Turn a 2-way into a 3-way?

[audiobomber]

I have an old pair of Energy Reference Connoisseur speakers. They're originally a full-range 2-way floorstander using proprietary drivers (1-1/4" mdf cabs, no parallel sides, slanted baffles, 80 lbs each).

I replaced the passive crossovers (1.5 kHz, 2nd-order on woofer, 3rd-order tweeter) with a Marchand XM44-3 crossed over at 2K. I was planning on crossing over from the mid-woofer to my UFW-10 subs @ 80 Hz, but I found I got better results by blocking the ports on the mains and running the mains full-range with one sub filling in the bottom (I use the second sub LFE only).

Now to my question, if anyone's bothered to read this far.

I'm using a 3dB notch filter on the tweeter at 4.5K and it is now very smooth. But I have a 3-6 dB hump from 630 to 1.6 kHz. The response at 500 and at the crossover point, 2K, are dead on. Should I try another notch filter, or does this look like a baffle step issue? It doesn't seem like a BSC filter is what's needed because it's not a gradually rising response, it's more like a plateau.

As an alternative, I'm considering moving from a steep-slope 2-way crossover to a first-order three-way, by putting a mid in between the original drivers. I thought a Jordan JX53 would be an interesting choice, crossed over at 500 and 5K.

Any thoughts?

[Andersonix]

Without addressing your hump between 630 and 1.6k, my first impression is that if your tweeter is very smooth above the new crossover point (2k), why mess that up and introduce all sorts of other problems by adding a midrange driver?
Another thought (somewhat off-topic) is that in theory and in practice, blocking off the ports on the main speakers will cause the cone excursion for those woofers to increase, as the driver is not loaded as it was intended (seems counter-intuitive, right?). But this would really only become a problem at high levels. I've ocasionally blocked off the ports on my "mids" to cut some bass output from them, but they seem to sound better overall when left ported as designed. If you could determine where these mid-woofers roll off (either ported or sealed), you could add a low pass crossover to the subs at around that frequency.
I think you are moving in the right direction by going with an active cross-over and keeping the system as simple as possible.
Perhaps you can measure the response of the mid-bass drivers by themselves to see if the hump is inherent to them? I guess you can't adjust the active crossover so that it crosses over at the original 1.5k? It seems likely that the original passive crossover accounted for this hump in the mid-bass response.

[audiobomber]

Originally posted by Tosh
if your tweeter is very smooth above the new crossover point (2k), why mess that up and introduce all sorts of other problems by adding a midrange driver?

I know what you mean. Address one problem, cause two more. My main thought in adding a mid-range is to be able to move to first-order slopes. Less stuff in the signal path.

in theory and in practice, blocking off the ports on the main speakers will cause the cone excursion for those woofers to increase,

You wanna see cone excursion, you should see when I play an LP without blocking the ports. It's scary! These little woofers seem to be designed for high power and a lot of movement, so I don't think it will be a problem. I've played them very loud a few times lately.

If you could determine where these mid-woofers roll off (either ported or sealed), you could add a low pass crossover to the subs at around that frequency.

The problem is, the main speakers go all the way down in my room (flat at 25 and 40 Hz, +6 @ 30 Hz, -6 @ 20 Hz). So they cover the same range as the sub. In fact they measure virtually the same with or without the sub. With the mains vented I can't get good coherence. The sub & mains sound too different. The main benefit the subs bring is to beef up the mid bass. My room has a trough at 50 and 63 Hz.

I tried rolling off the mains in the Marchand (4th-order @ 80 Hz). I didn't like the sound, the subs were too different to blend in. It may be possible at some other frequency, but it's kind of expensive to experiment. Besides I love the way the bass sounds right now.

Perhaps you can measure the response of the mid-bass drivers by themselves to see if the hump is inherent to them?

I could do that, but would it help to know whether the hump comes from the driver or the driver/cabinet together? It may be some of both, a resonance and a baffle-step response together. I could just order another filter from Marchand and be done with it, but I would like to minimize the numbr of crossover parts.

I guess you can't adjust the active crossover so that it crosses over at the original 1.5k?

Actually I can. I had the filters on order and then cancelled them until I decide where I want to go. I ordered a copy of The Loudspeaker Design Cookbook to learn more about speakers.

Thanks for the input Tosh.

Dan