Feedback on two-way design

[DeZZar]

Hi All,

I've been working for a while now on a floor standing two-way design using the Dayton RS180P-4 woofer and the SB Acoustics SB29RDAC-C000-4 tweeter. The box is a 35 litre design with as slot port tuned to around 37hz resulting in a f3 ~34-35hz.

I've gotten to that point now where I think its ready to purchase the final crossover parts but wanted to reach out to the more experienced builders here on the forum for your thoughts/feedback on the results and crossover design.

Here is the crossover schematic:


Here are the measured results with this crossover in place. Showing the woofer and tweeter on their own + combined response.
You can ignore the strange little dip and peak in the responses show here from ~15K up. This showed up when I had to switch mics to a behringer and wasn't showing on my previous calibrated dayton mic.


Total response with woofer near field and port nearfield shown.
Again you can basically ignore the strange results from 15K.


Here's the result with inverse tweeter.


And finally the total impedance measurement with everything in place. Min of 3.8ohm around 150hz.


I welcome any thoughts and feedback on this design.

Dez.

[DeZZar]

Here are some photos of the recently finished boxes

[rpb]

Speakers look nice, but I can't see any of the measurements. Might be because I'm using a chrome book.

[Chris Roemer]

Nice pics, but I don't see any screenshots either, and I'm not (Firefox).

[Wolf]

Look nice with nice drivers, but I can't see them either.
Wolf

[DeZZar]

rpb Chris Roemer Wolf - sorry not sure what I did there. I've tried re-uploading the images. Let me know if you can see them now.
​​​​​​​

[rpb]

I can see the measurements now. I don't see anything wrong, but I like to try lots off x-overs. I almost never finish a design for that reason. I keep changing it, and listening to see what I think sounds best. After you have some time with this one, I'd try a few changes. Maybe a notch on the woofer at 5.5k. A cap across the coil for a tank type notch might work. The other change would be to increase the order of the woofer slope. That might be a 3rd order electrical on the woofer, or it might be a 4th. You might try 3 x-overs that measure different, but sound the same for 90% of the music you play. On the other hand, sometimes one stands out in some aspect.

What will be different with the "final xo parts"? Were you planning on higher quality parts?

Have you listened much to these speakers in your intended room as a pair?

[DeZZar]

I actually had a parallel 0.5uf + 10ohm resistor combo across the coil on the woofer which according to sims would push the woofer breakup way further down. In actual fact when I measured with and without the notch actually made that ~5K spike about 6db louder. So I removed it.

I agree though - have been changing and tweaking crossover and listening to these for the past few months now and I would probably continue - but I have another 3-way speaker project waiting so I'm leaning towards wrapping these up now and moving on. I thought about increasing order on the woofer or as you say a parallel notch - might try that as a final tweak to see if it makes a difference to the top.

[rpb]

The on axis response can be identical, but subjectively they may sound different with the notch, or steeper slope. On the other hand, they may sound the same!
It's fun exploring the possibilities.

[DeZZar]

What will be different with the "final xo parts"? Were you planning on higher quality parts?

Yeah basically just purchasing higher quality caps and resistors.

Have you listened much to these speakers in your intended room as a pair?

I've had them setup for a while now with a few different crossovers. Current setup is sounding the most balanced.

It's funny because with some prior versions they sounded really clear with some music which worked really well but on other tracks were bright and harsh sounding which would sort of overpower the bass output. They now sound more balanced and don't have that shrill sound but when I play the tracks where they sounded super clear before, they sound less so now. I guess sometimes good things are emphasised for the wrong reasons!

[LOUT]

Just for fun (practice?) I tried to see how close I could get to that FR using as few parts as possible, resulting in this:

I'm not suggesting anything is wrong with the original XO (though I agree that a notch or steeper slope on the woofer to tame its ~5K spike even farther down might be helpful if you think you hear some harshness near the highs).

[DeZZar]

Just for fun (practice?) I tried to see how close I could get to that FR using as few parts as possible

Yeah nice - and thanks for throwing a suggestion in ring . Its pretty close to some of my earlier attempts and I'm generally biased towards less parts as well. It would work well I think if the drivers were a little better behaved on this baffle.

The reason I went with 1.5mh and a notch on the woofer is that 2mh pulls the 300-500hz region down a little too far trying to compensate for the emphasised humps around the 1K region. 1.5mh+notch helps to flatten this out and we don't loose as much mid bass.

I would have liked just a 12db slope on the tweeter but because it needs to cross low with the woofer (~1500, the woofer simply as no usable output at all beyond 1500) it needs the higher order to both avoid pushing it too far and to avoid adding extra energy to the woofers 1k peaks. We also have to deal with some peaks in the tweeter at 2.5k followed by a wide dip 4-5k. The transfer function on he 3rd order actually tends to exaggerate this whereas the 4th order helps smooth it out (more rounded knee and peaks where the tweeter is weak). The Zobel is mainly addressing a tendency in the tweeter to rise in response (2-3db) from 9k.

[LOUT]

I was particularly curious about the 4th order on the tweeter instead of 3rd, but that makes sense now. Thanks for taking the time to explain all of that.

[Chris Roemer]

Just using mfr. data, I came up w/this -

2nd order HP w/"L": 15uF series cap, 0.40mH shunt coil (to gnd), L-pad: SR=3ohms / PR=3ohms, (reverse polarity indicated)
2nd order LP w/Zobel and notch: 1.5mH (low DCR) series coil, 25uF shunt cap, Zobel (to gnd) = 6ohm resistor + 30uF cap
the "notch" (tank) runs "across" (in parallel w/) the 1.5mH coil and consists of 4ohms + a (tiny) 0.47uF cap targeting woofer's 5k breakup

[Prana Ferox]

What's the veneer and finish? Looks nice.

[DeZZar]

What's the veneer and finish? Looks nice.

Thanks . The veneer is Tasmanian Blackwood and its been finished with a few coats of satin polyurethane.

[Geoff Millar]

They do look beautiful; we once bought a house built by a timber merchant, the rooms were lined with Tasmanian Blackwood panels: lovely timber.

Your finish looks marvelous, well done!

Geoff