Please help with freq dips on Tempest Sonotube sub (1 Viewer)

[Scott Page]

My newly completed Tempest sonotube has a significant dip in in-room response between 22 and 35Hz. Can you suggest anything to try to smooth the response?

Here are the numbers:

11.6-21.6Hz > 95dB

23.2 91.5

24.9 90

26.7 90.5

28.6 87

30.7 87

32.9 89

35.3-76.0 > 95

Thanks, Scott

 

[Bob Ahlberg]

First a couple of comments:
1. That really isn't a very bad ripple in freq. response. Fortunately, it's at a fairly low frequency and not up around 50hz to make your bass either lean or boomy.
2. It is most probably room related.
To smooth it out you have several options...most of which cost money.
1. The free one first, move the sub to different locations in the room until you find the one that has the least variance in freq. response.
2. Buy an EQ...either a 30 band unit or one of the purpopse made audio control units to deal with the bass response.
3. Some HT preamps have a built in parametric eq. My B&K ref 30 allowed me to smooth out a major dip between 35 and 50hz.
4. At the risk of sounding blasphemous, you could also play with the tuning, which would roll off response below 25hz, but smooth out the response up to 35
Good luck,
Bob

 

[Scott Page]

Thanks Bob.

I do plan to get a Parmetric EQ at some point. I also have a second sonotube under construction and am considering tuning it higher to see if that would fill in the gap somewhat, but I'm not sure if having two different designs is advisable.

Anyone have thoughts on that? This first one was tuned to 15.8, but has slightly rising output (in room anyway) to 11.5Hz and doesn't really drop much till below 10Hz. Think I should tune the second one to 18Hz or so?

 

[ThomasW]

This is a null, it's function of the speaker interacting with the room. If you move the speaker the null will likely change frequency.

The electronic cure is not to try and boost the null. It's to cut the frequencies around it to flatten the over all response.

Don't build another sub unless you need more output. Buy a Behringer DSP1100P or DSP1124P and cut the surrounding frequencies.

 

[Wayne A. Pflughaupt]

Scott,

I agree with Bob, this doesn’t look too bad. However, it’s hard to tell for sure because you did not give us readings all the way up to 100Hz or so.

Believe me, you don’t want two subs of different design. Can you say, “equalization nightmares?”

Regards,

Wayne A. Pflughaupt

 

[Brian Bunge]

BTW, if anyone's interested, www.musiciansfriend.com is selling both versions of the BFD for $129. Only the DSP1100P is listed on the site but if you ask for BEHRINGER DSP1124P FEEDBACK DESTROYER PRO item number 182467 then you can get the new one.
Brian

 

[Scott Page]

Thanks Wayne.

Here are all the numbers:

FrequencyCorrected Reading

572

5.475

5.777

6.283

6.685

7.189

7.691

8.191

8.793

9.495

10.095

10.895.8

11.699.7

12.498.5

13.3100.5

14.2101.5

15.3100.5

16.4101.5

17.6100

18.898.5

20.296.5

21.695.5

23.291.5

24.990

26.789.5

28.687

30.788

32.989

35.393.75

37.895.5

40.696.5

43.597.25

46.697

50.095.5

53.696.5

57.597.5

61.698.5

66.197.5

70.996.5

76.095.5

81.590.5

87.490.5

93.793

100.590

107.791

115.588

123.987

132.888

142.490

152.794

163.796

175.696

188.295

201.995

 

[Tom Vodhanel]

Scott, have you tried all the key listening positions in the room?(mic placement).

Also be sure to place the mic as close to *ear* level as possible for each seat you try.

Sometimes you'll see a moderate dip like this at one seat...but much more severe issues to worry about at alternate seating locations.

TV

 

[Wayne A. Pflughaupt]

Scott,
Your sub peaks at 16.5Hz?? WOW!!!
I must say, these readings overall are a bit strange. Primarily, there is no evidence of room gain—the natural rise in gain as the frequencies descend. (Okay, I guess you could say there is some –e.g., that peak at 16Hz. However, it is really to low to be of use.)
I have to ask, is this a really big room? What is the volume (i.e., cubic ft., including any rooms or areas opening to the HT room)? Do you have the sub in a corner? If you are planning to get an EQ, you won’t have to mess with moving the sub around the room looking for the best response. The inevitable compromise is a response curve you can “live with,” at the expense of maximum SPL level and extension (although the latter probably won’t be a problem for you! ).
If you haven’t already, put the sub in a good corner (pick the one with maximum uninterupted wall length in both directions) and take new readings.
However, based on the readings you’ve posted, here is how I would set up a parametric EQ.
Starting at the bottom, the first problem is the “hole” centered on 28.5Hz. Normally response dips are a problem, but since this one is not severe, I think we can try to fix it.
Set a filter for 1/3 octave bandwidth and boost 5dB at 28.5Hz. This will fill in the hole between 40 and 20Hz, and leaving a dB or two boost at 20Hz.
After EQ the numbers will be:
21.6 95.5
23.2 92.5
24.9 91
26.7 92.5
28.6 92 (filter center - 1/3 octave bandwidth)
30.7 91
32.9 91
35.3 95
37.8 95.5
The next problem is two slight humps centered on 43.5Hz and 61.6Hz. The former is a bit wider than the latter. Set a filter for 1/3 octave bandwidth and cut 4dB at 43.5Hz. This should result in the following readings:
32.9 91
35.3 94
37.8 93.5
40.6 93.5
43.5 93 (filter center - 1/3 octave bandwidth)
46.6 94
50.0 93
53.6 95.5
57.5 97.5
Next, the smaller peak centered at about 61.6Hz. Set a filter for 1/4 octave bandwidth and cut 4dB at 61.5Hz. This should result in the following readings:
50.0 93
53.6 95.5
57.5 94.5
61.6 94.5 (filter center - 1/4 octave bandwidth)
66.1 94.5
70.9 94.5
76.0 94.5
81.5 90.5
Here are the adjusted readings, 20Hz-100Hz. The response is ±5.5dB from 23-100Hz:
20.2 96.5
21.6 95.5
23.2 92.5
24.9 91
26.7 92.5
28.6 92 (filter center - 1/3 octave bandwidth)
30.7 91
32.9 91
35.3 94
37.8 93.5
40.6 93.5
43.5 93 (filter center - 1/3 octave bandwidth)
46.6 94
50.0 93
53.6 95.5
57.5 94.5
61.6 94.5 (filter center - 1/4 octave bandwidth)
66.1 94.5
70.9 94.5
76.0 94.5
81.5 90.5
87.4 90.5
93.7 93
100.5 90
Pretty good, but I think we can do a little better. Response is still a little high overall between 26-93hz. This is pretty close to 2-octaves wide, and the entire range can be easily attenuated by 3dB with a 1-octave filter centered on 50Hz. Now, it is unusual to overlay a filter on top of two others. Some might cry “phase shift,” but phase shift is not audible at low frequencies (indeed, it’s debatable exactly how audible it is at higher frequencies). In any event, I feel the improved response far outweighs any concerns of phase shift.
This should result in the following readings. Overlooking an isolated execption at 81.5Hz (which is too narrow to be of any consequence) this would give us an improved response variance of ±3.5 from 23-100Hz:
20.2 96.5
21.6 95.5
23.2 92.5
24.9 90
26.7 91.5
28.6 91 (filter center - 1/3 octave bandwidth)
30.7 90
32.9 89
35.3 92
37.8 91.5
40.6 90.5
43.5 90 (filter center - 1/3 octave bandwidth)
46.6 91
50.0 90 (filter center - 1-octave bandwidth)
53.6 92.5
57.5 91.5
61.6 91.5 (filter center - 1/4 octave bandwidth)
66.1 92.5
70.9 92.5
76.0 92.5
81.5 88.5
87.4 89.5
93.7 92
100.5 89
At this point is that we have achieved very flat response, but that does concern me. The problem is that flat response sounds “thin” and unnatural. What is missing, as noted above, is the (usually) natural room gain. This could be EQ’d in, but I’m hoping better placement of the sub will give it to us for free.
I’m also curious, Scott, what are you powering this with?
Regards,
Wayne A. Pflughaupt

 

[Scott Page]

Wow Wayne! Thank you so much for all your effort. The sub also has the same peak at 14.2hz! From your numbers I've learned something that I was curious about. It appears that the boosts and cuts from an EQ are not straight across but "sloped" from the filter center, which I take it makes it blend in better than a rigid boost/cut across the bandwidth. I was wondering about that.

The sub is a 12 cubic foot sonotube (24" dia) with an 8" top-mounted port about 36" long (tube is 35" but is inset about an inch below the top plate which is cut out 12" in dia to flush-mount the grill. Design is Adire EBS design. Tempest is bottom mounted (flush). Wired to 4 ohm's and driven with one channel of an Adcom 545 (150W into 4 ohm). Tune was supposed to be 15.8Hz as per Adire design. I know that is a little underpowered but will have to do for now.

Sub is in the left-front corner about 8 inches from the left wall and about 4 inches from the back wall. Room is about 16.5 wide and 25 long with ceiling average of about 7, so about 2900 cubic ft. A hallway to bedrooms opens up midway along the right wall. The crossover from my Denon 3801 reciever is at 80Hz. All speakers are set as small.

The numbers are from warble tones. Maybe I should use sine waves instead. I did do some early testing with sine waves and the numbers seemed simular that I recall. All numbers are from ear position at the center seat. I will have to do some testing at other seats. Also, when I get my second sub finished, that will require a new round of testing.

Yes the sub really does extend low. You don't hear it but I must attest that output from 5 to about 18 hz might not be heard but it does have an effect. I felt really funky will testing the sub even though I couldn't hear anything. I guess that must add something to the realism of explosions and so on.

I will have to get a Feedback Destroyer as soon as I can. Not sure why no room gain. But I don't really know what room gain looks like. I guess I thought that the rise in output to peak at 14 and 16 hz was room gain. But you say it is too low. What should it look like? I'm glad to know that EQ can improve the situation. Not sure how to add room gain? Should I be trying to get what I hear is called the "house curve" response? Would changing the sub channel level up or down add to room gain or help with the 26-93hz problem? I do not have the sub channel level set high. In fact it is at -2db's. I know that doesn't mean much, but using AVIA, it is set about equal to the front left. No increase provided for meter compensation (not sure if AVIA tones already compensate for Radio Shack meter attenuation?). Not sure if increasing overall level would increase response across the board or would it have a bigger effect on the lower frequencies? Anyway, with sub level higher, it sounds a bit fake to me.

 

[Wayne A. Pflughaupt]

Scott,
Happy to help. Hmm, maybe I should start an “EQ your sub” consulting business…”

I do not have the sub channel level set high. Anyway, with sub level higher, it sounds a bit fake to me.

This is what I would expect. As I noted, flat response sounds unnatural.
Regards,
Wayne A. Pflughaupt

 

[Jack Gilvey]

I just wanted to thank Wayne for taking the time to provide such informative and detailed answers. It obviously took some time and thought to do so, and I personally appreciate the effort.

 

[Wayne A. Pflughaupt]

Thanks for the kind words, Jack

 

[Scott Page]

Thanks again Wayne. You should write an EQ your sub FAQ. I appreciate all your work on this but I'm starting to feel a little guilty at the amount of effort you have extended. You really know your sub EQ stuff! I'll try to answer a few of your questions.
The single sub does put out plenty of bass. I had no idea one would do so well and with only 150 watts. I've already purchased everything for its twin including the driver, and it is half constructed, so I'll have two anyway. My wife has been pretty good about the huge round black refrigerator in the HT (two might be another story :frowning. It fades into the dark pretty good. Of course TWO of them in the same corner might be a bit overwelming! I was planning to put them on different sides of the room, but have since heard that is a no no. She didn't think we needed a sub at all until I hooked up the Vandersteen 2W we had. I sold it (kept bottoming out) to fund the sonotube subs and now she can't live without one. I might sell one in the future. Those Tempests really put out.
The windows weren't open during testing but the HT is in a basement of 1850 square feet with an open hallway off the HT and a door to the upstairs off of the HT (slightly open during testing because of ratteling at certain freq). So that might limit room gain I guess. I'll have to retest with all the doors closed tight. Maybe the top mounted port also produces less room gain than a bottom mounted port due to the extra space around it.
I do plan to upgrade the amp at some point. I'd like to get to 500-750 watts per sub with a Pro-amp. But that will have to wait awhile. I'll get a Feed-back Destroyer first and some other goodies.
The mains are away from the back wall about 4 feet, slightly in front of the RPTV. They are about 3.5 feet from the side walls. They are full-range Vandersteen 3A's. You are likely familier with these towers, but if not, they provide a good amount of bass vol and extension, but I have them designated as small as per recommendations of many on line.
Does anyone reading this thread have any recommendations for an inexpensive EQ other than the BFD that might work better for this situation? Does anyone know if the BDF does the shelving thing Wayne talks about?
I wonder if more power to the Tempest will improve the room gain thing by increasing the output down low? Has anyone tested the effects to the response curve of more amp power? Maybe Dan Wiggins could chime in here if he sees this? (Love the Tempest Dan, its a 5 star product).
Thanks to all for your help. I would not have built this sub without all the great info on this site. I would have bought an SVS instead (sorry Tom).

 

[Richard Greene]

To Scott Page:

Use a slow sine wave sweep test tone to measure the actual frequency response at your listening position including

room resonances.

I recommend the Stryke audio Basszone CD with its five minute long sweep from 1Hz. to 100Hz. (please turn down the volume for the first minute, especially with a ported sub) You can roughly estimate the center frequencies of room standing waves using a stopwatch (each minute of the sweep covers a 20Hz. range)

Warble tones won't excite room resonances much -- they will reduce frequency response variations by at least 50%

(compared with using a slow sine wave sweep or lots of individual steady sine wave tones).

It will be a rare subwoofer that measures better than

+/- 10dB using a slow sine wave sweep tone -- if you use a parametric equalizer to reduce peaks, set your goal as

+/- 5dB (but you may never get that tight a frequency response -- I have never measured better than +/-6dB myself

using a slow sine wave sweep)

Frequency peaks above 40Hz. will be most audible because they will be more frequently excited by the music content -- One reason: the common four-string bass guitar goes down to 41Hz. so it cannot excite room modes at lower frequencies

Also, don't trust measurements below 20Hz. -- you are probably measuring quite a bit of harmonic distortion at higher frequencies and "room noises" -- a microphone can't tell the difference between the primary tone, harmonic distortion at higher frequencies and room noises.

If you insist, measure deep bass at very low volumes to limit the harmonic distortion level.

You may want to keep your Tempest driver at least 8" away from all walls -- my own Tempest tube sub really rattles the

nearby walls at 18-20Hz. even with two heavy concrete slabs leaning against each wall in my room corner.

Your room has potential room gain of +12dB/octave below 23Hz. -- that's an estimate of your lowest frequency room mode (caused by the 25 foot long surface-to-surface reflection) -- that's a potential of +12dB at 11.5Hz.

(your ported sub probably does not have enough output below the tuning frequency, so the room gain may not make much of a difference -- also, the threshold of audibility at 10Hz. is about 95-100dB) Room gain is an acoustics phenomenon -- it's affected by open doors, room openings, open windows and the flexibility of room surfaces -- not by your subwoofer.

Even the subwoofer's location will not affect room gain

because for frequencies below the lowest frequency room mode (566/largest room dimension in feet = lowest room mode in Hz.) sub location is not relevant.

To Thomas W.

Moving speakers will not change standing wave frequencies

-- the frequencies are based solely on room dimensions,

not speaker location.

To Wayne P.

Equalization is used to reduce standing wave peaks

-- these peaks are related to room dimensions and

not to subwoofer tuning frequencies -- that means using

two subs with different tuning frequencies is no big deal.

However, if you use two subs and place them in different locations, there will be destructive interference that will cause a hole in the frequency response -- that hole can not be equalized away.

Equalization can also be used to extend bass response of subs ... but not if the lack of output is caused by room standing waves. Trying to boost a standing wave minima

(aka "null") will overdrive the amplifier and accomplish nothing else.

Room gain and the house curve are not the same thing.

The house curve refers to treble roll-off needed for a natural sound quality in a specific venue (flat treble is too bright for most people) that is often obtained using equalization.

Room gain (or cabin gain) refers to bass ramp up below the lowest room mode caused by room pressurization.

To Brian Bunge.

$129 for a Behringer Feedback Destroyer = incredibly cheap

 

[Scott Page]

Thanks Richard.

Much to think about.

 

[Wayne A. Pflughaupt]

Scott,

I wonder if more power to the Tempest will improve the room gain thing by increasing the output down low? Has anyone tested the effects to the response curve of more amp power?

No, it won’t affect room gain. That’s a function of the room. However, increasing amplifier will get greater SPL if the speaker is under powered. Never anything wrong with that.
Happy Holidays,
Wayne A. Pflughaupt
P.S. I was only kidding about the open window thing. I know you’re smarter than that!

 

[Scott Page]

P.S. I was only kidding about the open window thing. I know you’re smarter than that!

Well, to be honest, I'm not, at least about this stuff. I did know you were kidding. But this IS my first HT, and the first sub I have had (other than the Vandy I had for a couple of months), and my first DIY speaker. So the sub testing is all new to me. So I'm not sure what effects to expect from open doors, spaces under doors, open hallways, and so on other than: larger room = more power/sub needed for same effect.

Thanks again to everyone for their generous help, especially Wayne! What a great forum and a great bunch of people.

 

[Jack Gilvey]

Wayne,

Not to derail the thread, but that Ashley 571 looks very interesting, particularly the LF shelving filter. It's tough for me to picture the curve it would apply, though, it just specifies "40-400Hz". Would that be where the "knee" in the curve would be (where the "shelf" portion of the boost appears), or is 40Hz where the boost begins? Sorry if that's unclear.

Thanks.

 

[Wayne A. Pflughaupt]

Jack,

Just like with the regular filters, you can select any frequency between 40-400Hz as the turnover frequency. I know it’s hard to tell because the PQX-571’s front panel is really cluttered (they pack more into this parametric than any other I know of--5 fully parametric bands, plus high and low shelving filters--7 bands total).

For the benefit of anyone who might not know, “shelving” means that all frequencies above (high shelving) or below (low shelving) the selected turnover frequency are affected by the filter. (Typically the bass and treble controls of a receiver or pre-amp are shelving filters.) As with regular EQ filters, the selected (i.e., turnover) frequency is the “peak,” and response gradually “ramps up” to the turnover frequency, and remains flat beyond that point rather than “ramping down” again.

The actual “curve” will depend on how much boost or cut you apply. It does affect a pretty broad area, however, considerably beyond the selected turnover—at least this is the case with the Ashly PQ-16 model I have in my bass guitar rig (which is the 571’s predecessor). I have the high-shelving band set for 9kHz, cut 6dB, and it is affecting down to and beyond 3kHz which is about 2 ½ octave down. They may have changed the parameters on the 571; can’t tell you for sure, because I’ve never used the new one.

However, you for sure would want to set the low shelving at 40Hz for a home theater sub. You would probably be able to boost 4-6dB at that point before affecting frequencies at the crossover point (assuming a crossover of about 100Hz).

Hope this helps!

Happy Holidays,

Wayne A. Pflughaupt