Xover comments for 830656 and XT25SC90-04

[Havoc08]

Hi all,
I am making a 2 way using these two components and have started from scratch because I want to learn a bit more about Xover design which seems to be the most important part .

I have simulated a 15L box for the woofer with a port tuned to 49hz which seem reasonable (taking into account the volume I can afford).

I have used the T/S obtained by zaph for the peerless 830656 (except sensitivity when calculating L-pad, I referred to both manufacturer datasheets for this).

I have followed Passive Crossover Network Design thoroughly and calculated by hand, supported by Crossover Design Chart and Inductance vs. Frequency Calculator(Low-pass)

The Xover is a 2nd order electrical LR crossed at 2500hz.

Firstly I made a zobel for the woofer. C= 26,7uF (this value is 17uF on some online calculators for some reason) and R= 7,375ohm.

With the new woofer Zmin of 5,9 ohm I made the 2nd order LR which was L2=0,75mH and C2= 5,4uF for the woofer and C1= 8uF and L1= 0,51mH for the tweeter (polarity inverted).

I then calculated the L-pad taking into account the loss in the woofer coil which gave me a Rs= 2ohm and Rp = 5,6ohm equal to a 5,37dB attenuation (calculated a wanted attenuation of 5,2dB). The L-pad is placed closest to the driver like the zobel.


Now only thing I haven't done is calculate the BSC which I presume is needed? The formular from true audio is L=(Wb*R)/1,021 but as I am building an iPhone speaker I can't figure out the width to use? Trial and error or other ideas?


Any comments of whether my math seems correct and/or if there is anything I have left out or done wrong?

Should the BSC be at the very first before the 2nd order LR and Lpad? If so does that alter any of my calculations when I introduce 4ohm more additional to the 4 ohm of the tweeter + coil resistance?

P.s. I have not converted my values to standard values yet to make it easier to follow my math and thinking (or westhosts more truthfully ).

Thank you in advance

[giralfino]

I'm not an expert, but I did built a speaker with the same woofer. You can search in the forum. I'm leaving home for vacation so I will give a very short answer.

Baffle step: I think you have to simulate the effect with some software, and then you choose the xo. For example with a xo point of 3.5K Hz I used a 1.35mH inductance.

Tweeter: Zaph said that this tweeter should be used LR2 at 3K Hz or LR4 at 2.5K Hz, so probably you should set a higher xo point, IMHO the woofer is able to arrive to 3-3.5K Hz.

Volume: I designed a 12L box, and is probably too big. I would choose a 10L box and tuning to 50-55 Hz.

Ralf

[Havoc08]

Hi giralfino.

a 10L ported box tuned at 50hz has a mild 1db gain in the bottom and could work. What is the argument for choosing this and not my box of 15L?

I read what zaph wrote, but I feel 3khz for the woofer is too high with a 12dB filter because of the 4khz dip and 5-6khz peak.
Perhaps I should make the tweeter impedance flattening filter also If I want to cross at 2,5khz with 12dB filter (I presume this Xover point is too low considering the Fs of the tweeter, but the impedance flattening would help)?

Regarding software I only use winisd, but it is very basic. I tried lspcad and speaker workshop, but they seem quite complex. Maybe I should give unibox a try or try speaker workshop again?

[Havoc08]

p.s. have a nice holiday. When you get back we will discuss your filter as you have different zobel values than me and only a 1st order on the woofer?

[Havoc08]

Does anyone know why you cannot just take into account the baffle step loss in the L-pad instead of making the BSC filter?

[giralfino]

a 10L ported box tuned at 50hz has a mild 1db gain in the bottom and could work. What is the argument for choosing this and not my box of 15L?

A too big box does sound boomy. In my speaker (12L) this effect is present, but fortunately only noticeable with speech.

After the realization I found on the vifa.de site another project with the 830656 midwoofer and the Vifa BC25TG15-04 (but they are named differently for the German market). The box is 8.8L tuned to 57 Hz. See StartAirKit 3R.

I read what zaph wrote, but I feel 3khz for the woofer is too high with a 12dB filter because of the 4khz dip and 5-6khz peak.

See the attached figure for the effect of the filter on the woofer response: the peak is almost flattered, and a little off-axis vanishes.

Does anyone know why you cannot just take into account the baffle step loss in the L-pad instead of making the BSC filter?

The baffle step effect is in the range 100-1000 Hz, depending on the baffle, so it alters the woofer response. You lose 6dB on the lower end because of the fact that the driver radiates in a 4pi space, whereas on higher frequencies it will radiates in a 2pi space (i.e. only in front of the baffle). Again, the attached figure shows the effect of the baffle step on the driver response (look at the infinite baffle response at the Zaph site). The main inductor of the filter here works not only as a low pass, but also reduces the baffle step effect.

Regarding software I only use winisd, but it is very basic. I tried lspcad and speaker workshop, but they seem quite complex. Maybe I should give unibox a try or try speaker workshop again?

See FRD Consortium tools guide for a tutorial on using FRD software.
I then used Passive Crossover Designer for the filter: Loudspeaker Design Software

Have fun, Ralf

[wolf_teeth]

If you find the 2.5kHz LR2 xover is a bit 'nasal' sounding, drop a 25-33 ohm resistor in across it. This will nuke about half the Fs magnitude, and regain it's smooth un-nasal character, as well as give you about 1-2dB of padding tops. It's a useful tweak with a dual-purpose result.

This works in both SXO and parallel types.
Later,
Wolf

[Havoc08]

Hi both,
@ Giralfino
The box would only sound boomy if it was tuned for too high a Q right? When simulating a 15L box ported it looks nice and flat (can't remember the port tuning, but can check when I get home).

Anyways I made the enclosure 2x15L last weekend so we'll see . Otherwise I'll have to stuff it with some volume absorbing foam, i.e. not open pored.

3khz looks not half bad, I'll use that if 2,5khz or Wolf's addition doesn't sound good.

I understand what baffle step effect is, but not why you cannot use the L-pad to reduce the tweeter 6 dB. I found out in the mean time though (I think). The BSC filter gradually lowers the upper range proportional to the X-axis (hz). If you just lower the tweeter the 6dB then you would get a 6dB bump in the midtone-highmid region as the woofer only looses in the low section due to **, but not the high section and the tweeter is lowered 6dB in its entire range.

_____
________-------- \
_______--------- \______________________
Woofer -6dB ** Tweeter -6 dB


@ Wolf
Thx for the diagram, I will give it a go if it sounds nasal after the initial burn in.

[Havoc08]

Well my ASCII art got destroyed and the ** says baffle step in short, which is banned because of its relation to a bull I guess



This is what I mean.

[wolf_teeth]

The resistor was my intent, not the SXO as a whole. Just to clarify...

Hope it works for you!
Wolf