Tweeter phase for first/second order crossover

I am working on some speakers that have had the bass driver swapped out and sound like they are lacking in mid range.

The tweeter is a Vifa D19SD which seems to be a higher power version of the D19TD that I understand is usually crossed over at around 5kHz.

The crossover has a series 3.3uF cap and inductor (unknown value) across the tweeter, and a series inductor (unknown value) for the woofer. So a combination first order and second order.

I don’t know what the original woofer was, so am only guessing re what the crossover frequency is.

The replacement woofer looks designed more as a subwoofer and has a stated range to 4KHz, 91dB efficiency (vs 89 for the tweeter).

So….I wanted to confirm that for this crossover design, swapping the tweeter phase is not going to make any difference?

If the previous statement is true, then the only way to recover some midrange is to drop the crossover frequency, so I was going to try a 6.8uF cap to start, and adjust by ear. Potentially lose some power handling capacity.

These are freebies for my niece so not trying to get them perfect, or to invest too much into it ;-)
 
I think 6.8uF is too much for this 3/4” dome tweeter. Try 4.7uF capacitor with 0.22mH inductor.
 

Attachments

  • 3950F292-FFCD-42AA-9988-9D2B2179113A.jpeg
    3950F292-FFCD-42AA-9988-9D2B2179113A.jpeg
    357.1 KB · Views: 65
  • Like
Reactions: Audio>X
You're probably right there. But the SD version of this tweeter is rated at 350W instead of the 100W of the TD.

To my other question - first vs second order phasing. If I add a cap across the woofer, does that then create an additive effect and a peak at the crossover point? Counter intuitive but I might replace the series cap with a 4.7uF and the try moving the 3.3uF to put across the woofer terminals.

I'm guessing the inductor is 0.33mH based on the Vifa recommended design for a 5kHz filter.
 
Well, there could well be a hole in the midrange. Manufacturer's claims for driver frequency range are often pretty meaningless, and the woofer certianly looks like it might roll off well before 4 kHz. The trouble is, how do you know for sure unless you measure?

If there is a midrange hole, a TC9 might well be a solution, but then you have the problem of coming up with a suitable 3 way crossover, when you don't know where the woofer is rolling off. That's hoping for a miracle, really.

While normally you wouldn't want to push that 19mm tweeter too low, you're up against it here, and it might be worth trying a bigger cap. I'd go for a small increase, though, at first - something like 4.3 uF would give a decent increase in level in the lower treble - maybe about 3dB in the 2-3 kHz range. The danger is,that if the hole is below that, you'll get even more boom and tizz. But it's worth a try, maybe. (I'd try @Galu's suggestion first, mind you.)
 
The midbass - tweeter distance is too big. Especially if you cut it at 4kHz. Try turning the phase on the tweeters. Find some frequency range app for android phone (Spectroid etc.). Using an online generator and amplifier, you play white noise on the speakers and watch the response on the phone. It is mostly accurate from 100Hz to 15kHz.
 
Nobody can answer that, without knowing more about the actual response. If you have second order acoustic slopes, which match nicely, you'd want the tweeter polarity reversed. But that's a big if.

Try it, both ways round, and see which sounds better.
 
  • Like
Reactions: AllenB
Try some measurement. Because no one can tell you straight away what and how. By removing the series inductance on the midbass, both the phase and the impedance will change. Also, without changing the crossover, just turning the phase on the tweeters will maybe add and maybe take away some of the midrange frequencies. That's why you should try it first.
 
For testing correct tweeter polarity you can use same method as in post #11.
Play some frequencies around crossing point, say 3-6kHz. If polarity is wrong, the drivers will cancel each other giving a certain "hole" in response at crossing point. If polarity is correct they will sum, with more/less same output.
 
Well, that was a learning exercise.

I tried shorting the inductor on the woofer, putting in a different capacitor, and also swapping the polarity of the tweeter.

I ran a couple of frequency plots and they were consistent in each scenario, even though I suspect there’s some factors that influenced them overall. (The regular dropouts were from the generator, not sure why but it kept pausing)

There seems to be a large peak in the tweeter around 13kHz. That also appears to be the limit of my hearing range, so it’s not apparent when listening. There’s also a dip between 4-5kHz in the starting setup.
IMG_0529.png


- Shorting out the inductor actually created a larger dip around 4kHz. Makes no sense to me.

IMG_0564.png

- Changing the cap to 6.8uF (only value I had at hand) created the smoothest response but added a peak around 3-4k

IMG_0582.png

- Swapping the polarity of the tweeter also improved things, with a smaller peak at 3kHz. I think I’ll run with this, and test an increase in the cap to 4.7k.
IMG_0576.png


Thanks for all the suggestions!
 

Attachments

  • IMG_0564.png
    IMG_0564.png
    303 KB · Views: 32