Salerno Subwoofer: SI HT18D4 11Hz LLT... times TWO!

Salerno Subwoofer: SI HT18D4 11Hz LLT... times two!

Building has begun!

I am building two subwoofers, each of the following design:

Driver: Stereo Integrity HT18D4
Box Volume: 22.06 Cubic Feet / 625L (24Cu Ft raw volume minus 1.94Cu Ft for bracing, port, and driver)
Port: 6.25" x 24" w/ 1.5" flare on both ends
Resulting tune: 10.9Hz
Material: 3/4" MDF
Power: Behringer iNuke3000DSP (possibly two)
Estimated weight: approx. 300lbs... EACH!

These subwoofers are named after Forward Operating Base Salerno where I was deployed when I initially designed these beasts. I name them after FOB Salerno because at this location, infrasonic bass was not uncommon, be it from one of our helicopters unloading missiles into the nearby mountain, the howitzer being fired, or the occasional rocket attack courtesey of the Haqqani Network. Initially, these were going to be two sonotube subwoofers. However, I've encountered difficulties in sourcing the sonotube in Germany, and even the German home theater forums have people who either can't get it, or have to pay a lot to get it. And so, the design changed to a box design.



These subwoofers are being designed and built for home theater use, although I expect they will also help music to sound a bit fuller also. The design goal is to achieve flat frequency response to 10Hz at reference levels, although I hope room gain will get me below 10Hz. At the very least, I hope to have meaningful output into high single digits.

The rest of my speakers are Klipsch RF-7 IIs, RC-64 II as LCR, RS-62s for side surrounds, and RB-81s for back surrounds. And so to start, I'll probably cross over my speakers at 40Hz for the mains, and 60Hz for the center and surrounds as a starting point. And so these subs won't be expected to output much above 60Hz.

Due to budget and build area restraints, the subs are being built one at a time. I hope to have the first one operational by 15 Mar, and the second one by 15 April.




Edited by DanLW - 3/14/13 at 2:54pm

DIY Circle Jig


The Drivers


The obligatory child in the driver hole picture


That looks scary, daddy!


Gluing the baffles together


Finished baffle. The blocks around the driver hole secure the T-nuts in place so that they will never (in theory) come off.


Hard at work...


MDF is Dusty!


Test fitting the sides.


CLAMPED! The top is not being glued on in this picture. The weights are to hold the sides to the base.


Like father...


Like son.


Bracing! All the pieces are glued to each other where they cross to help obtain maximum rigidity. You may notice the white things to the right of the long braces. This is paper folded over several times with glue between each layer. My 48" braces somehow ended up just a tad short.


Stuffing.


Another angle of the stuffing.


Ready to be glued!


"Clamps". At least those weights are being used for something...



Almost looks like something out of Lord of the Rings! On the outside, there were several runs. However, I was prepared with a damp wash cloth. Immediately after weighting the top, I wiped off all the runs on the outside so that I wouldn't have to worry about scraping/sanding them off later. Yes, I did use "excessive" glue, but I wanted to make sure I had a good glue bond all around, and on all the braces.



The subwoofer with carpeting. A grill will be added soon...


A wide shot of the back of the theater.


A wide shot towards the front, taken from atop the sub.


A detailed shot showing the system I used to keep the T-nuts in permanently.


The inside end of the access panel.


The bracing system for the second sub. This one should be more mechanically secure in that the vertical braces were cut so that they would go inbetween the horizontal braces.


A closer shot of the internal bracing.


The "Third Row Riser". Two Salerno Subwoofers nestled side by side.

Edited by DanLW - 4/5/13 at 12:38pm

16 March 2013

First measurement: One sub powered with PAMP1000, highest setting without exceeding XMax, One doorway into the room open. No EQ.



17 March 2013 (inaccurate - <30Hz is ~15db hot)

10Hz THD measurement @ Listening Position (Rectangular Window)


10Hz THD at max safe output before distortion (Hann Window)


More THD measurements



Subwoofer response from 10-1000Hz at three positions. 1m from the slot (pink), right in front of the slot, and handheld ~6 inches from the center of the cone (gold). The purpose of these measurements was to look for resonance issues. Looking at them, I notice that the measurement furthest away seems to have a lot of room influence, as evidenced by the dips at about 24 and 58Hz which aren't in the other measurements. As for the other stuff, other experts here will have to interpret for me. Right at the speaker, which presumably has the least room influence, the plot looks pretty good, with the first major swing at 70-80Hz. It looks to me like a +/-4db swing. While I'm sure a measurement of this beast in the middle of an open field would be very interesting, this sub is NOT going anywhere until I move!


I thought that my Klipsch RF-7 IIs sounded better in "Pure Audio" mode, and now I see why. I'm having some sort of issue integrating the subs into my system. There's a huge response hole from 60 to 90Hz. I did have my front XO set to 40Hz. As you can see from the "Pure Audio" trace, the RF-7s have absolutely no problem playing flat down to 30Hz. But once I switch my receiver to "Stereo", which activates the XO, all sorts of weirdness happens. 60Hz is where my XO is set now as that is the "best" response I could get, but it's still not as good as "Pure Audio", other than the fact I don't have any real output below 30Hz. Ideas, anybody?


18 March 2013

LR Front Speakers with receiver set to "Pure Audio"


It was suggested that I try changing the phase of my sub by 180 degrees. These two graphs are with the sub at 0 and 180 degrees. Looks like that wasn't it.


And so then I decided to try running Audyssey. Much better! Red is with DEQ off, green is with DEQ on.


Before and after Audyssey.


Subwoofer only sweep (soundcard output direct to sub amp input) at listening position.


Sweep with the meter held about 6" from the center of the cone.


Pure Audio with laptop (red) vs HTPC sound card line in (green). Red trace is also with Rat Shack meter cal enabled.


Subwoofer laptop vs HTPC sound card line in. The laptop trace (the higher one) was also done with the Rat Shack meter cal enabled.


As far as the laptop vs built in sound card, the biggest difference was probably the use vs non-use of a cal file for my SPL meter. Also, the meter was not in the *exact* same location for both measurements. There is at least a few inches, but less than a foot difference in mic placement.

So looks like I am actually perhaps 15db lower than where I thought I was. Makes me want to bet the second sub built all the more!

Also, I should be getting my iNuke3000DSP in sometime this week. Once that gets in, I'll be doing another round of excursion tests, applying a high pass, and will do more sweeps.

5 April 2013

These measurements were taken using a Radio Shack SPL meter as a microphone. No microphone calibration file was set in REW, per the advice of some in this forum. A sound card cal file was in place. The output of the meter was fed into the line input of the sound card integrated into my HTPC. (VIA VT1708S, which is the sound integrated onto my Asus P8B75-M LX Plus motherboard)




The average response of the front and rear row listening posistions.


The lower sweeps are subs A and B individually. The highest sweep is both subs working together. Combining them indeed does add 6db of SPL.


Effects of the "Phantom 10Hz HP filter". The filter was crated by combining a 20Hz -12db/oct HP filter and a -12db HS filter set at 20Hz.



Phase comparison. Setting either sub to 180 degrees out of phase completely trashes the LF response.


Two sweeps at -5 and -10dbFS. These were done with the sound card output connected directly to the iNuke 3000 input.


This chart shows my room gain. The blue curve is a near field measurement taken from about 1 foot from the driver. The gold curve is at the listening position. Levels were adjusted between tests to compensate for the db loss associated with moving away from the subwoofer.

Edited by DanLW - 4/5/13 at 12:53pm

Right now, I have one big question. What do you experts think about the stuffing? It's mineral wool sheets stuffed as a block into the center braces of the box. I put the braces every 8 inches, so it's basically an 8" x 16" x 24" block of mineral wool in the center.

I put it in the center because I read on another forum that for vented, it's better to put stuffing in the center because it will better eliminate standing waves at this location, rather than if it was lining the sides. As it is now, the sides are unlined, and there is at least an 8" air gap on all sides of the mineral wool block. Does this look like a good solution, or should I change? I haven't glued the top on yet, but that is the next step. Once the top is glued on, I can modify the stuffing somewhat, but it will be much more difficult at that point, and there will be some areas within the box which are simply unreachable because of the bracing.

Thanks! The bracing took 3 or 4 days to complete. For some reason, I couldn't find my 4th quick grip clamp, so I could really only do one at a time. It was a process of fit, glue, wait 20 minutes, and move to the next one. In many cases I had to sand down the end to get the brace to fit, but that's better than it being too short. I purchased a 4th quick clamp, and also got a couple little C Clamps, so I should be able to brace the next one much more quickly. There's a lot of little lessons I learned in building the first, which should make building the 2nd go much more smoothly. I'll have to compile a list of lessons learned and add them to one of the top posts.

Can anybody confirm or deny whether my approach to the stuffing is at least theoretically sound? From what I understand, putting it in the center will remove standing waves more efficiently than lining the sides.

Update: I just glued the top on tonight! (See last 4 pictures in the picture post) I'll let it dry overnight, and tomorrow I hope to do some initial measurements. I don't have the iNuke3000 yet, but I do have a Pyle Audio PAMP1000 which will do an advertised 160W into an 8 ohm load in bridged mode. So while it won't test the sub to it's maximum potential, it will be at least enough to do some preliminary measurements, and see how the driver behaves. Theoretically, I should be able to hit 109db at 50Hz, 105db at 20Hz, and 102db at 11Hz. Hopefully room gain will boost my lows so that it's a bit flatter. We shall see...

Well, here it is, the first measurement!



The amplifier I'm using currently is a Pyle Audo PAMP 1000 in bridged mode. According to the specs, it delivers 160W into 8 ohms. The amplifier was suprisingly capable, and seems to have pretty darn flat response into single digits.

The above measurement was done at the loudest setting I could attain without going over X-Max. However, I am reaching X-Max below 10Hz, so I should be able to go louder once I have my iNuke3000 with a Shelf/Highpass filter set for 10Hz.

I also did excursion testing using the DIY Audio Test DVD's -.5dbFS test tones. First, I calibrated my sub using my Receiver's built-in test tones, and verified with the -20dbFS test tone on the DIY Test DVD. Next, I played the 1Hz -20Hz -5dbFS test tones. When I did this, I would keep my amp turned down, and would slowly turn it up to the setting which I calibrated to using pink noise. If at any point I went past XMax, I would turn my receiver down a little and stat over. To tell when I reached XMax, I used a piece of foam rubber banded onto the end of a piece of bracing which I had left over. (I really should have gotten a picture of the test rig). I then set the end of the foam so that it was 23mm away from the center of the cone. If I saw the cone start to contact the foam, I knew I was exceeding XMax.

After doing this procedure, I ended up with -16db on my receiver being my maximum "safe" volume. Safe in that even if a movie I watch has a 0dbFS spike in the single digit realm, I still won't exceed XMax. Once I build my second sub, I should get an extra 6db. And once I have my iNuke, I should be able to use a highpass so that I can get all the way to 0dbFS on the receiver, and still be safe from 0dbFS spikes in single digit ranges.

One thing I did notice during testing is that at 10Hz, I woudl start to hear some sort of noise before I reached XMax. It wasn't a banging or anytghhing like that. And it didn't seem like it was coming from the port. My assumption at this point is that I am clipping the output of my PAMP1000. I also did notice that once that point was reached, additional volume boost on the amplifier didn't seem to increase excursion, so this further reinforces my theory. Once I get my iNuke, I should be able to confirm or deny this hypothesis.

As far as listening impressions, I listened to the THX Flowers intro, Flight of the Phoenix crash scene, Master and Commander battle, and U571 Depth charges. Even at -16db on the receiver's volume, there was definitely a lot of visceral bass which was not there before. Even with just one sub, and that huge hole in my response centered around 43Hz, my LFEs are definitely a LOT better than they were before with my Klipsch RW12D subwoofer!

Oh, and when I was doing testing at 10Hz, my wife was upstairs. When she came downstairs, she asked if I was doing something with the sub, because she thought we were having an earthquake! WIN!

At this point, I'm not too concerned about the wonky room response. Once I get the 2nd sub in the mix, I'll see how much it helps my room response, and start doing bass treatments and EQ from there. Also, I have one side of the room with a doorway and no door. I plan to buy a door and install it so that the room can be pressurized even better. But even with that door sized opening, I'm getting more than enough room gain to keep me flat to my goal of high single digits!

It is a Listening Position measurement. When I first started, I was doing close mic, and the measurements were showing a lot less output below 20Hz. But at the LP, all is good.

I didn't save any of those, but I can do some tomorrow. I have the front plate on now, though, so the front port and speaker are covered. Should I measure from just outside one of the slots, or should I measure somewhere halfway between the speaker and the port? The cover is easy enough to take off, so I can measure inside the slot also. I'm curious as to whether or not my scheme of putting a block of mineral wool in the center helped or not, but I'm not sure what to look for.

Well, I did a bunch of measurements today. Feel free to geek out, everybody!

10Hz THD measurement @ Listening Position (Rectangular Window)


10Hz THD at max safe output before distortion (Hann Window) With just ONE 18" driver, I'm up to 114.2db at 1.75% THD at the listening position. Once I get my second sub built, I should be up to 120db, and I should get a bit extra once I get my iNuke3000.


More THD measurements



Subwoofer response from 10-1000Hz at three positions. 1m from the slot (pink), right in front of the slot, and handheld ~6 inches from the center of the cone (gold). The purpose of these measurements was to look for resonance issues. Looking at them, I notice that the measurement furthest away seems to have a lot of room influence, as evidenced by the dips at about 24 and 58Hz which aren't in the other measurements. As for the other stuff, other experts here will have to interpret for me. Right at the speaker, which presumably has the least room influence, the plot looks pretty good, with the first major swing at 70-80Hz. It looks to me like a +/-4db swing. While I'm sure a measurement of this beast in the middle of an open field would be very interesting, this sub is NOT going anywhere until I move!


I thought that my Klipsch RF-7 IIs sounded better in "Pure Audio" mode, and now I see why. I'm having some sort of issue integrating the subs into my system. There's a huge response hole from 60 to 90Hz. I did have my front XO set to 40Hz. As you can see from the "Pure Audio" trace, the RF-7s have absolutely no problem playing flat down to 30Hz. But once I switch my receiver to "Stereo", which activates the XO, all sorts of weirdness happens. 60Hz is where my XO is set now as that is the "best" response I could get, but it's still not as good as "Pure Audio", other than the fact I don't have any real output below 30Hz. Ideas, anybody?


If anybody would like to see a specific measurement with a specific chart setup, let me know what you would like me to measure, and how you want me to set the graph up.
Edited by DanLW - 3/17/13 at 9:53pm

I'm just using the Rat Shack digital SPL meter. I also have the cal file loaded into REW, and there is also a soundcard cal. I'm using my laptop's built in sound card.

Hmm, that makes sense. I'm pretty sure the problem existed with my Klipsch sub also, as it is also on the back wall. After I get home from work today I'll have to reverse the polarity and see what happens.

Next time I measure, I'll ditch the cal file and see how it goes. But I'm pretty sure I am getting a pretty decent amount of room gain as I'm only pressurizing a 2000 cubic foot room. I'm currently limited to about 160W from my current amp, so per WinISD, I should be seeing roughly 104db... if I were measuring anecholic. But given my room's dimensions and volume, I should be seeing room gain starting at around 30Hz, and the measurements do reflect that. And if our screening of How to Train Your Dragon last night at -16db is any indication, I am definitely experiencing room gain since even at that relatively low level (I usually watch at -5 to -10db on the receiver) the sofa was shaking pretty good! I can't wait to get my 3000DSP in so that I can put in a highpass filter and turn the volume up a bit!

I'll also try hooking into the line in on my HTPC and see if there is any significant difference in measurements. On my laptop, I do keep the miv volume turned down really low (5 out of 100), and have no boost applied in order to not experience input clipping. But I'll try on my HTPC. Perhaps I'll get even better THD measurements!

Okay, new round of measurements which should be a bit less unbelievable. This time I used the built in sound card on my HTPC, and no cal file for the Rat Shack meter.

LR Front Speakers with receiver set to "Pure Audio"


It was suggested that I try changing the phase of my sub by 180 degrees. These two graphs are with the sub at 0 and 180 degrees. Looks like that wasn't it.


And so then I decided to try running Audyssey. Much better! Red is with DEQ off, green is with DEQ on.


Before and after Audyssey.


Subwoofer only sweep (soundcard output direct to sub amp input) at listening position.


Sweep with the meter held about 6" from the center of the cone.


Pure Audio with laptop (red) vs HTPC sound card line in (green). Red trace is also with Rat Shack meter cal enabled.


Subwoofer laptop vs HTPC sound card line in. The laptop trace (the higher one) was also done with the Rat Shack meter cal enabled.


As far as the laptop vs built in sound card, the biggest difference was probably the use vs non-use of a cal file for my SPL meter. Also, the meter was not in the *exact* same location for both measurements. There is at least a few inches, but less than a foot difference in mic placement.

So looks like I am actually perhaps 15db lower than where I thought I was. Makes me want to bet the second sub built all the more!

Also, I should be getting my iNuke3000DSP in sometime this week. Once that gets in, I'll be doing another round of excursion tests, applying a high pass, and will do more sweeps.