Taming the High Frequency of Tangband's Titanium W4

[borispm]

Hi guys,

I am going to order a pair of Tangband's W4-1337SD titanium cone full range drivers and I plan to run them full range i.e. no aiding tweeter or woofer.
The problem is there's a peak from 10kHz all the way up to 20kHz in Tangband's published graph and from other measurements I've seen (e.g. John a.k.a Zaphaudio) there's also a rising from 6kHz onwards.

I am pretty frustrated by that peak but I still want to give those drivers a try, which filters should I try? Should I give them a single inductor to roll off the rising response or something like an L-R trap or series notch filter?

Regards,
Boris

[ttan98]

Hi,

try this site, MiniStatements

look at net 3

the L3011 and C3011 parallel with remove your peak.

[borispm]

Hi ttan98,

First of all, thanks for your reply.
I am curious why I need the capacitor to be parallel with the inductor. It looks like a low pass parallel with a high pass now.

Regards,
Boris

[ttan98]

Originally Posted by borispm

Hi ttan98,

First of all, thanks for your reply.
I am curious why I need the capacitor to be parallel with the inductor. It looks like a low pass parallel with a high pass now.

Regards,
Boris

You need a notch filter, the one shown is a variation, I use a resistor, inductor and capacitor in parallel(in another driver). To do it yourself you need to measure the rising bump and design a notch to compensate.

[borispm]

Hi guys, I managed to get a clearer official response graph, so the problem should be more visible now. If I opt for a single inductor, I may be losing much of the SPL above 15kHz in order to get a flatter response, and if I use an L-R trap to get rid of the broad yet abrupt peak at 10kHz to 20kHz, the rising response from 6kHz onwards will not be solved.

[ttan98]

An inductor is not sufficient to remove bump, only a notch will remove a bump effectively. I suggest use this circuit first to see whether it will work effectively.

[hm]

Hello,
i think you overestimate "Musik" at 10 kHz, do nothing, i prefer for this driver
my KORNETT, double horn fulrange, no crossover.
over 10 kHz much SPL is for ages over 50 years not bad.
if you don´t want it change the angle 10° step by step.

[borispm]

Hi ttan98, can I use the L-R approach used in BSC circuits? Or will the slope be too gentle?

I want to keep things as simple as possible. I have seen some elaborated filters for this driver but I want to keep the essence of minimal components.

[borispm]

hm, thanks for pointing that out. To most grown ups it's not an annoying problem, but I am still quite sensitive to high frequencies.

I just can't bear the 15kHz noise that CRT TV generates, so I guess the 10dB bump will give me pretty unpleasant experience.

I understand your point on the off axis, but I guess I can have moderate amount of toe in...

[borispm]

hm, by the way the w4-1320 is a paper cone and seems to have more gentle breakup than the w4-1337sd so no wonder you do not need filters of any sort.
I wanted to try the bamboo cone too, but I like how the titanium cone looks and the area I'v living in is very humid so I do not want the cone to soften in spring and sound different, depending on seasons.