for Zobel's usefulness in other applications, I'll leave that to more experienced engineers than myself. ....and that would be just about anybody!

[Lo-E]

The crossover is done and is a success!

After much head-scratching and help from this forum, I ordered a bunch of parts and rolled a nice, beefy crossover for my old EV cabinet.

I ended up with a 3rd-order HP, a 2nd-order LP with no attenuation for the HP and a crossover f of 840Hz.

The end result? Fantastic! It sounds really, really smooth. Much better than the original EV x-over and I couldn't be more pleased.

Thanks everybody who helped me out here - especially Bill and Duke - with advice and information. It would have been a much harder task without you guys!

The rest of the project is documented here for those curious:
Speaker ideas for old EV B115-M cab (15" and 6")...

Best regards-
E