SME

I was not aware of electric absorption effects, but wouldn't a bleed resistor discharge that lingering voltage as well?

I agree with @peniku8. The high Z mode is probably for commercial/retail applications where several in-ceiling speakers may be strung together in series for one amp channel. That makes sense given the 70 Hz high pass filter, which keeps heavy bass from overloading little speakers. As I noted, you need a HPF down at like 16-20 Hz. Your sub will probably do just fine without it for most content, but a few movies have strong content at 15 Hz and below which could overwork the sub even at fairly moderate volume. I suggest being mindful of the situation and if you hear obvious distortion with bass, turn it down. When you get around to doing measurements, one of the first things you'll want to check is where the tuning frequency is. One way to do this is to use a tone generator (like the REW software has), starting at like 25 Hz. Set the volume so you can see the cone moving, but only just. You don't want to burn out the coil by running the tones too hot. Then gradually decrease the frequency and note the excursion. It might increase a bit, but as you drop the frequency, it should start to decrease until the driver is almost stationary. Below there, excursion will increase again. Where the driver moves the least is where the tuning frequency is. With your design, I expect it to come in roughly in the 18-22 Hz range. Have you had more time for listening? Feel free to share your impressions.

Congrats on finishing the build! It looks real nice under the TV. I apologize. I hadn't noticed that the Crown XLS DSP is limits the high pass filter setting to 30 Hz. Without a high-pass filter (HPF), preferably set a lot lower than 30 Hz (like at 16-20 Hz) heavy bass below the sub's lower limit can cause heavy distortion and damage to the sub over time. A high-pass filter is generally recommended for vented subs; although yours is tuned low enough that you probably won't encounter issues with most content and don't turn the volume up too much. One option to allow setting a lower HPF is an external unit from MiniDSP. I haven't looked at their stuff for a while, but you should be able to get something very capable for ~$100. In addition to allowing setting of a high pass filter, you can set other EQ which can help you tune the sub for the room. Either way, I'm glad you are enjoying the new sound. I'm pretty sure your new sub plays a lot lower and louder than the old one. +

That's a real bummer. So if the voltage decayed exponentially (i.e. as a series RC circuit), then the rail voltage after 2 minutes powered off was probably like 45 V. Of course the amount of energy involved in the discharge depends on how much capacitance there is, and I imagine 45 V with a lot of capacitance can do quite a bit of damage if it's released all at once.

Interesting. I wonder how fine-grained the IPAL system is here? Does it adjust EQ in real-time or just gain? If it adjusts EQ, how tight is it? Do the sound characteristics noticeably wander through a live performance? It's not too hard to monitor DC resistance and adjust gain in real-time, but that may not be good enough. I expect that thermal changes manifest over seconds to minutes or hours. If we assume all of the thermal effect is to raise the coil DC resistance, then the nature of the non-linearity is such that we can understand things in terms of linear response that changes with time. This is not possible for inductance because inductance can fluctuate much more rapidly, with each stroke of the driver. Using a linear analysis, we consider a sine wave at one single frequency at a time. The impedance (Z) describes the relationship between voltage (V) and current (I), and we can describe the behavior completely if we let V, I, and Z be complex quantities with *real* and *imaginary* parts. Each quantity can alternatively be described as having a *magnitude* and *phase angle* part. Both descriptions are useful depending on the circumstances, and one can convert between them using arithmetic formulas or geometry. A complex quantity can be represented as a point on a 2D X/Y with x = *real* and y = *imaginary* parts. Draw a line between this point and (X=0,Y=0). The magnitude is the length of this line and the phase angle is the counter-clockwise rotation from the positive part of X-axis to the line. (This article on the Complex Plane might be helpful.) For V and I, the magnitude is the absolute value of the peak amplitudes of their oscillations. The phase angle describes the *phase shift* which essentially describes the temporal shift while recognizing that a continuous sine wave is *periodic*. Periodic means it precisely repeats at the same interval. If you shift a sine wave by exactly 1, 2, 3, etc. periods, the result is exactly the same. So, it makes sense to represent the amount of time shift as a *phase rotation* on a circle. Analogously to Ohm's law, the relationship at a single frequency is: V = I * Z or (rearranged) I = V / Z or (...) Z = V / I. The rules for multiplication/division of two complex quantities are as follows: (1) multiply/divide the magnitudes to get the new magnitude. (2) add/subtract the phase angles to get the new phase. Therefore the meaning the impedance phase angle is the phase difference or change between V and I. Note that this math is generally useful for analyzing oscillating signals including audio acoustic transfer functions (i.e. frequency response magnitude/phase). So if we can easily multiply and divide complex quantities using the magnitude/phase description, what is the purpose of the real/imaginary description? The latter is for adding subtracting such as when analyzing a series electrical circuit or acoustic interference effects. The rule for adding and subtracting complex quantities is to add and subtract the real and imaginary parts independently. For impedance, the *real* part is called the *resistive* part, and the *imaginary* part is called the *reactive* part. As expected, a pure resistance (i.e. straight a wire) contributes only to the real/resistive part of impedance, and the tendency for oscillating energy storage/release manifests only in the reactive part. Ideal (as in zero resistance) capacitors and inductors only contribute to the reactive part. Do note that the resistive part of the impedance is not always equal to the DC coil resistance. The acoustic properties of the subwoofer contribute to impedance (both resistive and relative) as well. As such, we can conclude that increasing the DC resistance by heating the coil will alter the frequency response by different amounts depending on both the magnitude and phase of Z. The increase in Re adds directly to the resistive part of Z. The peak current is reduced, but the amount of reduction depends on how much the *magnitude* of Z changes. So you can plot the initial value of Z on the Complex Plane and then plot the *new* Z shifted to the right by the increase in Re. Then, the current will decrease in inverse proportion to the change in distance between each point and (x=0,y=0).

Yes. Their solution is very interesting, but it's not implemented on many drivers. I wonder why not? Acoustic Elegance designs their woofers with a long copper sleeve in very close proximity to the voice coil which dramatically cuts inductance in their drivers. AE also claims that their sleeves improve thermal performance, but I have not seen independent verification of this claim. Some other sub manufacturers (Funk audio, maybe others?) also use full sleeves, albeit not installed as close to the voice coil as in AE's designs but close enough to yield impressive performance and linearity.

I'm not 100% sure on whether or not the energy that would have been stored in the field is dissipated, but I believe it is. I'm guessing it's typically a small amount, which may reduce the driver power efficiency slightly but I don't know. No. Shorting rings actually *increase* voltage sensitivity by allowing higher current, specifically for high frequencies. Higher coil inductance increases the lag time for current to respond to changes in voltage. The higher the frequency, the faster voltage is oscillating and the less time there is for the current to get up to speed before the voltage reverses again. I'm not sure which aspect of the IPAL system you are referring to here. For amps that can maintain consistent power output into very low driver impedances, voltage sensitivity (such as inductance causes for high frequencies) is less likely to be a limitation, but that doesn't correct all the problems that inductance causes. EQ / signal shaping can be used to compensate for high sensitivity frequency droop, as is often done in passive crossovers, but the correction is likely imperfect and may allow some "sound signature" to persist. Furthermore, inductance does not behave especially linearly with medium-to-high signal levels, and this can cause significant compression and distortion effects. Well-designed shorting rings directly reduce and linearize inductance and so help with all of these problems.

I will try to explain as best I can here. Shorting rings don't act mechanically, at least not directly. They magnetically interact with the voice coil to reduce and/or linearize inductance through the driver's stroke. The voice coil is a natural inductor. Inductors store and release energy via the magnetic field in their proximity. For a straight wire with pure resistance, the current responds in perfect lock-step to changes in voltage. If an inductor is subject to a sudden increase in voltage, however, some time and energy are required for the current to "spin up" the magnetic field. At that point, if the voltage is suddenly cut to zero then the current flow continues for some time while the magnetic field "spins down" and releases stored energy. Hence, rapid fluctuations in voltage tend to be smoothed out in current, e.g. high frequencies are reduced. This is a major cause of high frequency loss and sometimes "humping" in a speaker driver's response. At the same time, the inductance itself is likely to vary, not just with frequency but with changes in instantaneous current and/or driver stroke. This is *non-linear* behavior, which causes distortion, including inter-modulation distortion, which may be particularly undesirable. Though the linear aspects can also be degrading if not precisely corrected with EQ. The response "humping" alters the spectral balance and likely imparts a non-neutral characteristic sound. A shorting ring is made from a material that is both magnetically and electrically active. The magnetic field generated by the current induces current flow within the ring. The ring "shorts" this current to the rest of the driver assembly, allowing some of the energy in the magnetic field to be transferred to the shorting ring and dissipated as heat instead of stored. If designed correctly, this effectively reduces the inductance of the coil), and depending on the position of the rings vs. the voice coil, may also keep inductance from fluctuating as much throughout the stroke. The relationship between inductance and damping is via the electrical impedance. Impedance is essentially a 2-dimensional quantity which can be described in terms of a pair of parameters: either *magnitude* and *phase angle* or *real* (resistive) and *imaginary* (reactive). Inter-conversion is possible via basic trigonometry; see the "Complex plane". Damping is a property of the resistive / dissipative (non-energy storing) component of impedance. Pure inductance and capacitance both contribute only to the reactive (energy storing) component of impedance. Speakers using a composite electrical circuit that has effective resistances, inductances, and capacitance contributed by several different factors including the mechanical and acoustic properties of the system. So needless to say, inductance and "damping factor" both contribute to the system behavior in a way that's not simple to describe. To answer your last question: No. I mean, if the resistance of your speaker wire is high enough be a problem in the absence of shorting rings, then shorting rings probably won't fix that problem.

Servo technology reduces subwoofer non-linear distortion substantially when the sub is being pushed to its limits. Below that point, the difference is not as likely to be appreciated, and above that point, the sub runs out of travel completely just like any other sub does. If you want lower distortion, it's usually better to have enough displacement headroom so that you aren't driving the subs near their limits much if at all. Another thing is that most servo solutions probably control lower harmonics better than higher harmonics, but the higher harmonics are probably much more important, perceptually speaking. Of course sometimes you have to compromise, and servo technology has its niches. Just don't expect any major SQ improvement from it. The whole sub system design, especially including SPL capability is important. I would say there is a primary three-way compromise between *cheap*, *compact* and *low-and-loud*; pick any two. Low-and-loud are one in the same because it's pretty easy to make something tiny play loud if you don't care how low it goes, and you can technically EQ any system to be "flat" as long as you never turn up the volume high enough to hear anything.

"The maximum output in dB is not very important. " But it *is very important*. It's always very important because it has by far the greatest impact on other trade-offs like size and cost. With that said, there are a variety of DIY and commercial options that can get you started, and you shouldn't have much trouble getting decent output for moderate music listening for the range you indicated in a fairly small room like yours. I'm reluctant to name brands because I'll probably leave out someone relevant, but off the top of my head, I can think of Power Sound Audio, Hsu Research, Rhythmic, and Monoprice as being potential sources of high-value/budget consumer offerings, similar to SVS. I don't know what kind of DSP capability any of them offer these days though. Depending on how you value your time and labor, a DIY sub can get you a lot more for your money than anything you can buy. It also gives you the opportunity to focus the design better for your needs. For example, a larger sub may be more feasible for your room if you can design it to fit the footprint you have available. As far as sub type, vented typically offer more output for the money than sealed but at the cost of size. Vented subs also roll-off much more rapidly below the bottom end of the bandwidth, and typically require a high-pass filter to protect them from excessive content below their capability. Subs from the aforementioned vendors provide DSP for this, even if they don't provide user accessible controls. My last piece of advice is to give serious consideration to two subs instead of one. When placed in different parts of the room, two subs typically offer much better sound quality than just one, for a few reasons having to do with room interactions.

Oh, I forgot about the VBSS, and when I was looking at vented PA460s, I was looking at the outcome of using 2 x PA460s in a single small vented box. No wonder it didn't work. If you don't mind the size and footprint, VBSS may be a great way to go. Four of them at the 20 Hz tune will probably handle just about anything you throw at them, and they don't need much power at all.

The sealed wedge cabinet I suggested could be built using 2 x PA-460s each with the drivers in a diagonal orientation. The PA-460 doesn't appear to do vented well unless it's tuned higher than you want. Any vented cabinet will tend to take a lot of space vs. displacement of the drivers installed, especially if you tune lower. The MBM options from DIYSG are definitely worth a look, and that MBM-15XL driver looks real nice and provides a little more displacement than 2 x PA-460s. I'd probably opt to build bigger cabinets and/or make the ports longer in order to tune though lower. I believe the 30 Hz bottom end spec assumes a lot of roll-off by that point. IMO, you should tune lower so that you don't roll-off at all until you're under 30 Hz. Lots of psy/ambient/bass music I have goes to 30 Hz if not lower, and I notice extension to 30 Hz and below is becoming a lot more common with popular music as well.

When you find an inexpensive high displacement strong motor driver, let us know. Also if let us know if you find any unicorns. ;) I think the 21DS115 is one of the better choices if you're looking for "affordable", compact sealed ULF and don't want to compromise mid-bass or use separate sets of subs. Is its distortion below 15 Hz uniquely bad? Almost all sealed subs have bad distortion under 20 Hz, and IMO the Data-Bass measurements look pretty good in the ULF compared to alternatives. I think the 21DS115 and the LaVoce knock-off actually hit a kind of sweet spot. Multiple sealed cabinets are most useful if you intend to spread them out in the room. If you only have two locations for subs, then building only two cabinets will probably save on materials, labor, and weight and will also make slightly more efficient use of the available space. If you go with the wedge cabinet design I suggested, you need something deep enough for the SM60fs to fit. They are 16" deep, but still have some width at their back-ends. I'm not sure how much. I'll take a wild guess that they need 18" depth. So build 18" deep corner cabinets which are 36" wide and 36" tall. External volume: 6.75 cuft. I'll make a rough guess that you end up at 4.75 cuft internal volume or 135 liters. This is what that looks like in Hornresp with a SAN214.50 with 77.5 V (750W @ 8 ohm / 1500W @ 4 ohm) applied and simulation with complex inductance: One SAN214.50 21" isn't quite as good as 4xPA460s in the deep bass. For mid-bass, the 460s are a lot better. However, I doubt you'll actually use that extra mid-bass output. Also realize that the 4xPA460s were simulated in ~8.5cuft internal volume vs. 4.75cuft for the SAN214.50. That's a much bigger cabinet. If you want the same deep bass output from the 460s in the smaller cabinet size, you'll need a lot more power---probably more like 2-2.5kW @ 8 ohm for each set of 4. The SAN214.50 makes better use of less cabinet space. Lastly the SAN214.50 has better build quality with a cast frame vs. stamped frames on the PA460s which may have practical performance consequences as their limits are pushed, but I don't know. The price on the 460s is hard to beat, especially if you don't pay any additional shipping charges per driver. If you have a bigger amp and/or don't mind making bigger cabinets, they look like a good choice. Otherwise, I think the SAN214.50 looks like a good choice.

Is $1k the budget for just drivers or the whole subs? Amps are expensive, so hopefully you have that covered. If you already have the amps, what are their power specs? Here's a possibility: Build a pair of wedge shaped sealed cabinets, i.e. with a 45-45-90 triangle footprint. Make them about 36" tall and just wide enough to fit a LaVoce SAN214.50 (21") on the front. Install one in each corner and stack the SM60fs on top. Maybe add some custom wood trim pieces to go between each front edge of the SM60f and each wall. Use EQ to tweak for the room and to shelve deep bass up to taste. Recommended amp power: 1500W @ 4 ohm each or 3000W @ 2 ohm with both in parallel. You should be able to get bass that goes as low as you want for music at moderate levels, especially if the room is closed off. Edit: So I guess the SM60f too deep for the above design, so just make the cabinet wider and deeper until everything fits into the corner with the horn front flush with the sub cabinet. I think this design is hard to beat because both the speaker and sub will probably do well corner loaded like that, and you'll get an attractive clean / space-saving look.

I think you'll want extension to at least 30 Hz if not lower. Either vented or sealed probably makes sense for your setup rather than horns. What placement locations are available in the room and how much space is available in each? Will you have any kind of DSP available for EQ?

Good points about live kick drums used in rock and metal. When you talk about the 100-1000 Hz decade being attenuated are you talking about live music shows too? Or just content mixed for release? FWIW, cinema content tends to be quite a bit hotter below 250-500 Hz. I'm curious in general as to how much 100-500 Hz "mud" is a problem in a large room live setting. My experience is that the fundamental cause of "mud" is resonances in that range, and that the effect of broad boosts and cuts over that range is primarily to accentuate or suppress the effect of those resonances. These resonances can be caused by the speaker, by its interactions with its surroundings, or by EQ problems. OTOH, if response in that region is clean, there is no "mud" even without applying broad cuts. FWIW, my home system crosses around 100-150 Hz (it's a custom "sloppy" XO) as that seemed to work best with my setup. I also like subs that can play higher. That PDF is where the joke came from? The recommendations in there are curious. Why reduce upper-midrange with a Q=4 filter? That's really narrow isn't it? I don't put much stock in generic EQ recommendations. Every situation is different, and it's good to be able to identify and fix the major problems at whatever frequency region they occur in. That takes a lot of skill and experience though. I don't think you need 120 dB for TR at 500 Hz, which is great because 500 Hz @ 120 dB is bad for the ears. IMO the best TR is *balanced* TR (or balanced FR, if you don't have a dedicated TR device), and I find balance is even more important for TR than for sound. When bass is extremely accurate, it can actually be surprisingly *intense* at relatively modest levels. This is very good for reducing SPL requirements, reducing equipment requirements and protecting health.

Every kick drum sound is a bit different. I note plenty in the 50s and 60 but also some that go down to 30 or 40 or up to 80 --- or 100. Furthermore, depending on genre or mix/mastering choices, kick drums may also be "doubled" creating a subharmonic an octave below the fundamental. This is quite common with music in movie tracks and is becoming more common with music masters too. Kick fundamental is also almost always accompanied by a strong 2nd order harmonic an octave higher. Just 93 Hz? Let me point out two things. First, different people come with chest and vocal physiology of different sizes which likely have different resonant frequencies. Second, the body exhibits tactile responses for a very wide range of bass frequencies, roughly anything below 500 Hz in my experience. However the subjective quality of the tactile bass response is extremely sensitive to linear frequency response, and some otherwise minor flaws can mask away most of the tactile sensations. Otherwise though, I agree that crossing up at 120 Hz is often beneficial. Music often has a lot of energy by that point.

This is a complicated question, and another variable concerns the nature of the load. A highly reactive load may behave very different on an amp than a purely resistive one. Most amps provide substantially less power output below a certain frequency. Most should lose little power at 40 Hz, but I think there are some that are already down below like 100 Hz. ("Burst" power for "very short" bursts, heheh.) Max power output is a lot more likely to drop off below 20 Hz. I think beyond 10s, you're just testing the protection circuits, which is nice info to have but not worth the cost of amps and the fire and health risks. What might be more interesting are some "real-world" tests using some of that EDM. Like, can you actually do -6 dB average with full-scale kicks for a significant duration? Maybe that still risks blowing any amp, but it's more likely to be encountered in real world use. I think the purpose of the CEA is to keep the test narrowly focused on the frequency region of interest while keeping it transient. Without a smooth on/off ramp applied to the sine wave, the test signal takes on a lot more high frequency content at the start/stop points. Almost all real world signals that "hit the limit" are going to decay more like the CEA sample. Think a kick drum tone that is a decaying oscillation. The CEA isn't so realistic in that it has a slow "on" rather than a hard transient like you'd get from a drum. If you don't like CEA, one option would be a signal that is 0 for t < 0 and a cosine wave for t > 0, multiplied by an exponential envelope (to simulate smooth decay). A cosine wave is a sine wave shifted so the peak is at t=0, so this will be like a hard kick drum transient. If the amp has a LPF, you might just turn that on before feeding this signal, or you can apply your own LPF to it. It's not completely realistic because real-world transients have a lot less HF energy, but it should be a lot closer to a real drum hit. For a sub-only amp, 16kHz seems pointless. For a general purpose amp, I think it would be more helpful to know how much if any response roll-off there is in the top. THD+N type stuff may also be helpful. You're right that no one is playing 16 kHz+ at 10 kW or even 200W. My home speakers have ridiculous boosts (20 dB+) above 17 kHz, and it's never a problem. As far as which bass frequencies to test, octaves seem to be a reasonable compromise. I don't think 15 Hz is representative of single digits though. A lot of amps will do 15 Hz just fine but seriously struggle with 5 Hz. Even the SpeakerPower amps have quite a bit less to give down there. Of course performance at 5 Hz is only of interest to a small number of people. Maybe do 10, 20, 40, and 80 Hz?

My apologies, I was rushed when I posted my last reply and did not read your detailed post. Yeah, the numbers you posted look a lot better. I don't really like the blowing up part though. Good luck with that!

Lots of people are using that for home theater subs that are meant to go to 10 Hz or below, so I think it's an important detail that you don't get peak power long enough to complete a full cycle at 10 Hz. I know this kind of short duration peak power output is typical for amps which are designed to give their best for kick drum peaks and not a moment longer. In a HT setting though, this amps doesn't look like it's going to hold up to one that can sustain the output for several hundred milliseconds at least. If we're talking about highly dynamic medium duration "bursts" (e.g. on the order of 500 ms long), the "AC line" should handle it just fine. I mean I'm sure there are exceptions. The length of the run is important, and especially how the amp handles voltage droop. The better amps don't care much about voltage droop and will suck every extra bit of current they need to keep the output going full-tilt for a medium to long "burst" duration (i.e. up to a few seconds with the SP amps). I find it kind of lame that so many amps on the market now don't have decent protection circuits.

Only 80 ms? That's not long at all considering that people are using those Sanways for ULF HTs.

Very interesting. How well do you think it will work to limit output of high-harmonic distortion and noise?

Two notes: First, I have found sine waves to be a terrible test signal to assess the bandwidth limits of human hearing. I don't hear sine waves too well above 16 kHz, but with real world content, I hear contributions from frequencies to 20 kHz and beyond even at fairly quiet volumes. My wife once complained about an annoying treble harmonic in a music passage, which I identified to be at 20.8 kHz, and she confirmed that the annoyance was gone after I filtered it out. Yet in sine wave tests during the same session her ears didn't respond above 10.5 kHz. (!) Second, frequency response shape in the upper treble is tricky to interpret. In-air absorption is quite profound at the upper reaches. For example, at typical room temperature (T) and relative humidity (RH), response at 20 kHz drops ~0.5 dB per meter of distance. The particulars of the in-air roll-off shape also depend a lot on RH differences, which could vary dramatically from day-to-day in some climates, like mine. Any published anechoic speaker response measurement will have been performed at some distance (1 or 2 meter or more) from the speaker and should exhibit at least some droop at the top. If the distance, T, and RH are all published, it's theoretically possible to estimate what the speaker itself is doing, but otherwise all you can do is guess. Also, evidence suggests that humans adapt very well to air-induced roll-offs and probably also to roll-offs that "look like" an air-induced roll-off. Harman did binaural room scanning (BRS) studies of distance effects in cinemas. The BRS used measured impulse responses at different rows inside a real cinema to assess perceptual consequences of in-air roll-off. The study concluded that the treble sounded essentially the same at the different distances. A thing to look for in a top-end roll-off is whether it's smooth and gradual or it's abrupt. A lot of compression driver + horn combos drop abruptly above a point somewhere below 20 kHz (often a break-up resonance) , and those speakers will obviously not reproduce the content above the cut well and will also likely audibly ring at the abrupt transition point. Another thing is that when listening indoors with reflections, off-axis response likely contributes a lot to the sound, and compression driver + horn combos very often have significant polar response variation in those upper reaches even if they look pretty smooth on-axis. So just because a response looks smooth on-axis doesn't mean it will sound smooth if the off-axis is rough.

This is an excellent point, and I agree whole-heartedly. Getting the *overall* response (with mains crossed) smooth is crucial. Along these lines, if your mains when measured *outdoors* have a response that rises toward the bass , then you want the subs to maintain that slope if possible. Or if output is limited, at least transition the shape as smoothly as possible and expect emphasis around the transition area.

For outdoor use, a good starting point is response that measures flat. Whatever EQ you used to do that is probably a good starting point to use indoors too. In both cases, it's entirely reasonable to make tweaks to suite taste. Rooms often have ugly resonances that are helpful to EQ out, but I suggest going by ear here more than measurements. Indoor measurements involve a bunch of caveats and are difficult to interpret. And note that it's normal for a sub that's flat outdoors to have a rise toward the bottom when measured indoors.