Tempest Woes

[richie00boy]

PS- In the case of open baffle (dipole) , does it means that the woofers are overdamped?F

Only if you stray below the target Q of 0.5 recommended by Linkwitz. If your woofer has Qts below 0.5 you can use a Linkwitz Transform to bump it up to the desired value.

 

[jbandpc]

Thanks!

Thanks for the great discussion!

If poor integration can be preceived as loose bass, then I'm sure that's it. My speakers are in odd places (spouse thing).

I have a radio shack meter, so I'll test it out. If I find I have a room mode or two, I guess I'll try to get the Feedback Destroyer or the Ultracurve. I can't move the sub (spouse thing). I can't re-build the sub (lazy thing).

Thanks!

 

[Tim Moorman]

Crossover

Just a quick final comment:

The midbass is now working down to 60 Hz, as you stated previously. You might try adjusting the plate amp crossover upwards to the 80 Hz THX standard, or even higher to see if the mids improve, and the bass tighten up some.

The specs on the amp show considerable effort to optimize the crossover slopes for better integration, so that should not be an issue. The crossover point may be, however.

I'd also try every conceivable sub location and crossover point to isolate the problem to either the room or the box construction. If the box is shaking, clamp up some 2x4s or 3/4" panels on it to see if that helps. If it doesn't, the thin sheet goods won't help.

Tim

 

[Luke]

cheap SPL meters

Ive looked at cheap SPL metes, and they typically have an accuracy of +-3dB. Is this good enought to use for speakers?

 

[Tim Moorman]

Cheap Sound Meters

In a room, you can have swings of 10 - 12 dB from a speaker with even response, particularly in the bass region. Anything that helps, even the cheap unit. Plus, there are re-calibration sites on line to make them accurate.

Tim

 

[paulspencer]

Re: Cheap Sound Meters

In a room, you can have swings of 10 - 12 dB from a speaker with even response,

I'd say that's in many cases a big understatement! You can easily get 20 db

 

[Luke]

If a room has 20db swings then how can measuring response be a usefull tool for speaker design.Surely calibrating something that is not linear is a bit of a waste of time isnt it?

 

[Tim Moorman]

Sound Meter

Luke,

You still need to identify problems to correct them, from the speaker design or from the room.

Not all rooms are bad, but bass has a way of exciting all manner of spurious nodes and resonances due to the long wavelengths (20 Hz=56+ft). Sometimes just moving the sub will have a quite a dramatic effect. Room corners really reinforce output, or if you are seated against a back wall the bass will gain significantly.

Why abandon a design when the room will not allow you to hear it properly?

Tim

 

[kelticwizard]

Just to clarify things, I posted some charts from Small, (of Thiele-Small).

First, this is the response chart for various values of Qtc in relation to Fc, the response in the closed box.

At Qtc = 1, the response is right on the midband.
At Qtc = .707, the response is -3 dB from the midband.
At Qtc = .5, the response is -6 from the midband.
http://www.diyaudio.com/forums/attachment.php?postid=88030

Next, I posted the step response for each value of Qtc. Before you go overboard, remember that even Qtc = 1 gives significantly better step response than any vented box you are likely to run across.
http://www.diyaudio.com/forums/attachment.php?postid=88711


And here are the step responses for vented alignments. Remember that a = (Vas/Vb).
http://www.diyaudio.com/forums/attachment.php?postid=366557

And here is a thread where we did all this before, plus some interesting charts by Rob Wells on Qtc interacting with room gain:
http://www.diyaudio.com/forums/showthread.php?threadid=8114&highlight=Qtc+tightness

 

[jonz]

I have 2 tempests subs in a sonutube configuration...they are both around 155 liters or 5.7 cu.ft. sort of mid to high Q...I am powering them with a crest 2600 amplifier in a 4 ohm config..at 750 watts per....I realized they don't need all that power but that was the only amp I have presently.

I have them postitioned on opposite ends of a couch ..one is in a corner and the other is not....The one in the corner sounds completely diffrent than the other..to the point that I have to lower the volume on it by about 2db .I have experemented and everytime I move them they sound diffrent..

I am very happy with them...to me they sound very deep..and tight...I realise those two words can have a different meaning..to every listener...

 

[Tim Moorman]

Tempest Differences - Sealed

kelticwizard,

Agree with all posted with the exception of the first chart, which gives the impression that the lower "Q" tuning of .5 will roll off faster than the higher "Q" tunings. You have stated the same. The roll off is -6dB at the tuning frequency, but a much lower tuning frequency is allowed.

Adire Tempest:

Q of 1 yields a sealed box of ~1.7 ft^3 (?) with 54 Hz Frc
Elevated response of 2dB at 75 Hz to flat at the tuning frequency. Down 9 dB at 30 Hz; down 17 dB at 20 Hz.

Q of .55 = 12.4 ft^3 (!) box with 27.6 Hz Frc (down 5 dB).
Down just slightly from midband at 75 Hz. Down 4.5 dB at 30 Hz; down 8.6 dB at 20 Hz.

The lower Q yields a slope of about half the drop in amplitude as the 1.0 tuning, not twice the rate. Box size naturally increases as Hoffman stated at ~8 times/octave.

Phase and group delay slightly better with low Q tuning, but power handling drops significantly at lower .55 Q to about 40 % of rated at 20 Hz.

Tim

 

[paulspencer]

If a room has 20db swings then how can measuring response be a usefull tool for speaker design.Surely calibrating something that is not linear is a bit of a waste of time isnt it?

You can't design a speaker to compensate for such dramatic peaks and dips. Most would deal with speaker design and room issues separately.

However, in this situation I suspect that room issues are more significant than sub design issues that relate only to the speaker itself (box, Q etc), to the point that the room issues are likely to overshadow everything else until they are fixed.

You can lessen room modal problems with placement and treatment. You can then smooth the response further with eq "calibration." The best bass I have heard is in a system where this was done. In my experience a low cost sub can outshine a cost-no-object high end design if the low cost sub is integrated properly and the expensive sub is not.

 

[paulspencer]

Jonz,

Nice setup. The way you have your subs set up seems to me the best for most rooms as far as room modes go. Also they are tables as well so they integrate well with the decor and are practical.

 

[Luke]

You can lessen room modal problems with placement and treatment. You can then smooth the response further with eq "calibration." The best bass I have heard is in a system where this was done. In my experience a low cost sub can outshine a cost-no-object high end design if the low cost sub is integrated properly and the expensive sub is not

So would you recommend something like rods http://sound.westhost.com/project84.htm to get the eq right? Will this help iron out the rooms issues? If you cant rely on the room in the first place how do you know the woffer is ok and issues are room related?

 

[paulspencer]

You can isolate room issues by taking it ouside.

I don't think the ESP eq is sophisticated to really deal properly with room issues. Parametric eq is needed. Like the Behringer Feedback Destroyer, often discussed on this forum. It's quite cheap.

I was intending to build that project but after looking into it more I've concluded that you can't match the power of units like BFD or Ultracurve and they don't cost that much. When you hear a well set up sub it is very impressive. There is a world of difference between a sub placed indiscriminantly in a room and a sub that is well placed and correctly calibrated and set up.

 

[directservo]

Grey,
You bring up several good points.

The bass will not get appreciably tighter. Assuming that you mean what I mean when you say tighter (you'd be amazed how many people say tight when they mean deep), you're pretty much stuck with the sound that you've got now. I'm not saying that it won't change at all, just that the "tightness" isn't going to get better just by stiffening the cabinet.

Another type of tight bass is from small size bookshelf speakers that have high Q and limited bass extension. BTW, this type of bass is fake. The reason it sounded tight is because it filters out all ambience information and it almost sounded artificially clean. Having a good bass extension runs against having a tight bass. "Articulate bass" is more appropriate in the case of good bass extension.

Okay, now on to "extension." I've said it elsewhere, but people keep buying these drivers and ending up dissapointed, so I'll say it again: They ain't flat. Period. They have a broad peak in the 80-100Hz range and slope off at about 6dB/oct south of that. The end result is that your 20Hz response is a good 12-15dB lower

The cause of the hump at 80-100hz is the high voice coil inductance (or to be technically correct, semi-inductance) . This is not just with this particular 15" driver, several other 4-layer, 6-layer voice coil drivers also exhibit similar characteristic. People didn't realize because they never put the inductance into computer simulation. To correctly simulate for these humps, do not put in the inductance value published by manufacturers because most likely they are measured at 1khz. Instead, put put in a value that will produce the correct impedance mimimal point around 80-100hz as in the published impedance curve(that means you also need to have the impedance plot of the driver).


Brian D.
Rythmik Audio