Correcting 'valleys' in FR

[gainphile]

I know that to correct frequency response peaks notch filters are used. My question: is there a "reverse notch" to correct valleys, perhaps an active one ?

[boggy]

Originally posted by gainphile
I know that to correct frequency response peaks notch filters are used. My question: is there a "reverse notch" to correct valleys, perhaps an active one ?

Hi gainphile, one possible way to build an active "reverse notch", you can find on attached schematic.


Hope this helps!

[gainphile]

Thanks Boggy,

I could not see the attachment. Did you miss it ?

[ttan98]

Gainphile,

I notice you using the Alpha 15" driver as the OB bass driver.

These fairly large drivers sitting on top of your table seems a little daunting for me maybe not so much for you.

Cheers on a successful implementation.

[planet10]

Originally posted by gainphile
Thanks Boggy,

I could not see the attachment. Did you miss it ?

It was there... attachments are moderated in a different queue than messages, and if a mod is in a hurry he can forget to moderate the attachments attached to a post, in which case the message shows up & the picture doesn't.

dave

PS: you should be able to see it now.

[gainphile]

Originally posted by ttan98
Gainphile,

I notice you using the Alpha 15" driver as the OB bass driver.

These fairly large drivers sitting on top of your table seems a little daunting for me maybe not so much for you.

Cheers on a successful implementation.

They don't seem to big after awhile . From many experiments only dipoles can live in my small study.

Originally posted by planet10


It was there... attachments are moderated in a different queue than messages, and if a mod is in a hurry he can forget to moderate the attachments attached to a post, in which case the message shows up & the picture doesn't.

dave

PS: you should be able to see it now.

Yes. Seen it now ... thanks heaps I'm studying this circuit.

[David Gatti]

The passive way would be to design an enclosure shape where diffraction peaks occur at those valleys.

[boggy]

Originally posted by David Gatti
The passive way would be to design an enclosure shape where diffraction peaks occur at those valleys.

IMHO, this can be useable, if you build a loudspeaker which measurements (only) are "flat"....

But, if you have a some "drop" in speaker sensitivity in relatively small frequency range (one octave, for ex.), caused strictly by electro-acoustical parameters (impedance, ... ), not only acoustical (diffraction,... etc.), ... this problems cannot be cured with each other... Electrical problems must be treated with electrical circuits (filters, generally...)... acoustical problems... with better acoustical construction... (dampening, shaping, etc.)

[holdent]

gainphile:
I don't think correcting the "valleys" or nulls is a good idea. Some nulls can be 10 dB, 20dB, or more. Providing equalization for nulls in the bass region can require substantial power and may quickly overload your driver. I think a better approach would be as David Gatti suggests -- change your enclosure choice and knock down the response peaks as necessary.

...Or simply put up with the peaks. Some enclosures, such as the Nagaoka family of TL/horn hybrids at the Frugal-Horn website, have a number of nulls inherent in their design and still get excellent DIYer reviews.

[martinv]

Hi gainphile,

Some time ago I played around with using a parametric equaliser in order to tame peaks and valleys. For those who don't know, a parametric equaliser typically has three controls; one for selecting the centre frequency, another for setting the boost/cut level, and a third for adjusting the width of the frequency range being affected. Kind of like a graphic equaliser with just one slider, but you can adjust the frequency the slider takes effect at as well as the 'broadness' of the range being affected. Maybe a circuit for such a device would be worth looking for. I think Silicon Chip recently had a sub-woofer filter that incorporated such a design. Let me know if you need to know the issue and I'll have a search.

OK, that's your question answered... Now my advice - don't do it. :-)

As opposed to peaks, it's popular belief that valleys are nowhere near as noticeable.

Here's another thought - what is causing the valley? You may be fighting a losing battle if it's a room interaction problem.

Yet another thought - this valley - does it occur at every listening point? (I know that's difficult/impossible to measure) If not, you might be fixing up one point at the expense of other listening points. A bit like the experiment I did to decide whether to offset a driver - www.gadgets.co.nz/mjv/loudspeakers/exp01.shtml

Anyway happy open baffle building.

-M