I'm trying simulations for converting 3-way speakers to 3.5-way. I couldn't find any guidance about it via online. So, I just did it my way. And, I need advice from experts to consider whether my design went correct way or not.
Let's start with the original 3-way system.
Note: the woofers are set the SPL at 74dB (-6dB from 80dB) in order to exhibit the graph was tuned flat. But the spike at 2.5kHz was intentionally left because of my preference.
What I did for the conversion was I doubled the inductance and halved the capacitance on the low-pass filter of the upper woofer and added twice larger inductor to the lower woofer, as seen in the picture below.
The criteria randomly set by myself was the attempt not to let the upper woofer's response exceed the original woofer's region. As you can see, the upper woofer, purple line, won't go higher frequency than the original woofer's response, red dotted line.
The green dashed-dotted line is the response when the original woofers were played at their actual level, woofer's SPL at 80dB (not at 74dB) equals to midrange's and tweeter's.
Finally, the impedance plots are shown below.
Are the frequency response and the impedance response presented acceptable?
Let's start with the original 3-way system.
Note: the woofers are set the SPL at 74dB (-6dB from 80dB) in order to exhibit the graph was tuned flat. But the spike at 2.5kHz was intentionally left because of my preference.
What I did for the conversion was I doubled the inductance and halved the capacitance on the low-pass filter of the upper woofer and added twice larger inductor to the lower woofer, as seen in the picture below.
The criteria randomly set by myself was the attempt not to let the upper woofer's response exceed the original woofer's region. As you can see, the upper woofer, purple line, won't go higher frequency than the original woofer's response, red dotted line.
The green dashed-dotted line is the response when the original woofers were played at their actual level, woofer's SPL at 80dB (not at 74dB) equals to midrange's and tweeter's.
Finally, the impedance plots are shown below.
Are the frequency response and the impedance response presented acceptable?
No, because these results require a flat impedance and frequency response from the drivers, but approx. 99,9% of real drivers don't have this, not to mention the effects of the baffle of the speaker to which the drivers are mounted. So your simulations doesn't represents what would happen in real.
On the other hand yes, if you satisfied with the results.
I can't see your baffle response but usually with a 0.5 way, one driver plays right through the baffle step and the other plays only up to it. If your woofers don't reach that high then should you be doing it with the mids instead?
Even as a theoretical exercise (which is all this can be), for a true apples to apples comparison the SPL of the woofers should be the same in the 3-way as the 3.5-way. If you're "virtually converting" the upper design from 3 to 3.5, keep the woofers at 74 as in the original, otherwise you have 6 dB more low bass in the 3.5 vs the 3 (as shown), when it should be the same.
In fact, the baffle step frequency has been covered by the woofers already in the 3-way system. But, in this case, the objective for the 3.5 ways isn’t intended for the baffle step but the more flexibility on bass tweaking. I discovered that when the mid and high were reduced 6dB (-6dB), or the woofers will be played 6dB louder (+6dB), some mid-bass will be raised as well causing a rich voice symptom. At the moment, I found it’s best when there’s a mild boost at only 3dB on the woofers, in the 3-way system. Nevertheless, I got an idea it would be better if the +6dB can be occurred without mid-bass affected. This is the reason for this 3.5-way conversion project.