Doubling up on BFD boost filters.

[chrisbee]

You may remember that Thomas has occasionally mentioned doubling up on BFD boost filters where the bottom end is thought to be lacking in an IB.

I usually run a full width 120/60 +16dB boost @ 20Hz on my 4 x AEIB15 array. Smaller cut filters at higher frequencies help to flatten the hump produced by the large boost filter.

The other day while playing around with REW I thought I'd finally try Thomas' suggestion.

So I just added another boost filter of +16dB @ 20Hz to the existing filters on the BFD. No attempt was made to flatten the resulting curve with new cut filters.

The results were absolutely insane! ;D

The bottom end lifted by at least 5-6 dB as can be seen on these "All measured" curves: The violet is the double filter trace with the much more rapid roll-off at higher frequencies.



I seem to have no room gain whatsoever. Traces can be easily separated into with and without BFD LF boost. Nearfield is much the same as the listening position. Ignore the sub-10Hz area.

If you imagine the violet trace boosted overall to follow the upper roll-off of the other curves then the sub 30Hz area lifts to insane levels! A house curve which would blow the windows out!

I tried a few minutes of Limp Bizkit "New old songs" and some other bassy CDs and quickly decided the extra boost was not for me.

I will try some more double filters @ 20Hz when I get the time but will use much less boost on each to see what happens. I'm hoping for a wider spread below 20Hz without completely overdoing the resultant house curve.

Be very careful out there if you think you'll try boosting low down or you might say bye-bye to your lovely drivers!

[ThomasW]

Actually you can thank Tom Nousaine for this tip. It's something he told me he was doing when we talked about EQing our respective subs, since we were both using the same Symetrix equalizer.

However the tip was with regard to coupling the filters and the effect of that creating a lower filter hinge point. There was no recommendation to use 2 filters to double the amount of boost

It's unwise to dial in this extreme amount of boost. The problem with that much boost is you're one high amplitide transient away from having the amp clip very hard. If that happens, the amp can dump DC into the speakers. Then instead of speakers, you have 4 large paperweights with the smell of toasted voice coils.

[chrisbee]

Thomas,

That much was understood but thankyou for the clarification as a warning to others.

I am far too conservative in my volume settings to do much damage.
The moment I coupled the second boost filter the doors started rattling like mad at almost zero volume. Cone movement was still modest but it was obvious that there was suddenly a serious house curve at work.

My IB is rather well balanced with the single boost filter. I will try a few coupled filter variations (using REW for guidance) with much lower boosts just to see what happens. A slightly wider plateau around 20Hz would be nice for films.

[chrisbee]

The use of Sonnie/brucek's cal file to 5Hz using the Galaxy 140 SPL meter now allows measurement with reasonable precision right down to 5Hz.

I have since played with much smaller doubled booster BFD filters @ 20Hz.

Even down to 2 x 20Hz +3dB 120/60 works at least as well as 1 filter @ 20Hz +16dB.

The whole response graph flattens with increased output above the roll off slope. This was quite unexpected. The massive +16dB boost at 20Hz had very little effect in reality in my own set-up except below 20Hz.

I'll post some comparison REW graphs later of various doubled boost filters.

Here's my response curve using the Galaxy 140 SPL meter.
It shows strong extension to around 8Hz. This is with the 20Hz +16dB boost. Doubling filters actually lifts the bottom end as well as higher frequencies.

[ThomasW]

chrisbee said:

I have since played with much smaller doubled booster BFD filters @ 20Hz.........

Even down to 2 x 20Hz +3dB 120/60 works at least as well as 1 filter @ 20Hz +16dB.

Good news....

That was the primary idea of using ganged filters. My guess is that you could use another 2dB or so of boost if you want to.

[chrisbee]

Thanks Thomas.

It is fascinating how the whole curve changes as boost is applied low down with these doubled up filters. I might find time tomorrow. (and remember to save the comparison graphs this time)

[htnut]

I wound up overlapping mine. One at 10 hz peaked and the next at 10 or so and rather broad. That worked out great in my room. Both at 10 peaked seemed to make the whole curve just tilt up toward the higher frequencies. Not sure why that. But with my Rane and the knobs to twiddle to my hearts content it doesn't take long to get just what I want.

[chrisbee]

The Rane is foolishly expensive in Europe so I'm stuck with the BFD.

Working with 20Hz filters is like fishing with a very long rod in muddy waters.

You never know what you might catch.

And if you do catch something it may not be very tasty. ;D

[chrisbee]

Quarter of an hour's quality time with my Galaxy 140 SPL meter, REW and my BFD resulted in the following:

Rather than display all of the images I will show the first run with "ganged" [ie. two identical filters set at exactly the same frequency (20Hz) on the BFD]

Here's the 2 x +3dB @ 20Hz first run: The black lines represent the calibration files loaded into REW. The thick black (drooping) Galaxy 140 cal file came from HT Shack. It covers 5-200Hz with C-weighting unchecked in REW.



Here's the "All measured" showing the progression as the double BFD boost filters are stepped up 1 dB at a time from 2 x +3dB to 2 x +10dB. All are based on a 20Hz filter setting: Filters were set at maximum width of two octaves (120/60).



Only the 20Hz filters were adjusted. No attempt was made to shape the curves with the existing higher frequency filters. No adjustment was made to level between runs. WYSIWYG.

Note the increasing noise around 5-7 Hz due almost certainly to the double doors to the enclosure rattling with increasing SPLs and increasing lower frequency excitement. I wish I could decide whether to build a manifold in place of the present array to see if it would reduce such mechanical noise.

[ThomasW]

Interesting stuff, thanks for providing the comparison...

Shows just how persistant the null at 15Hz is....

Regarding a manifold, I'm not sure if this is a viable option but could you build it so it's depth is split, 1/2 in the listening room and 1/2 out into the other space?

[chrisbee]

admin said:

Interesting stuff, thanks for providing the comparison...

A pleasure. I thought it might be useful to others with BFDs and a lust for more infrasonics.

Shows just how persistant the null at 15Hz is....

Probably distance dependent. I'll move my meter tripod forwards a bit and check if it makes any difference.

Regarding a manifold, I'm not sure if this is a viable option but could you build it so it's depth is split, 1/2 in the listening room and 1/2 out into the other space?

What would this achieve, Thomas? Wouldn't the wall would get in the way of the drivers(?) I was thinking about a standard 4 driver manifold with two 15" drivers in vertical pairs each side. If I turn the magnets inwards I won't hit the magnets when I open the door.

[ThomasW]

chrisbee said:

What would this achieve, Thomas? Wouldn't the wall would get in the way of the drivers(?) I was thinking about a standard 4 driver manifold with two 15" drivers in vertical pairs each side. If I turn the magnets inwards I won't hit the magnets when I open the door.

The idea was simply to not have the entire depth sticking into either area.

[chrisbee]

Okay. Thanks, Thomas.

I haven't had a chance to listen properly to these ganged filters yet so I'll report back with some SQ opinions when I have. The 2 x +16dB was certainly way over the top. I did listen to some Loreena McKennitt using my final 2 x +10dB settings and she sounded very nice. Though there's not a lot of deep bass involved. Time for some more organ music I think. ;D

[htnut]

Off topic, but that Loreena gal has some great pipes. I have never heard quite such a pure voice as she has. And that band has some very weird instruments.

Darned if I can remember the link, but here is my link to a program that lets you plot the response at the listening position for up to two subs. If I had any visual basic skills I would try to extend it to at least four subs.

www.digphoto.net/HT/FamilyRoom.XLS

[chrisbee]

Very off-topic for the IB Cult! I am very fond of Loreena McKennitt's music. I think I found her on the Celtic Circle series. Not only has she a range of completely different voices but remarkable breath control and expression. No two moments of her singing are equally loud. I believe she writes her own material as well as playing various instruments. I can listen to her CDs for hours without ever becoming bored. It must be the IB, BFD and the active crossover that Thomas recommended

[peted]

admin said:

It's unwise to dial in this extreme amount of boost. The problem with that much boost is you're one high amplitide transient away from having the amp clip very hard. If that happens, the amp can dump DC into the speakers. Then instead of speakers, you have 4 large paperweights with the smell of toasted voice coils.

Hi Thomas:

This always confuses me. It seems that it does not matter whether boost or large cuts everywhere else are used to create a hump. It is really that shape of the curve (tuned in the BFD) and overall volume the sub is played at that will determine amp clipping. Right?

Boosting or cutting in the BFD just changes the risk of clipping the input or output of the BFD when you get everything tuned right - sub level with mains. The "shape" of the frequency response curve leaving the BFD is no different. So, the amp shouldn't know the difference, I think....

Pete

[ThomasW]

Pete,

Yes it's a function of available displacement, amplifier power and the overall output/EQ settings.

However........

Chris was running +16dB of boost at 20Hz, that's extreme by anyone's standards.

My statement was to warn people that having a pair of filters at 20Hz each with +16dB of boost probably isn't the best idea.

[chrisbee]

I monitor my BFD and the EP2500 carefully during my filter experiments and listening tests.

I have only had the BFD above 4 bars lit when playing at foolish levels on certain material. WOTW Machine Rising is a good test. I ought to monitor this on my new Galaxy SPL meter.

My EP2500 orange lights usually only flutter without coming on steadily.

My stereo preamp also has a low output and has to be boosted by at least 6dB in my active crossover to bring the IB level up to match the speakers.

The 2 x +16dB ganged filters were just foolish with the obvious bottom end swamping everything else. It was unlistenable on music.

I have settled on 2 x +6dB for the moment for music. 2 x 10dB is strong on classical organ music and may be better for films. I haven't tried films using ganged filters yet.

It may simply be the easy 8 Ohms per channel load which my EP2500 enjoys which protects my IB from meltdown.

[canaris]

Chris, I keep my IB at 8 ohms also..

How do you double the filters.... you set 2 identical filters one after the other...?

[chrisbee]

You just set two identical filters at 20Hz, maximum bandwidth and chosen gain.

Remember to activate the second filter by selecting PA. Then double click the Store button when you've entered frequency (20Hz), bandwidth (120/60) and chosen gain.

Don't go crazy with boost because the infrasonic lift is quite severe as can be seen from my all measured graph.

I'd strongly suggest you have REW running and "target level" set when you set your ganged filters. Check your results with a quick sweep before you load up a DVD.

It is probably much safer to try a bass-heavy music CD to check cone movement at increasing volume before you finally test with a "heavy" film DVD.

Try raising your boost very slowly in 1dB steps from 2 x +3dB then sweep with REW each time. Then you can do an "All measured" graph to see your IB's response progress.

Have a listen to your music CD after each gain increase. I cannot emphasise enough that you place your drivers at risk by using ganged filters. My own system seems unburstable with very little cone movement despite very high levels at very low frequencies. Your own IB may be very different and go into destructive overexcursion on the very first listening test! You have been warned!

Every IB will behave differently with boost so I can accept absolutely no responsibility for wrecked drivers. Besides it was Thomas' idea, not mine.