Active circuit to compensate for a dip ?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi !

I've built sealed boxes for Fostex F120A drivers, but I got a bad surprise when measuring the driver : it exhibits a -10dB dip at 5kHz !! (not a problem with the enclosure)

The speakers will be active filtered, so I can build an active circuit to compensate for that.

The problem is that I only found notch filters which do the opposite of what I want.

Is there a circuit that could help me shape this weird response ?


Thanks in advance ! :)
 

Attachments

  • f120a frequency response_thd_group delay.pdf
    8.8 KB · Views: 68
No, it is not an issue with the room or the enclosure or the driver.

Tested in a different room, on an open baffle, the result is the same.

And it is no problem with the mic either, other speaker measure normal...


Fostex really cheats on us in their datasheet...
 

Attachments

  • response_1m.png
    response_1m.png
    23.2 KB · Views: 137
The dip is here because the drivers are like this, and I don't know why... Bad design for sure.

The solution would be to replace the F120A with some FX120, but now I've bought the F120A I have to deal with this sucking dip...

To get rid of it, I can use the equalizer of the audio player, the problem being that these speakers won't be used with a computer...

I could also build a horn with a very narrow response, but I have already built the enclosures, and I'm not sure about the result anyway.


As I will active filter these speakers, the only solution I see is an active circuit.

However, if you have another solution, I'd really welcome it !!! :smash:
 
I don't really know anything but I'd like to make a guess.

The dip is around 5k and could be a number of things. One thing that comes to mind is at that frequency it could be from a reflection about an inch or so behind the cone. These speakers have a big magnet that covers(reflects) a lot of the backwave. You could try putting a little wool felt on the inside of the basket legs and over the outter edge of the magnet (without touching the spider) and make another measurement.

I'd try a mechanical solutions before adding more circuitry, but that's just me.

Scott
 
consort_ee_um : nice circuit ! I have plenty of ne5532 that I could use for this purpose. Hope it doesn't degrade the sound too much ; anyway it seems I may have no other solution.

scottw : that cancelling due to standing waves could be a point ! If it is the case, that would be a simple solution that doesn't involve circuitry.

That's strange the FX120 don't have this issue, they have exactly the same chassis, spider and cone dimensions.

I'll give it a try when I'm back home ; as I don't have the speakers in my flat unfortunately...

Strange Fostex designed such a bad driver, and I don't see how they can lie like this in the datasheet, apart from including another driver's frequency response...
 
Compensating for dips with an eq is usually a bad idea. If you move the mic a bit the dip most probably looks different. This means that you will end up with a peak in some directions if you compensate for the dip, and a peak is much more audible than a dip.

Furthermore, it seems as if the driver has a large amount of distortion at the dip frequency; amplifying that region will not reduce the distortion.
 
forr : I'm not sure but, a bandpass filter doesn't change the gain at one frequency (what I need), it only allows some frequencies to pass through it. Is there a way to "parallel" this circuit with another in order to introduce a peak ?

Svante : what you say about distorsion is right. It is quite obvious in the PDF i posted at first. However, I've listened to these speakers with and without equalization (done by the computer) and it clearly needs it (10dB is huuuuuge).

Off axis, the dip is a bit less pronounced, but still here (45° off axis, on open baffle, attached)


Well, I see no other solution except selling the drivers and buy FX120 instead :D
 

Attachments

  • offaxis_45°.png
    offaxis_45°.png
    23.7 KB · Views: 70
Hi Youyoung,
You can certainly use the Deliyannis circuit in a parallel configuration to keep the gain flat at the other frequencies.
However, there are many adjustable equalizing circuit, at least to test the validity of the correction you think of. For example, there is an old Pionner equalizer where you can adjust the frequency, the damping and the amount of correction. Once you are sure of the parameters of the required equalisation, it would be rather easy find a simple circuit with fixed components to mimic it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.