Using a high Q filter to fill a dip

[AllenB]

I want to fill a 4dB hole between 400Hz and 150Hz on a pair of mids, and then roll them off second order below 150Hz. I have been using a passive shelf along with a passive first order high pass to do this so far. If I go active and combine the job into a single high Q high pass filter that has a 4dB peak, will it sound OK?

[tsmith1315]

If the passive components are solving the issue, a single active filter should be a great solution. So, will you just center the peak at 275Hz or so? Elliptical?

Would you share your implementation and results with us?

[AndrewT]

could you use a Linkwitz transform to flatten the response?
If you flatten it all the way down to 150Hz, you will need quite a bit of gain.
Will that overload the Mid driver?

[tsmith1315]

It sounds like the response actually extends below 150Hz, and the irregularity really is a dip. Additional excursion is still a concern, but not as critical as when a LT is used to pick up a half octave or more.

[DSP_Geek]

Is your mid cabinet about 12 to 14 inches wide? The dip might be a baffle step, in which case a shelf is the right thing to correct it. We'd be able to figure out what was happening if you showed response curves.

In other news, a high Q highpass won't solve your problems: a 4 dB peak woud be much narrower than the hole you're trying to fill. Without calculating, I'd guess it'd be about a half octave wide at best,while your hole is three times that.

[tsmith1315]

It's really not that high-Q for 4dB. Looking at a 2nd order high-pass, Q around 1.5, F3 around 140Hz will give +4dB about 230Hz and is effective from about 150Hz-500Hz. Trouble is, it's not symmetrical. Roll off is steeper on the low end, and more gradual above the peak.

I wonder how low the mids need to go...

Not bad, though. And after all, this is car audio. Lots of unpredictables, close is often good enough.

[AllenB]

OK, these are midwoofers in the front doors. I tested them by sitting in the drivers seat with an RS meter and a sine wave generator. The dip at 400Hz was sudden but I didn't expect measuring a car to be easy. It looks to me just like a baffle step.

The linkwitz transform sounds ideal except for one thing. The passive crossover is working well for me except that the drivers bottom out at higher volumes (they are not perfectly sealed in the doors and are compliant drivers). I just need a few more dB to be happy. While I am crossing them over, I really need the 2nd order for power handling.

I tried simming it with passive components but it isn't easy to get a good peak at 400Hz and a significant drop by 80Hz using only passives.

[richie00boy]

I would imagine the sudden dip is a simple wavelength null rather than baffle step, in a vehicle. Try measuring from a slightly different position in the car and you will probably find the response changes.

[AllenB]

About the possible wavelength null, do you think it could be the way I've mounted them? These are six inch drivers and I've drilled and tapped holes in the rear of the magnet then epoxied and bolted them to a bracket that mounts on to a reinforced section of the door. They do not touch the trim and the fronts of the drivers are free. There is a small gap in front of them that could allow front to rear cancellation.

[richie00boy]

Hmm thats a very bad way to mount the drivers, they need at the very least a baffle to prevent front to rear cancellation.

This could be your problem but equally standing wave nulls at the measurement position could be too.