One hell of a 12” woofer!

[AOR]

Here's a peek at the woofer I designed for an upcoming 12" powered subwoofer at PS Audio. I was shooting for low and linear inductance and we have 30 mm Xmax and 45 mm suspension travel that is very symmetrical.

View attachment 435159View attachment 435162
View attachment 435161

Very nice. Reminiscent of a car audio sub with that woven surround foot.

 

[Chris Brunhaver]

Very nice. Reminiscent of a car audio sub with that woven surround foot.

Yes, that cone/surround glue joint is reinforced in a few different ways on different kinds of subs for long term reliability reasons. We also have a small paper reinforcement at/under the triple joint.

The basket was a tooling that used to be used by TC Sounds for the Eclipse car audio titanium series.

 

[egellings]

Not a ton of Xmax by modern standards, but it can take lots of power.

Xmax is just the excursion one way, not both ways totaled, right?

 

[Salt]

Xmax is the distance from resting positon to max extension of VC within the magnetic field, and therefore sometimes marked by +-.
It is not the most possible extension of the driver, as that is limited pyhsically and beyond Xmax.

 

[Chris Brunhaver]

Xmax is the distance from resting positon to max extension of VC within the magnetic field, and therefore sometimes marked by +-.
It is not the most possible extension of the driver, as that is limited pyhsically and beyond Xmax.

In the last 20 years, this has been refined and the accepted standard is that which Klippel recommends.

https://www.klippel.de/fileadmin/klippel/Files/Know_How/Application_Notes/AN_05_Displacement_Limits.pdf

For a subwoofer, this is the least of the three limits where the BL hits 70% of rest, Cms hits 50% or Le(vs)X changes by 10%. This equals 20% distortion.

The mathematical Xmax of voice coil overhang and of when a coil is "in the gap" doesn't take into account pole piece geometry and if the coil is aligned to the magnetic center etc.

I would replace the word "extension:" above with "excursion" and this is often called Xmech, or the mechanical limit of the driver (not considering sound quality/distortion.

 

[jamescarter1982]

Sealed is the better choice for your application then.

Something that is often overlooked is that, given a suitable woofer, a vented box sub can be sized and tuned such that its native response is the approximate inverse of anticipated room gain from boundary reinforcement. Taking the room into account results in less low end "on paper" than if you had sized and tuned the box to maximize low-end output.

You can think of it this way: Instead of building a sealed box and raising the very bottom end with EQ, you build a vented box whose in-room response is about where the equalized sealed box would have ended up. This type of vented-box tuning results in less group delay than more typical tuning strategies, theoretically often below detection thresholds, though still not as low as a sealed box.

is this an extended bass shelf your referring to here pls

 

[Head_Unit]

is this an extended bass shelf your referring to here pls

Well no I don't think @Duke is. Yes EBS is "undertuned" but usually when I see those they are not monotonically decreasing. They start to roll off, then flatten and maybe have a bit of a bump, then decrease fast. I feel that indicates time domain behavior that is less than optimal (though I must admit that is based on my engineer's intuition and not hard data). I prefer something that keeps decreasing monotonically, kind of like fattening the bottom of a sealed box curve. Sealed and ported can both become WAY too obsessed with the -3 dB point, to the detriment of both time response and actual in-room frequency response. The F3 point is just a convention borrowed over from the electric filter design theory which Thiele and Small and Benson took speaker design out of the dark age by adapting to physical speakers bless them bless them bless them. I believe the -6dB and -10dB points are much more important due to room gain; when I mentioned that once to Richard Small on a business trip to Harman Indiana he agreed with that.

 

[Head_Unit]

woofer I designed for an upcoming 12" powered subwoofer at PS Audio

Damn Chris that is nice work! But I think you need a bigger surround. And more cowbell!

 

[Duke]

is this an extended bass shelf your referring to here pls

No; see @Head_Unit's explanation in post #127 above. He nailed it better than I could have.

 

[jamescarter1982]

Sealed is the better choice for your application then.

Something that is often overlooked is that, given a suitable woofer, a vented box sub can be sized and tuned such that its native response is the approximate inverse of anticipated room gain from boundary reinforcement. Taking the room into account results in less low end "on paper" than if you had sized and tuned the box to maximize low-end output.

You can think of it this way: Instead of building a sealed box and raising the very bottom end with EQ, you build a vented box whose in-room response is about where the equalized sealed box would have ended up. This type of vented-box tuning results in less group delay than more typical tuning strategies, theoretically often below detection thresholds, though still not as low as a sealed box.

is this an extended bass shelf your referring to here pl

Well no I don't think @Duke is. Yes EBS is "undertuned" but usually when I see those they are not monotonically decreasing. They start to roll off, then flatten and maybe have a bit of a bump, then decrease fast. I feel that indicates time domain behavior that is less than optimal (though I must admit that is based on my engineer's intuition and not hard data). I prefer something that keeps decreasing monotonically, kind of like fattening the bottom of a sealed box curve. Sealed and ported can both become WAY too obsessed with the -3 dB point, to the detriment of both time response and actual in-room frequency response. The F3 point is just a convention borrowed over from the electric filter design theory which Thiele and Small and Benson took speaker design out of the dark age by adapting to physical speakers bless them bless them bless them. I believe the -6dB and -10dB points are much more important due to room gain; when I mentioned that once to Richard Small on a business trip to Harman Indiana he agreed with that.

thankyou so I think you are saying a ported cabinet can be designed to roll off in a similar way to a sealed box . how is this achieved. if its a chore for you to explain it could you link me somewhere I could read about it . im interested in making loudspeakers and this idea interests me .ive heard manufactures describe this I think they were saying tune really low and then low cut the bass below the system resonance I think . not sure if im getting the two ideas mixed up
thankyou in advance

 

[Duke]

is this an extended bass shelf your referring to here pl

No, as I said before, it's not an "extended bass shelf" alignment.

It might be called an "inverse room gain" alignment. If the speaker's native response is approximately the inverse of room gain then, to a first approximation: Your speaker's frequency response + room gain = your desired in-room frequency response.

im interested in making loudspeakers and this idea interests me

Is this for a DIY project, or for a commercial venture?

 

[jamescarter1982]

No, as I said before, it's not an "extended bass shelf" alignment.

It might be called an "inverse room gain" alignment. If the speaker's native response is approximately the inverse of room gain then, to a first approximation: Your speaker's frequency response + room gain = your desired in-room frequency response.




Is this for a DIY project, or for a commercial venture?

thankyou that makes perfect sense now thankyou .
Excuse me i didnt ask if it was an extended bass shelf again I think that i mistakenly replied to the post when I originally asked you the question or something and it put it again. I was listening to your first reply just wanted a bit more info . thanks again

 

[Salt]

?

 

[Head_Unit]

thankyou so I think you are saying a ported cabinet can be designed to roll off in a similar way to a sealed box . how is this achieved. i

To answer this, and why it isn't done more, we have to go back to The Dawn Of Time. OK not that far back, maybe 1960s? In the 1950s speaker designers did not know what the hell they were doing, they were literally like the early alchemists, playing around and trying to derive rules of thumb. One of my former bosses, Jim (James) Novak, achieved some speaker fame by quantifying some rules of thumb. Meanwhile, Down Under, J. E. Benson of Amalgamated Wireless Australasia analyzed loudspeaker systems mathematically, analogous to electric filters, whose theory was by then well understood. Some of this was published in Amalgamated Wireless Australasia Technical Review.*

I *think* Benson was really the first, but in those pre-internet days those insights did not circulate much so it was Neville Thiele and Richard Small who picked up the baton and really cemented the idea of taking the physical properties of speakers (magnetic power, moving mass, surround/spider compliance/stiffness, resistance) and transmuting them into electrical filter elements. This gave rise to loudspeaker designers using common electrical filter alignments as design targets, including infamously rather slavishly chasing the lowest -3 dB point which is an electrical filter convention.

This work was really seminal stuff. It was the difference between going to a bowling alley and just throwing the ball down the lane, often guttering, versus taking lessons to learn about stance and throw and what the marks on the lane mean. Purposeful design became possible. Nowadays there are simulation tools available and we don't have to be slaves to electric filter alignment tables any more.

SO back to the actual question, I do this kind of design empirically. I'll start with a sealed box for roughly a Q=0.7 kind of target and say hmmm let's make the box bigger and tune a port to 30 Hz.** I look at the response to see that it monotonically decreases. Usually there is something wrong-like as big a port I want does not fit-and I iterate, iterate, iterate until I've extended the bass as much as possible. I feel like I am missing something in this description so if not clear please ask Note: in some way this is kind of an Extended Bass Shelf, except I smoothen it out to keep decreasing so there is no "Shelf" at all.


** (I don't like to tune much higher than that probably due to mental audio scarring from the line of Bazooka subwoofers. Inventor Dave Profitt (sp?) told me the design came to him in a dream from God, and his insight was you did not need to tune lower than 41 Hz because that is the lowest note on a bass. That is true and efficient for loud output with low excursion BUT sounded droningly awful due to time decay problems.)

* as referred to somewhat tangentially here, super math-dense stuff https://audioxpress.com/article/voice-coil-focus-quasi-alignment-families

 

[LSPhil]

Not to mention the low quality copper windings on the voice coil formers of the Daytons. People don’t realize that when you’re dealing with a company that’s been making drivers for 100 years, they learn some tricks of the trade along the way. Such as ways to wind the voice coil wire on the formers, adhesives used, materials used etc. These generic Chinese companies (like Dayton sources their drivers from) entire driver portfolio is based on copying everyone else. And they cut corners in many ways regarding durability. Last time I tried a beefy looking Chinese woofer in a small sealed cabinet with DSP EQ, the voice coil burnt up within 5 minutes of testing that wasn’t even that hard.

Some may also argue that some of the old school European manufacturers may have an in house owned Chinese manufacturing facility. Even if they do it only means they moved the same quality manufacturing equipment , same quality knowledge, and same everything somewhere the labor is cheaper. Still all R&D takes place in Europe. Not so when you’re dealing with the big Chinese OEM’s. Everything they know was ripped off, and typically corners cut. For DIY ok. Not ok if you’re putting your business rep on the line.

I’m curious how the voice coil burned within five minutes, what test signals and operating conditions led to that failure?

 

[ChrisG]

Easy enough with sine waves, especially in a small sealed box - it's probable that most/all of the sub's bandwidth was below Fc. ie, low-impedance.

 

[LSPhil]

Easy enough with sine waves, especially in a small sealed box - it's probable that most/all of the sub's bandwidth was below Fc. ie, low-impedance.

I agree with that explanation and would like to add some context.
Pure sine waves are extremely stressful for a voice coil because they represent a continuous 100% duty cycle with no crest factor. That means the driver is exposed to full RMS power all the time, which leads to very rapid thermal buildup.
Music behaves very differently. Even bass-heavy material has significant dynamics, so the average (RMS) power is far lower than the peak power. A rough comparison illustrates this well:

100 W sine 100 W 100 W Pure continuous operation, typical measurement condition

Pop music ~20–30 W ~80–120 W Amplifier operates near its limit, clipping possible at full volume

Classical music ~1–2 W 200 W+ Very high crest factor, requires substantial amplifier headroom

So even music with strong, artificial bass typically averages only around ~20% (or less) of the rated RMS power, whereas a sine wave delivers 100% continuously. This alone explains why a driver can survive music playback but fail quickly during sine testing.
In a small sealed enclosure, the situation becomes even worse. Below Fc: acoustic loading decreases
current draw increases (often a low-impedance region)
At the same time, during continuous sine operation the air inside the enclosure heats up very quickly and cannot be exchanged. Instead of cooling the voice coil, the trapped air becomes a thermal insulator. A large portion of the heat normally leaves via: the cone and convection to surrounding air
You’re essentially lucky if the cone is metal, as it can conduct some heat away. Otherwise, thermal runaway can occur very fast.
With bass-reflex designs, one could expect slightly better thermal.
So yes, destroying a voice coil within minutes is entirely plausible when using high-level sine waves, especially in a small sealed box and below Fc. Music playback at the same nominal power rating is far less demanding thermally.

 

[ChrisG]

Just a small note that only DC or square waves are 100% duty cycle, since they're always at +/-100% signal.

We use RMS calculations on sine waves to find the DC voltage that delivers the same heating power.


I generally agree with the rest of your post, although I'd use different language around "classical" vs "pop", and examine the crest factors of specific pieces of music, and using that to relate back to amplifier power outputs. I'm sure there's some popular music that has a decent dynamic range, and I'm sure there's some classical with relatively little.

 

[egellings]

100% duty cycle for DC is okay; for a square wave, no. Duty cycle is defined as the ratio of the on-time (or off-time, if that's what's being considered) during one cycle of the waveform to the time needed for one complete, whole cycle. It is a ratio of two time lengths.

 

[ChrisG]

A symmetrical square wave (ie, no DC offset) certainly is 100% duty cycle.