my Tumult...depressed at 22 Hz

[kramskoi]

I would welcome any observations and tips on the following dilemma at 22 Hz:


Image not available


I've tried moving the sub up, down, backwards, forwards and every which way and the null remains. I've done manual sine wave sweeps with the same result. This sweep was done at 2m. The same sweep is 5 dB better (22 Hz) at 6m near the rear wall, but it still dips.

The following is a picture of the front part of the room where the sub resides (it's in the left corner):

Image not available

This little cubby hole is 6 feet wide and 10 feet deep and it opens into the larger part of the room which is 13 feet wide and continues to 22.5 feet at the rear wall. This room can be totally sealed so i'm getting good room gain but also wacky frequency response.

This is a sealed Tumult 15" which is 23" x 21.5" x 21.5" so it must absolutely stay behind the LCD. I plan to use a Bassis and one of the feedback units from Behringer.

I also have some disarray in the 60-70 Hz area, but i think this will be easier to deal with. I'm using Klipsch SB-2's, which are rated to 54 Hz (-3dB). The sub does'nt have phase adjustment, as it is passive and driven by an ep1500. No matter what i set the crossover to, the null appears in various places between 60 and 90 Hz. I'm using a RS digital spl meter (set to 80 dB range) in conjunction with roomEQ wizard for the measured sweeps (10-200 Hz). I can provide more sweeps if desired. Thanks

[Paul H]

[FONT="Times New Roman"] ... The sub does'nt have phase adjustment, as it is passive and driven by an ep1500. No matter what i set the crossover to, the null appears in various places between 60 and 90 Hz. ..

While you don't have phase adjustment, have you tried flipping the plus and minus speaker wires to your sub? It will give you a 180 degree phase change. The revised sub and mains interaction could possibly give some improvement.

Paul

[kramskoi]

While you don't have phase adjustment, have you tried flipping the plus and minus speaker wires to your sub? It will give you a 180 degree phase change. The revised sub and mains interaction could possibly give some improvement.

Paul

Sorry Paul, i neglected to mention that my receiver has subwoofer phase reversal (180 degrees) and it's worse at the REV setting. Thanks anyway. :T

[ThomasW]

You can't fool father physics. Put a sub in a tunnel and you're going to have problems.

Pull the sub back to the listening position and substitute it for your coffee table, your null will disappear in nearfield operation...

[kramskoi]

You can't fool father physics. Put a sub in a tunnel and you're going to have problems.

Pull the sub back to the listening position and substitute it for your coffee table, your null will disappear in nearfield operation...

your input is noted and appreciated Thomas. However, putting two tumult 15" in a 1,800 cu. ft. bedroom is'nt feasible. I'm sorry that i did'nt specify.

The response results are uncorrected for the RS digital meter. I was under the impression that the digital meter loses sensitivity in the low bass regions, especially below 30 Hz. Is this correct? Perhaps i'll set the wizard for c-weighted compensation on future sweeps.

I would hate to lose the output i currently enjoy by moving the sub away from the walls toward the center of the room. I recorded 120 dB @ 2m (uncorrected) for the SW II fly-by scene.

Anyways, thanks again for the suggestion...maybe i'll try it tomorrow and see what happens. :T

[ThomasW]

In that small a space you're not going to loose any output by moving the sub.

The bottom end roll-off of the RS meter isn't responsible for that room null.

[jdybnis]

I can't see the where the sub is in that picture. Try moving the sub to next to the desk, against the wall on the side closer to the camera. How deep is that cubby? (just under 4m?)

[SteveCallas]

I believe it was found that the data out on the digital spl meter is already corrected, so no need to apply c weighting with roomEQ wizard. It's only if you are going off of what the display says that you need to apply correction.

[kgveteran]

With an enclosure of that size sealed, your F3 is about 35hz. That driver requires some help from a L/T circuit.

I'm sure the location doesn't help, but the big issue is a Tumult sealed and the rolloff.

Even with room gain you are going to be down at least -12db@20hz.

You'll set the BASSIS at about 30-35hz and add about 8-12db of (gain-extention) to it.

[ThomasW]

but the big issue is a Tumult sealed and the rolloff

That's certainly not what's happening in his room given the nature of his response plot.

Even with room gain you are going to be down at least -12db@20hz.

Generalizations that simple aren't helpful. I can post plots showing +10dB or more of room gain at 20Hz from a sealed box.

[Ilkka]

Could you do a nearfield measurement? Just put the RS around 2-3" from the cone and take a sweep. Set the range high enough, so that it doesn't clip.

I'm sure the problem is your room, but I would like to confirm it.

[kramskoi]

I can't see the where the sub is in that picture. Try moving the sub to next to the desk, against the wall on the side closer to the camera. How deep is that cubby? (just under 4m?)

the cubby is precisely 3 meters deep. The photo was taken before the sub was moved into position. It resides in the left corner of the picture currently. I may try it out in that position...against the wall, i would have 1/4 space as opposed to 1/2 in the center of the room. The problem is the chairs and the fact that this is a bedroom...there's just not much room in which to work. I'll post a larger picture of the room and maybe a drawing to scale. There is another corner placement in the room, but it is so far from the mains that i'm reluctant to position it there. Thanks :T

p.s. added a drawing of the room...to scale:

[Jack Gilvey]

I would hate to lose the output i currently enjoy by moving the sub away from the walls toward the center of the room. I recorded 120 dB @ 2m (uncorrected) for the SW II fly-by scene.

I think you could probably afford to lose a bit of output to get a flatter response. The capability of the sub likely far exceeds the capability of the L/R/C as it is, at least as you get into the midbass/crossover region for the sats.

[kramskoi]

Could you do a nearfield measurement? Just put the RS around 2-3" from the cone and take a sweep. Set the range high enough, so that it doesn't clip.

I'm sure the problem is your room, but I would like to confirm it.

yes... i can do this. I think you are correct though...it's the room. I've run probably 200 sweeps, and the same null every time. There is an alternative placement in the diagram further down in the thread. You can see in the pic what i'm up against. Like Thomas said, it's "in a tunnel". Thanks :T

[ThomasW]

Options, none very good......

1) Put the bed in the tunnel and move the A/V gear to the opposite wall

2) Get a Behringer DEQ2496 ($300), more $ that the other EQ's but it can replace a Bassis and give you standard EQ as well. Try to boost the null, after using shelving filters to pull down everything on either side of it. Usually one can get 'some' improvement doing that.

3) Place sub by "X" or behind the listening position with a VERY low crossover point (40Hz or lower)

4) IB sub in the ceiling over the listening position, if that's an option.

5) Live with it, be happy with what you have

6) Sacrificial offerings to the housing gods for a better room......:wink:

[kramskoi]

With an enclosure of that size sealed, your F3 is about 35hz. That driver requires some help from a L/T circuit.

I'm sure the location doesn't help, but the big issue is a Tumult sealed and the rolloff.

Even with room gain you are going to be down at least -12db@20hz.

You'll set the BASSIS at about 30-35hz and add about 8-12db of (gain-extention) to it.

I don't know what the f3 is but i do know that the fsc is 31.6 Hz, as measured with an impedance analyzer. :T

[jdybnis]

Yeah, try putting the sub directly across from the X, next to the desk. Don't worry about maximum output until it is demonstrably a problem. You are tying one hand behind your back by limiting your placement options. Solve one problem at a time.

P.S. That's a Big A** Closet!

[kgveteran]

That's certainly not what's happening in his room given the nature of his response plot.
Generalizations that simple aren't helpful. I can post plots showing +10dB or more of room gain at 20Hz from a sealed box.

I can only go by my sealed Tumults.If i post a plot , it would look just like the one in the original post.A tumult in a 3.0 cubed box will begin to roll off at about 30-35hz at about 12db/oct.Room gain will lesson that effect as did in my room.

The only way you could be certain of anything is to do a close mic.Since that has yet to be done, I suggested that the BASSIS would give him the boost he need out of a sealed system.

No need to throw around generalizations, I have four sealed Tumults and I have a BASSIS along with (4) 1500watt mono QSC's and this situation is very typical a sealed subwoofer.A sealed system needs four times the amplifier power of a ported system.I'm now measuring below 10hz @ about 100db with very little THD.With out the BASSIS this would never happen.Room gain came in at about 22hz down to 16hz.The rise was not enough to do it alone with out an L/T circuit.

[Exocer]

Definitely when you get the chance, run some sweeps with c-weighting compensation on. This is comming from an analog RS spl meter owner, although i do believe the same applies to the digital meter. Also get a hold of the latest and greatest .cal file before taking more measurements. My dip @ 30hz disappeared after applying the .cal file..from there the BFD helped to more accurately calibrate a house curve. I ended up with this:

[ThomasW]

Veteran,

and this situation is very typical a sealed subwoofer.

No a dip like that is certainly not typical of the performance of a sealed subwoofer. It is however stereotypic of a room induced null. And that conclusion is reinforced by the fact that there's another null approx one octave above it.

I can only go by my sealed Tumults.If i post a plot , it would look just like the one in the original post

Room gain came in at about 22hz down to 16hz

Then obviously your plot did not look like the one in kramskoi's original post, where there is a huge null at 22Hz.

[kramskoi]

???????????

this is the sub close mic'd at 2 inches from the surround. The sub is still in it's original position in the left front corner of the room. The crossover is set to 200 Hz and the RS digital meter is set to 80 dB. No c-weighted compensation is being used. This does'nt look good...

Image not available

[ThomasW]

Yep that's a pretty typical nearfield response for a smallish sealed box

Compare that to your other plot, and you see what the room placement is doing to the FR.

[SteveCallas]

This does'nt look good...

That is the typical output of a sealed box with no EQ applied and no room gain. Plug it into a modeling program and that is virtually identical to what you will see.

[JonMarsh]

It's EXACTLY what I'd expect to see nearfield with no EQ with a big high Le driver in a small box. Response always looks like a hill, peak at box Fb.

[kramskoi]

Thanks everyone...appreciated

I measured 31.6 Hz for the Fb with an impedance analyzer, so i'm not sure about the peak band from 38-43 Hz. Looks like i can lop off about 6 dB there! The internal volume is between 3.3 and 3.5 cu. ft. Anyways...

I did a sweep with the sub against the wall near the desk...not worth mentioning! I did a second sweep in position X on the room diagram...again not worth mentioning! Finally i did another NF @ 2" and recorded a smoother response. This NF was done out in the center of the room.


Image not available


Evidently, there's still "some" boundary interaction, even at 2" mic distance. The sweeps near the desk and position X are clearly worse! That said, the beast is going back into the left front corner. I'll leave it to BFD and the Bassis to iron out the ripples as best possible. I sincerely thank all who responded...much appreciated. :T :T :T

[jdybnis]

Not worth mentioning, why? Not enough gain, or big dips?

[ThomasW]

This NF was done out in the center of the room

Yep smoother. If you want it really smooth carry it outside and run a 2m ground plane.

I'll leave it to BFD and the Bassis to iron out the ripples as best possible

If you haven't purchased these then consider the DEQ2496 ($300 gets you both a BFD and an LT circuit). Or start with just the standard BFD.....and see what a bunch of cut gets you. The Bassis/LT may not make much sense given the issues with the room/placement.

[kramskoi]

Not worth mentioning, why? Not enough gain, or big dips?

yes...sharp dips in the 20 and 30's range. It reminded me of stage profiles of the tour de france! 8O

[Ilkka]

Your FR rolls off too fast being Tumult in a 3.3 cu ft box. Apply C-weighting (REW) and see how it changes.

And actually Fc at 31.6 Hz suggests around 3.7 cu ft and ~0.65 Qtc.

[kramskoi]

Yep smoother. If you want it really smooth carry it outside and run a 2m ground plane.
If you haven't purchased these then consider the DEQ2496 ($300 gets you both a BFD and an LT circuit). Or start with just the standard BFD.....and see what a bunch of cut gets you. The Bassis/LT may not make much sense given the issues with the room/placement.

Thanks again Thomas. If the sweeps are any indication, then system Q is nowhere near .57, which is what i'm aiming for. If the F6 bandwidth is the difference between Fb and F6, then there would appear to be more system ringing than i'd like. I read your short dissertation on using the DEQ2496 as a "defacto" Bassis. I must confess, i'm not sure that i totally understand the graphs you posted.

I'll look into the DEQ2496 but i've read about issues with reliability. It seems it's designed to do TOO many things at once. I may just go with the FBQ2496 and then get the Bassis. I don't know...i'll see. :T

[kramskoi]

Your FR rolls off too fast being Tumult in a 3.3 cu ft box. Apply C-weighting (REW) and see how it changes.

And actually Fc at 31.6 Hz suggests around 3.7 cu ft and ~0.65 Qtc.

that's correct Ilkka...that's what i computed a few weeks ago. The sweeps were done without c-weighted compensation. The cube is 19.5 X 19.5 x 19.5 internal. A lot of the displacement (probably 1/2) is cut for the driver because the front baffle is 3.5" thick and the driver is not countersunk into the cabinet. There are 4 internal braces which are 3/4" thick and i've got it lined only with 1" foam and no polyfill. This should theoretically make the box appear larger to the driver.

I used an impedance analyzer to find the exact Fb after it was sealed up. I've read that there are deviations in RS analog meters but maybe not as much for the digital ones. It seems that these meters should be tested against one another and a calibrated, professional microphone.

Ilkka, this a compensated sweep from 10-100 Hz:

Image not available

Thanks :T

[Ilkka]

I used an impedance analyzer to find the exact Fb after it was sealed up. I've read that there are deviations in RS analog meters but maybe not as much for the digital ones. It seems that these meters should be tested against one another and a calibrated, professional microphone. Thanks :T

Yes, all RS meters are different. I have a few comparisons made against a calibrated mic and some are off only 4 dB at 20 Hz, some as much as 12 dB. So we can't really use any common correction factors if we want accurate results. ECM8000 is pretty cheap and much more accurate mic for the job.

[Scott Simonian]

Wow. Ilkka, should I throw out all the test I have done at home with my RS meter? Don't know if I can afford such a nice mic.

[Ilkka]

Wow. Ilkka, should I throw out all the test I have done at home with my RS meter? Don't know if I can afford such a nice mic.

You can use those general corrections or C-weighting corrections with it and get relatively accurate results, but dead accurate? Not without a calibration.

[steve nn]

I may just go with the FBQ2496 and then get the Bassis. I don't know...i'll see.

I've had a chance to work with mine (FBQ/BASSIS) today with some help from Ilkka and I got to tell you.. it's quite a tool in shaping your curve or going ruler flat if that be the case. I'll post a few before and after graphs tomorrow Kramskoi if I get things wrapped up?

[kramskoi]

I've had a chance to work with mine (FBQ/BASSIS) today with some help from Ilkka and I got to tell you.. it's quite a tool in shaping your curve or going ruler flat if that be the case. I'll post a few before and after graphs tomorrow Kramskoi if I get things wrapped up?

That's fine steve. I'll just have to live with what i have. I don't like it being depressed at frequencies this low. This type of situation is usually only solved by a second subwoofer. I think i'll get good results with the above mentioned combination, though it may not end up being ruler flat.

I like the DEQ2496 that Thomas mentioned but i'm thinking that the Bassis would be less complicated for me. Thanks :T

[ThomasW]

I ask Dennis H to type up a LT DIY for the DEQ, here's his text....

The PEQ function of the DEQ2496 can be used to duplicate the LT with three filters: two parametric band filters and one 12dB/octave shelving filter. You'll need to measure the response of the woofer as you apply the filters to get them tuned. The pic shows an example using the LT to covert a 40 Hz, Q=1 box into a 20 Hz, Q=0.5 box.

Image not available


The first parametric filter is used to adjust the Q of the existing rolloff to 0.7 so the response is -3dB at Fb. If your box has a low Q, you'll need to boost a bit at Fb and if the Q is very high, you'll need to cut some.

The 12dB/octave shelving filter extends the bass. You could use a low shelving filter to boost the lows or a high shelving filter to cut the highs. Unfortunately, the lowest F you can set is 20 Hz so the low shelving may not get you as low as you would like. F is the lower of the two corner frequencies with the low shelving. Using the high shelving, F is the higher of the two frequencies, so just set that to Fb and cut the highs until the response looks like you want it to.

The second parametric filter is optional. After the shelving filter is applied, the Q at the new Fb will be 0.7. If you want to change that, you can use a parametric filter to boost or cut at that frequency and shape the rolloff.

If practice, you don't really need to worry too much about calculating Q and all that stuff. A perfect theoretical LT will work perfectly in an anechoic chamber but we use our subs in real rooms with nasty peaks and valleys, room gain, etc. We'll also be using the PEQ to correct for the room so we can forget the two parametric filters of the LT to start with. Just use the 12dB high shelving filter at your in-room F3 and cut the highs as much as you'd like to extend the bass. Once you have the extension you'd like, you can start using the band filters to flatten out the peaks and dips. That's really all there is to it.

The benefit to using a Bassis is that you can EQ lower than 20Hz and somewhat simpler operation. The downside is the price more than the DEQ and all you get is the LT function.

[steve nn]

I don't like it being depressed at frequencies this low. This type of situation is usually only solved by a second subwoofer.

So many variances when it comes to dialing in FR. I'll hope that the second sub doesn't cause a whole new set of issues to deal with. One thing is for certain though, your going about things in the right way. If you can get your FR to taper off at 20, you'll be in there.

[ThomasW]

This type of situation is usually only solved by a second subwoofer.

Where would you put it?

[HMenke]

I feel out of my league here but I have a few ideas FWTW.

1) RoomEQ Wizard is very useful in sweep mode. I moved my sub around all over the room to find the flattest "natural" location before final EQ'ing with the Behringer Feedback Destroyer Pro DSP1124P.

2) Alternately, put the sub at the listening position and move the RS meter around all over the room doing REQW sweeps to find a place where the null might disappear or become less severe. This location is a candidate for a new sub location.

3) Don't forget the upper half of the room. It might be an option to try locating the sub on a shelf in a ceiling corner. Considering the ceiling gives more different corner options. I have a small rectangular and symmetrical HT room and it is amazing how only one corner is "best". I would have tried the ceiling if none of the floor corners had worked.

4) DIY corner bass traps to try to kill or moderate the null.

[kramskoi]

Where would you put it?

just musing to myself Thomas...generalizing mostly. I'll just make the most of what i have and ditto the rest. Thanks :T

[steve nn]

I'll just make the most of what i have and ditto the rest. Thanks

We're all in the same boat here trying to get the best we can in our imperfect rooms. I got so many stories! Anyway here is the graph I said I would post... it reflects incorporating the BASSIS with the FBQ. We cant be sure about 14 Hz and under.. but it seems reasonable another three to five in FR Hz is on track. This was a hr. eq job, so admittedly I could improve it a little.

[Jack Gilvey]

I ask Dennis H to type up a LT DIY for the DEQ, here's his text....

That's a great write-up, looking forward to the rest. I especially like the last paragraph, which is a point I'm always making about real-world vs. excel results for FR, Qtc, etc. It looks like the DEQ will be a sweet unit, and should allow extremely precise real-world results.

[Mark Seaton]

stevenn,

Good job there. That resulting graph is exactly why there are still compelling reasons to use sealed subwoofers. Once you get there, you just need to use enough drivers for your desired playback level. Now how does that response sound subjectively?

Jack,

Dennis's write up is very good. The only thing I would add is that you can easily extend the response of most sealed subs to well below 20Hz with any DSP based EQ that has shelving filters, even if the lowest frequency setting is 20Hz. Most I've measured have little trouble passing and affecting the response down to 10Hz. So long as the corner of the un-EQ'd response is in the working range of the DSP/EQ, you can push the response down as low as you like. The only thing you loose below 20Hz is the ability to adjust Q to anything other than that of the shelving filter. Some shelving filters do have a small range of Q adjustment. As an interesting example of doing exactly this, anyone who's read about how I set up "thebland's" bDeaps saw reference to some very low frequency extension. For all practical purposes, what I did on his and other bDeap installs was execute a L-T as Dennis describes for the bDeaps.

[ThomasW]

There was a bit of miscommunication between Dennis and myself.

I thought he owned a DEQ and I was going to tap him for step by step instructions. Turns out he was relating how to make a 'generic' LT filter after playing with a DEQ owned by a friend.

Anyway, I have an empty Woodstyle small sub box in the basement and a few 12" TC2+ drivers. So I'll cobble up a sub, and use it to document how one goes about using the DEQ to create an LT for that particular system. It should be generic enough for anyone to copy the technique.

To keep costs to a minium I'll document the performance with John Mulcahy's software. But I won't be using a RS SPL meter ..... :wink: