Sealed Shiva for a small room, is this make sense ? (1 Viewer)

[Jones_Rush]

I want to build a sealed sub using the shiva driver.
My HT room is small: 13x9x8feet.
Since the largest room mode is 13feet, I will get room pressure below 43hz (566/13) 12db gain per octave,
which mean that I will have +12db gain at 21.5hz.
I need to build the sub so it will have -12db at 21.5hz, so I will have "flat" frequency response, this mean building a sealed box (for the shiva driver) which has about 45 liter volume (according to "WinISD" that will give me -12db at 21hz).
Am I making sense ? I mean, expecting flat frequency response (except for the room modes) down to 20hz with only 45 liter design ??? can it sound good ?
In theory, going with a larger box will give me boomy sound, since I will have too much pressure (room gain) at low frequencies which will make the bass intolerable...

 

[Greg Monfort]

Unless you've measured this much gain, I wouldn't assume it, as it requires a well sealed room with stout construction. 3-6dB is more realistic.
Another thing to consider is the Qtc of the sub under power. This small a sub, even stuffed, will have a fairly high Qtc in operation, so group delay will be worse than a vented design.
Generally speaking, modeling a low Qtc with F3 = ~1st mode works well. ~262L stuck in a corner looks pretty good on paper.
GM
------------------
Loud is beautiful, if it's clean

 

[Tom Vodhanel]

Right, 6dB octave seems to be about the average--but it depends on the openings/window area of the room...and the overall construction.
BTW---a slight boost

 

[Jones_Rush]

Greg,
Tom has a 20x14x8 foot room with one large picture window and one 5 foot opening into the room and he measured about 5dB of gain at 20hz. My room is 13x9x8 foot, with NO openings at all (both the door and window are closed during listening sessions), so my gain at 20hz must be over 5db!
maybe not 12db, but might be 8-10db (my room construction is not very sound proof, but still, we are talking about brick and cement walls, not papers!.
12" Shiva with a sealed 45 liter enclosure might be too small. Adire Audio recommends 2 main sealed choices:
1) Q = 0.707 , 54 liters design.
2) Q = 0.6 , 88.5 liters design.
Can you guess what will sound better in my small room ?

 

[Greg Monfort]

So the walls are solid cement, with a brick exterior facing? No voids, no drywalled studded walls to dissipate the energy? The door is solid hardwood, with weatherstripping? The window is small?
The fact that you don't consider it very soundproof pretty much tells the tale.
Anyway, all I can do is speculate based on personal experience, and the most gain I ever measured was ~11dB/octave in a converted 50's bomb shelter. Talk about room modes!
I still think the larger box is the way to get the flattest response in-room based on what you've told me, as even the larger box sims a ~3dB roll-off, so any extra gain will pump up the extreme bottom end. But if you want to use what Adire recommends, I would go with the 88.5L, and then EQ it as required/excursion permits.
GM
------------------
Loud is beautiful, if it's clean

 

[Rick Steverson]

I suggest you may wish to simulate your Shiva in-room using LspCAD, a design package setup you can get free from Adire Audio at http://www.adireaudio.com/software.htm
I don't disagree with the suggestion that 12db/octave room gain is high. You can look for a thread here from Julian Data, if I remember correctly, who compared the in-room simulation with his actual reading and got really close results.

 

[Jones_Rush]

Greg,
Here are the Anechoic measurements Adire measured for the 88.5 liter,Q=0.6 sealed design(with 16 ounces of polifill):
Fb : 34.4Hz
Anechoic F3 : 30.7Hz
Anechoic F8 : 20.5Hz
According to these measurements, -8db will happen at 20.5hz,
since my room can surely give ~ +7/8db gain at 20hz, we are talking about flat response.
Do you think I should still opt for the ~262L enclosure ?

 

[Tom Vodhanel]

I just gave what my room measured at as a *FYI*...as Greg notes...it's nearly impossible to predict the exact amount ot *gain* you'll get.
Also, fwiw---I agree with Greg on the enclosure suggestions...build the biggest option you can.
in the end---your response is NOT going to be (minus)1/(plus)1 regardless, so worrying too much about a gain difference of 3-5dB at 20hz...could be much ado about a little.
TV

 

[Jones_Rush]

Rick,
I tried to use LspCAD, but something is VERY wrong with it,
the parameters for the Shiva 2 are completely different than anything I have ever seen. I uploaded these parameters to WinISD, and look what I get:
Driver : Shiva 2
Vas : 151.0
Qts : 0.73
Fs : 21.00
SPL : 84.20
--
Number of drivers : 1
Box type : Closed
Box size : 515.8 l
Relative
Freq Gain SPL
[Hz] [dB] [dB]
20.00 -3.54 80.66
25.00 -1.24 82.96
30.00 -0.21 83.99
35.00 0.19 84.39
40.00 0.33 84.53
45.00 0.35 84.55
50.00 0.34 84.54
55.00 0.32 84.52
60.00 0.29 84.49
65.00 0.26 84.46
70.00 0.23 84.43
75.00 0.21 84.41
80.00 0.19 84.39
85.00 0.17 84.37
90.00 0.15 84.35
95.00 0.14 84.34
100.00 0.13 84.33
I don't know what it is, but this is NOT a Shiva!!!
This doesn't even seem like a new Shiva prototype. NO company will create a driver that needs over 500 liter of volume for Q of 0.73, this is crazy!

 

[Jack Gilvey]

It's ok, relax, take a breath...
It looks like the default value for the Shiva in that program is with one coil connected,one open, doubling the Qts. Can you choose an option to parallel them?
An interesting option is the use a potentiometer across the unused voice coil, varying the Q of the final sub and "tuning" the response.
------------------
Link Removed

 

[Greg Monfort]

Since small rooms need either an IB or EBS sub IMO, the 262L is in most cases the practical physical limit for an in-room cab, so my suggestion hasn't changed.
WRT LspCad, you have to select how the VCs are wired in the Driver Unit/Configuration window to get the lower Qts.
Also, to get a better idea how the sub will respond under power, you have to list max peak power available in the Measurements Setup window, and check the three boxes in the Additional Analysis section of the Driver Unit/Configuration window. The first thing you'll notice is that for a given cab volume, the Qtc will go up considerably, and why I chose the larger volume. Using programs like WINISD are fine to make quick driver comparisons, but as TV noted are next to worthless for much of anything else.
BTW, for the flattest in-room response without EQ, you will need a lower Q system, and why an IB would be required since the cab would be so large. If an IB is used, then JG's observation about only wiring one VC and fine tuning the Qts with a pot is spot on IMO, but I didn't suggest it because with only one driver, the up to 6dB loss of peak SPL is considered unacceptable for HT.
GM
------------------
Loud is beautiful, if it's clean

 

[Rick Steverson]

Jones,
I don't know the in's and out's of every part of the LsdCAD software, but I do know that there are somethings that don't work the way we may expect them. A case in point; the SPL, cone excursion, port airspeed can be simulated at any power you like with the measurement setup dialog box. The program uses the source voltage to calulate power. Every first time user enters their amp's power rating, which is ok except when the VC is connected in parallel. Then, we have to enter 1/2 the rated power to correct for the 4 ohms speaker load instead of 8 ohm that's assumed.
Anyway, Dan Wiggins, president of Adire Audio has preloaded the specs to work correctly with LsdCAD. He uses this software himself and know it well. It's got a lot going for it, including the ability to simulate in-room response. Try using the LsdCAD Box Optimize feature, with its table lookup set for a .707 Q and it will return 64 liters. Use WinISD with Qts=.39, Adire's published spec, and you will get close to the same. I don't know why Dan Wiggins has to use Qts=.73, but it will return the right simulation.

 

[Rick Steverson]

By the way, the thread that Julian Data posted about real world vs. LsdCAD simulations is http://www.hometheaterforum.com/uub/...ML/002288.html