Sealed test enclosure

**Ricci** · **10-08-08, 09:48 PM**

I have built a 10 cu ft sealed enclosure with a removable or changeable baffle, so I can easily remove extremely heavy drivers, or switch driver sizes. This will allow me to more accurately measure a drivers TSP's with the WFT3, to see the effects of stuffing, different volumes by adding material and to have a control enclosure for testing drivers. It's nothing pretty, or deadly accurate as it's not something that will be used permanently. Built it with a jigsaw, sabresaw and a power drill in 2 afternoons. Dimensions are 24.5" x 24.5" x 36.75". The driver mounts in with 12 1/4" x 3.5" bolts into a 2x2 with t-nuts and is sealed with weather stripping.

**thxgoon** · **10-08-08, 09:59 PM**

Neat! I'm interested in hearing the results of your testing.

**Ricci** · **10-08-08, 10:27 PM**

So far I had a small air leak in the front baffle and some more air leaks from the driver mounting holes, but other than that things work ok. I've taken the baffle off probably 5 or 6 times so far with no problems.

This is an impedence measurement of a XXX driver in the box with no stuffing. 24.9hz resonance peak.

This is the same test with stuffing added (11 Walmart pillows). 23.5hz peak.

This is a close mic fr also with no stuffing and no crossover. Notice the nastiness at 170hz. This is a box resonance apparently. The level seems high but these measurements are with the mic 3" from the cone.

And the same sweep after the pillow stuffing has been added...Notice the 170hz nasty's have been diminished greatly

.

Here are the before and after overlaid on each other. Orange is with stuffing.

**avaserfi** · **10-08-08, 10:31 PM**

Considering the box dimensions it seems as if the peak at 170Hz is related to the 1/4 wavelength internal reflection. It seems as if your introduction of the pillows was sufficient to stop this reflection. This is one of the many reasons I recommend using a high grade acoustically absorbent (OC705 or 8lb mineral wool) material in subwoofers and loudspeakers. Materials such as polyfill or eggcrate foam are not sufficient to properly absorb waves of this length.

Do note this isn't typically an issue with smaller subwoofers, but with units such as LLTs it easily becomes an issue due to their massive size.

**Ricci** · **10-08-08, 10:52 PM**

Yep. That's what I thought it was too. I'm kind of surprised that the pillows helped out with it so much.

**avaserfi** · **10-08-08, 11:57 PM**

Quote:

Ricci wrote:

Yep. That's what I thought it was too. I'm kind of surprised that the pillows helped out with it so much.

This is likely due to the large amount of pillows you added. A more efficient material would require less to have the same results.

Also, note that in the higher octaves, based on your cabinet design it is likely that there are audible cabinet panel resonances which do not actually show up on such measurements. The only solution to these resonances would be decoupling of the driver or a more efficient/dense bracing matrix. Of course, this is not an issue with most subwoofers, but is one with midbass modules or loudspeakers.

**Ricci** · **10-09-08, 01:09 AM**

Actually the resonances aren't too bad. There are a few that showed up when I did a loud full range sweep, but they were out of the SW range that this will be operated in. The box could use more bracing up front but I needed to keep it open for possibilities down the road. This is quick, dirty and rough around the edges. For sw testing below 200hz.

How would you decouple a big powerful driver?

**Mike P.** · **10-09-08, 07:57 AM**

The 1/4 wave length of 117 hz is 29 inches so is that the internal length of the cabinet?

**avaserfi** · **10-09-08, 11:15 AM**

Quote:

Ricci wrote:

Actually the resonances aren't too bad. There are a few that showed up when I did a loud full range sweep, but they were out of the SW range that this will be operated in. The box could use more bracing up front but I needed to keep it open for possibilities down the road. This is quick, dirty and rough around the edges. For sw testing below 200hz.

Actually, cabinet resonances will not show up in a typical measurement. These other resonances you saw are likely either other internal reflections or cone breakup. Large resonant peaks will normally show up as blips in the impedance plot (as the peak at 170Hz did), but other audible resonances can still be present and not show up on the plot.

There are two ways to measure cabinet resonance the first is using an accelerometer. The other, far more accurate method, is to take multiple far field impulse response measurements at varying angles in an anechoic chamber then average the results and waterfall.

Quote:

How would you decouple a big powerful driver?

Such a large driver isn't really reasonable to decouple. The method B&W uses would work though:
The driver is installed from the rear of the front baffle with a ring of low Fs foam attached to it. A piece of threaded rod is connected to the rear of the driver and inserted into a hole in the rear baffle where foam is placed. This tension holds the driver up while ensuring it is decoupled.

This would probably result in about a 6dB drop in cabinet resonances alone (not sufficient to remove audibility fully). Now, if the intended passband it just up to 80-100Hz there is no need as typical subwoofer bracing techniques are sufficient to attenuate resonance this low. Cabinet resonance is a more an issue from 150Hz-1kHz with most damping technics.

Quote:

Mike P. wrote:

The 1/4 wave length of 117 hz is 29 inches so is that the internal length of the cabinet?

The resonance is at about 170Hz 1/4 wavelength is about 20 inches

.

-Andrew

**Ricci** · **10-09-08, 02:43 PM**

Internal front to back measurement is 34.5". I think that the 170hz blip ( it was actually in between 165 and 170hz) was most likely due to a side to side resonance as it is 23.25" x 23.25" square and this is where the least amount of bracing is up front.

I didn't mean that I saw resonances like the one at 170 on the full range sweep graph. (Yes that is not a pretty graph either! I didn't expect it to be though and I just had to know.) I heard them or felt them with my hand resting lightly on the top during a loud full range sweep. My best guess is that there is another around 340hz and seemed like there was another one higher up. These are way out of the range of this driver though, which could never be confused with a woofer or MBM. It'll be crossed at 80hz or lower and anything past 150hz should be well rolled off.

I was interested what you would say about de-coupling, because I just don't see any way to do it effectively with a driver capable of exerting force like this one. I was all ears though if you did have a solid idea

.

**collo** · **10-09-08, 05:06 PM**

Nice work Ricci!

Your peak may be a front-to-back-wall half wave resonance...

You should also have a driver-to-rear wall quarter wave resonance at around 99hz. I see you have a slight ripple there, but it's not really stong enough to be conclusive.

If you expand your sweep frequency a tad, you may also see a side-to-side resonance and a top-to-bottom resonance of around 294hz.

I'm in the process of building a tubular sub with the driver occupying one end completely, and the other end being a moveable wall with the intent of demonstrating the existence of a quarter wave driver-to-rear-wall resonance.

The discussion is at this thread at diyAudio

**Ricci** · **10-10-08, 12:59 PM**

Collo,
That sounds like a great project. I just downloaded Boxnotes. I've seen a lot of people talk about it but for some reason I never checked it out. Looks cool especially the resonance estimations. I think the 99hz resonance may be greatly reduced in level because the driver magnet is 10.5" diameter, rounded like a bell and the frame lines up with the center hole in the bracing. There's not much of a straight line from front to back wall except for the small area of the secondary brace windows. I did take some full range measures but I'd have to use a different driver to really tell anything. This one gets rough above 300hz and really bad above 400hz. This is unsurprising given the intended use of it. It'd be hard to tell what's going on, but I didn't see anything that jumped out at me around 300hz.

**collo** · **10-11-08, 07:58 PM**

Thanks for those comments on boxnotes. As it turns out, they might be premature as the experiments linked to show that the driver-to-wall resonances are half-wave, not quarter-wave as I had always assumed. This doubles their frequency.

Boxnotes and sonosub are in the process of being fixed, meanwhile your screen would now look like...

I feel this is just the first of the apologies....

**avaserfi** · **10-11-08, 08:33 PM**

Quote:

collo wrote:

Thanks for those comments on boxnotes. As it turns out, they might be premature as the experiments linked to show that the driver-to-wall resonances are half-wave, not quarter-wave as I had always assumed. This doubles their frequency.

Collo. This is physics of pipes, a quarter-wave driver to wall interaction will create a null (destructive) in the response and a half-wave interaction will create a peak (constructive). Both will effect frequency response if not properly attenuated their effects will just be different. Clearly, something is effecting the test enclosure at 170Hz, this is a quarter wave cancellation. The reason for the following peak is likely due to the more complex interactions between wavelengths.

It seems your experiment didn't show the quarter wave null because there was no reflective surface to interact with the quarter-wave. This reflective surface would be present in a sonotube or typical box build.

-Andrew

**Ricci** · **10-12-08, 12:29 PM**

You may be right that the 170hz deal is more than just one simple resonance and is possibly the complex interaction of a couple different things. I'm just happy that it is out of the normal passband for a SW and that the stuffing helped tame it some.

Collo,
No need to apologize. It's a free program after all.

**collo** · **10-12-08, 06:06 PM**

Quote:

avaserfi wrote:

It seems your experiment didn't show the quarter wave null because there was no reflective surface to interact with the quarter-wave.

I had postulated that the driver, being in motion, was a region of high velocity and low pressure (because the air is free to move). This would be equivalent to the open end of a pipe, leading to a sustained quarter-wave resonance between the driver and its opposite wall. No additional interractions would be required to support this resonance.

The SPL meter was mounted through the wall opposite the driver. This location is a pressure node for both half-wave and quarter-wave resonances ie not a null point, and so would show either resonance if present.

Only a half-wave resonance was detected. This says that the driver is acting as a pressure node rather than a velocity node, as others have suggested.

I think now, that the only quarter-wave resonance present in a sub, is the pipe-mode resonance in a port.
(edit: Pipe mode resonance in a port is also half-wave)
All the driver and wall resonances appear to be half wave.

As for the 170hz dip / peak, it does indeed look like the result of some interractions. The size of the brace would provide reflections that would mix with any standard box resonances. Its good to see that the damping got rid of the problem.

**avaserfi** · **10-12-08, 06:26 PM**

So, was there was a wall opposite the driver through which the SPL meter was mounted? You seem to be implying this. Am I understanding properly? From the picture it seemed as if the SPL meter was free mounted with no wall.

It if there is, then it would seem there is no quarter wave resonance, oddly. If not, there seems to be a lack of allowance for reflective interactions to be accounted for.

As far as the 170Hz dip followed by a peak in this example, I believe my explanation was misinterpreted. The initial dip is directly related to the quarter-wave reflective cancellation within the cabinet. The following peak is related to a complex summation of varying interactions within the cabinet.

-Andrew

**collo** · **10-12-08, 06:50 PM**

Ah, that clears up the confusion.

Here's a picture of the SPL meter / movable wall...

The driver occupies the entire end wall....

For those who didn't follow the link, here's the results we're talking about...

No evidence at all of quarter-wave resonances, but conclusive demonstration of half-waves.

**avaserfi** · **10-12-08, 08:46 PM**

Is there any possible way you could repeat the experiment and measure the rear wall, as you did, as well as driver response? I believe the experiment you did measured the half-wave resonance (perhaps similar to port resonance), but due to microphone placement did not manage to capture the quarter-wave reflective cancellation.

I fail to see how there is no quarter-wave cancellation since it is a result of the source emitting a wave that hits a surface at its quarter point, 90 degrees, and then is reflected back to the source at another 90 degrees creating a cancellation.

I am not saying this quarter-wave anomaly is a resonance, just something that should be noted. After all, this cancellation is clear in Ricci's test box as well as other measurements I have seen similar to his.

**collo** · **10-13-08, 12:43 AM**

A quarter-wave dip, if it's there, would be measured back at the driver, however the energy reflecting off the far wall should still register as a slight rise in SPL for that frequency. This is not seen in the results.

I'm going to do another set of measurements to look at how the intensity of these peaks is reduced outside the box. This involves taking a measurement as before (just at one spacing, probably 500mm), and then removing the meter without moving the wall. A second sweep will then be done and the SPL measured outside the box at 1m from the driver.

This will show just how much importance should be attributed to these resonances. As an adjunct, any quarter wave dip should be visible in the second measurement. How does that sound..?

**avaserfi** · **10-13-08, 01:21 AM**

Quote:

collo wrote:

A quarter-wave dip, if it's there, would be measured back at the driver, however the energy reflecting off the far wall should still register as a slight rise in SPL for that frequency. This is not seen in the results.

Understood, I just want to cover all the bases. There are a variety of factors that would prohibit this from registering.

Quote:

I'm going to do another set of measurements to look at how the intensity of these peaks is reduced outside the box. This involves taking a measurement as before (just at one spacing, probably 500mm), and then removing the meter without moving the wall. A second sweep will then be done and the SPL measured outside the box at 1m from the driver.

When you do this could you also meausre 1" from the driver? This would satisfy the conditions I previously outlined.

Quote:

This will show just how much importance should be attributed to these resonances. As an adjunct, any quarter wave dip should be visible in the second measurement. How does that sound..?

The audibility of these peaks and dips will be related to Q, source material and frequency (there has been large amounts of research on the audibility of such issues). In the end it doesn't really matter since simple steps can be taken to fully remove these anomalies via proper damping

. Then again, my goal is always perfection...

**collo** · **10-13-08, 03:29 AM**

1 inch it is!

**Ricci** · **10-13-08, 11:13 AM**

Great tests Collo. I look forward to your further results.

Andrew,
I saw your anouncement and checked out the site. It looks good and I'll be waiting to see what you have in store in the coming months. Once you get things rolling I could possibly help provide some drivers for testing.

**avaserfi** · **10-13-08, 11:21 AM**

Quote:

Ricci wrote:

Andrew,
I saw your anouncement and checked out the site. It looks good and I'll be waiting to see what you have in store in the coming months. Once you get things rolling I could possibly help provide some drivers for testing.

Thanks! It has been a lot of work!

PM sent.

**Ricci** · **10-13-08, 11:29 AM**

Back to the sealed test box and the graphs I posted...The focus here so far has been about resonances and their effects, which is great. I also have a post in my big thread at AVS and the discussion there has been about the FR itself. The main one being that the roll off shown is greater than the expected 12db octave that a sealed box would normally show and is closer to 16db octave. Now this box & driver combo (XXX18 10ft) has a Q of about .875 or thereabouts because these model like a overbuilt IB driver and you can forget about getting a normal .7q. This can contribute to a bit faster roll-off and so could a bit of inductance, but not to this extent I would think. The other effects in play are that the CE4000 amp has a 3rd order butterworth HPF at 8hz and the Rane PE-17 has a 12db octave at 10hz. Together these affect everything from about 12-13hz down.

Now I also have a pair of SDX15's in 4 cu ft each, so I took a close mic FR of each of them and they matched each other very well. They exhibited a faster than 12db roll off too. Again it's more like 16db octave. These are in a different room, with a different system and I'm using a different CE4000 with them. Here is that graph. It does have a 24db octave filter at 200hz.

At this point I'm like

at the steep roll off which has been exhibited by everything so far, so I switched out the CE4000 for an AETechron 7560 which is a dc coupled amplifier that has been tested by Chuck over at AVS in the measuring amplifiers thread as being -0.2db at 10hz and took another measurement.

As you can see this did flatten things out quite a bit. Here are the 2 graphs together. Blue is the 7560 trace. Notice the 30hz area

. I have no clue what's up there. Better control of back EMF?

So at this point Illka chimed in with his measurements of the SDX which are much flatter and extended in the HF's and in a similar sized (smaller) enclosure, his show what appear to be a 40hz peak while mine are at 30hz. Also he has 2 CE4000 also and both of his measure nearly dead flat to 10hz. I find it hard to believe that both of mine are exhibiting the same weird response irregularity.

Here is Illka's 2m groundplane outdoor FR of an SDX15 in 100L.

At this point I'm trying to track down what if anything is skewing my measurements. Maybe my ECM8000 does not correlate well with the mic cal file? My testing is somehow flawed? Don't know. Next stop is FR testing all of my amplifiers, outdoor measurements and possibly getting a 2nd ECM and having one professionally calibrated.

Advertisement