What watt? Why do we need all that power?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

PRR

Member
Joined 2003
Paid Member
> You're leaving out cone breakup

An overused term. It can mean a whole lot of things. You may have run into a new "breakup" that I have not met.

> the cones on these things have a purely mechanical resonance all their own.

Some of Olson's older papers called this "second resonance". Few speakers escape it. A conical 12" will ring ~600Hz-1KHz; curvilinear significantly higher, paper not so much, some poly rings like a chime.

> you'll get something more like "thup-thup."

Ugh!

If it was a fish, or a melon, you'd throw it back.

I'm not even sure how they'd be plumping up 80-100Hz. It would of course be "commercially valuable": for every one buyer like you, there are 10,000 fans of boom-da-boom thud.

I grant that you would know and avoid the "obvious" mistakes: wall/ceiling resonances can be in this range (I assume you get the same from one driver in a test-box in the driveway), roof reflection (but a wood roof typically would not bounce 80Hz well), room standing-waves, etc.

One thing does come to mind: that much excursion on a rubber surround, it must be a BIG surround, significant radiating area, and possibly a rather low self-resonance. I'd call that bad driver design, but if it bomps-up pop-music thud, then bad is good for profit.

> A stiffer cone would be more massive from the git-go

Actually, a much stiffer cone can be done in 30 or 40 grams (OK, with a much shorter coil). We don't so much want stiff as a higher second resonance: adding mass doesn't raise resonance (it may damp it). What mass seems to buy is gross coil overhang, baroque dustcovers, sexy rubbish doping the paper/felt/poly.

> I knew I was going to need to eq to get as low as I wanted

Nah, hardly any. Several classic drivers with Fs around 20Hz and Q around 0.25-0.5. Put 20Hz and Q=0.5 in an infinite box, you only need a 6dB bump at 20Hz.

> and 15s were rare then (the few that were available were unacceptable due to foam surrounds). 18s? Whazzat?

??? The 15 inch formfactor has been with us since the early 1930s. And we didn't have no foam then, except in our shaving-mugs. The Eighteen is a newer invention, late 1960s? I know they were standard items by the mid-1970s. Both have traditionally been available in two efficiencies (loud or deep) and in two stiffnesses (pro-sound, and instrument-amp).

> Given the need for eq, I knew I was going to need a large Xmax

Fallacy. Xmax is not a function of EQ, it is a function of acoustic power and a (strong) function of lowest frequency. Decide how much air-slosh you want, how much surface you can fit; Xmax comes out of there. EQ is a very different issue.

> I went with Titanics

I confess that the flaws you discovered do not leap to my eye when I look at the Titanic specs. Big 80Hz hump in a semi-infinite baffle "shouldn't happen". Marketing...

--------------------------------

From R.H. Small "Vented box loudspeaker systems, Part II: Large signal analysis"

Pa = kp * Vd^2 * f3^4

Where

Pa is acoustic power, Watts
kp is a fudge; for simple sealed box about 0.8
Vd is displacement volume, cubic meters
f3 is lower frequency limit

Transposing to find displacement volume:

Vd = SqRt( Pa / (0.8*f3^4) )

Using 20Hz for f3, 1 Acoustic Watt design output level:

Vd = SqRt( 1 / (0.8*20^4) )

Do the math:

Vd = SqRt( 1 / (0.8*160,000) )

Vd = 0.0028 cubic meters

Since I measure things with my thumb, convert to old fashioned inches:

171 cubic inches

This says an 18-inch speaker would have to stroke an inch to do 1 watt at 20Hz, which I believe. At least I don't think I dropped a lot of decimal places.

I would expect trouble at 1 inch stroke. In my day, 0.15" stroke was a lot.

You apparently already have a dozen 12" holes in the ceiling. A Twelve has about 70 square inches area, so you have 820 square inches radiating surface. 171/820= 0.2" peak displacement. This is about the extreme limit of the old rubbered-canvas surrounds. It is far short of what they claim these days. On that other hand, a more-sane 0.14" stroke gives only 0.5 acoustic watts, which is still more than most home-size hi-fis will deliver honestly, even ignoring the bass and assuming ample amplifier power. So indeed a dozen 12s is the right ballpark.

As a cross-check: I know a pipe-organ, piped down to ~32Hz. Some time in the past, they were sold a sub-octave, a bank of oscillators running down to 16Hz, a (120W) amp, and two E-V 30W speakers. A 30-inch nominal speaker has radiating surface around 500 square inches, two is 1,000 SqIn, on the order of twelve 12s. Note that the 30W is a 1960s driver, Xmax around 0.15 or 0.2 inches. If such an array worked in a 660 seat cathedral, it should be enuff for home use.
 

PRR

Member
Joined 2003
Paid Member
> honour of my nordic heritage I would of course call it Tor (or Thor) - god of thunder.

I think that name has been taken commercially. Of course you can call your own children whatever you want, as long as you don't sell them.

> Mr. Pass, You seem to have a fablesse for MOS-FETs. Do you by that mean that one shouldn't bother with bipolars?

It is possible to make good sound with any device. But understanding the quirks of each device, in fine detail, can take decades. Many good designers "play the field" in their youth, and then settle down with one specific device family and explore its richness.

Of course this can go very wrong. Pity the designers working in tubes through the 1960s. Several fine tube-amp companies dried up and died. Of the companies who moved into the brave new world of semiconductors, I think most of them had new designers, moving the tube-guys up to Management or out to pasture. Something of the same may happen in BJTs: outside the audio world, BJTs are being displaced by MOSFETs, switchers,or switchers with MOSFETs, IGBJTs, or more exotic devices. Only a couple fabs still cast the big highly doped BJTs we need in large audio amps: OnSemi may be doing it the rest of our lives, and maybe not. Fortunately, usable (not perfect) MOSFETs are in vast production.

If I had to design a subwoofer amp today, I'd just use BJTs. Nothing wrong with BJTs for bulk power, all their flaws can be tamed well enough to be inoffensive in a subwoofer. My main reason would be that I know BJTs better than anything but tubes, and huge tube amplifiers are incredibly costly and heavy and have no audio advantage (and some disadvantage) over BJTs. If I knew more about FETs, I might go that way: they are harder to make "nasty" than a bad BJT amp, and somewhat less prone to self-destruction, though they do need more drive voltage which in a bulk-power amp leads to complication. Neither is "better": depends what the designer is comfortable with (or what supply contracts his/her company already has...)

> I think we might have some old tubes lying around at the department, and I was wondering if they could supply the voltage gain and then utilize darlingtons for current gain? I guess the problem is that they require very different supply voltages... Maybe the tubes won't allow for the Zen-like simplicity...

If you are going to do tubes, DO tubes. It takes a lifetime to really understand the audio aspects of any device. Tube driving sand buffer: been done many times. To my mind, it combines both worlds, including the bad parts of both worlds, plus the extra fun of completely different design approach, different supply voltage, utter lack of serendipity or simplicity.
 
Unfortunately, it's not a fish nor is it a melon. You order from a web company and you're stuck. They're not defective, per se, just not hi-fi. That's not necessarily a valid basis for return with some companies.
Actually, I did the test box on the back deck (21' off the ground), firing at the trees.
The surround is, guessing (I'm at work, so the drivers aren't handy), about 3/4" across. It's not like the drivers you see now where a third of the cone surface is surround. Is it a factor? Absolutely. Is it as much of one as you would suspect by looking at today's long throw drivers? No.
Somehow, you aren't reading the same things I'm writing. I'm not sure what the problem is. Perhaps I need more practice as a writer (I've only been at this for 15 years or so, perhaps clarity kicks in at 20 years...we'll have to wait and see). Let me try this again. Design criteria: Large diameter, long throw, low Fs. Rubber/butyl surround required; foam absolutely unacceptable (jeez, the things are already roughly ten years old, they'd be falling apart by now). Price was part of the equation, but not primary. Were 15s available? Sure. But not as long throw. It was another two or three years as I recall before the long throws 15s came out. Twelve inch drivers were the limit at that time and there were only two or three of them available. Long throw was necessary because of eq throwing the Xmax out in left field. Don't tell me that it's only 6dB, work it out as cone excursion. Then factor in the real world fact that the driver will actually poop out higher than you expect, no matter what you do. Plan on 12dB eq as a bare minimum. Work that out as cone excursion. Then back up and consider how little of that is actually linear. So double the cone excursion just to lower distortion. Then go to multiple drivers for the same reason. (I trust my position on foam surrounds is clear.)
The cracks in my ceiling aren't because the drivers are mounted in the ceiling--they're there because of the music I've played. The drivers are six per side in 7' free-standing cabinets.
Musical instrument drivers? Hmmm. I'm guessing that you aren't a musician, or at least not a bassist or keyboardist. Forgive me if I'm wrong. I've been playing bass for over thirty years. Without getting rowdy about it, I'll just encourage you to measure real world musical instrument speakers. Don't give me theory, go do your homework. In the real world, even the big 15" and 18" drivers aren't worth patootie below about 50-60Hz. The only company I can think of who has made much effort into getting lower than that is Bag End (the ELF stuff), and they use trick tunings and monsterous amounts of eq to do it.
For that matter, it wouldn't hurt to do a few real world measurements on hi-fi subs. I warn you now...abandon all hope, ye who enter here. Leave theory at the door. You'll be disappointed by what you see in the real world. Blame it on inflated specs. Blame it on incomplete theoretical models. Doesn't matter to me. But please...please...check the reality out before you start throwing so much theory around. It doesn't match what's really happening.

Grey
 

PRR

Member
Joined 2003
Paid Member
> Somehow, you aren't reading the same things I'm writing. I'm not sure what the problem is.

I know you hate foam.

> work it out as cone excursion.

Perhaps I need more practice as a writer (I've only been at this for 40 years or so, perhaps clarity kicks in at 80 years...we'll have to wait and see). Let me try this again.

PRR said:
From R.H. Small "Vented box loudspeaker systems, Part II: Large signal analysis"

Pa = kp * Vd^2 * f3^4

Where

Pa is acoustic power, Watts
kp is a fudge; for simple sealed box about 0.8
Vd is displacement volume, cubic meters
f3 is lower frequency limit

Transposing to find displacement volume:

Vd = SqRt( Pa / (0.8*f3^4) )

Using 20Hz for f3, 1 Acoustic Watt design output level:

Vd = SqRt( 1 / (0.8*20^4) )

Do the math:

Vd = SqRt( 1 / (0.8*160,000) )

Vd = 0.0028 cubic meters

Since I measure things with my thumb, convert to old fashioned inches:

171 cubic inches

This says an 18-inch speaker would have to stroke an inch to do 1 watt at 20Hz, which I believe. At least I don't think I dropped a lot of decimal places.

A Twelve has about 70 square inches area, so you have 820 square inches radiating surface. 171/820= 0.2" peak displacement.

I just don't see how you need large excursion, even at 20Hz, from 800 square inches of radiator.

> You'll be disappointed by what you see in the real world.

In my youth I did earn my supper by designing and modifying speaker systems. I done the homework, in the real world, and took the blame when it didn't work as predicted (mostly setup-error or shipping damage). As salaryman, I have often recorded large pipe organs and do know the difference between live and speakers. I recently recorded a modest organ playing Prelude and Fugue in A minor, BWV 543, Johann Sebastian Bach (1685-1750), Shea Velloso (96K, 7Meg MP3 file). About 25 seconds in: you should hear the fundamental, most folks can't.

> I'm guessing that you aren't a musician, or at least not a bassist... Without getting rowdy about it,... Don't give me theory, go do your homework. In the real world, even the big 15" and 18" drivers aren't worth patootie below about 50-60Hz.

PRR said:
...have traditionally been available in two efficiencies (loud or deep) and in two stiffnesses (pro-sound, and instrument-amp).

I apologize for not writing clearly here. Instrument speakers are STIFF; the Fs for a Fifteen is generally 40-50Hz for quality speakers, perhaps higher for PartsExpress-market MI speakers. Twelves will typically be tuned higher. No, I was NOT suggesting you use MI woofers: merely noting that several varieties of Fifteens have been around a while, and you have to be careful which you get.

Generating an honest 20Hz is not a new trick, but it has always been rather an outer-fringe system because of the vast area and volume required to do it at all.

> 7' free-standing cabinets.

I also apologize for thinking you had some semi-infinite baffle. The required back-chamber for 20Hz is room-size, not cabinet-size; almost exactly the size of the room I am typing in.
 
Once upon a time, I cobbled together a little circuit I chose to call the Aleph-X. I posted the thing here. The first page of the thread was filled with the usual 'Wow, cool, tell-me-more' kind of comments that any thread with a schematic seems to accumulate. The next ten or twenty pages were filled with a few hundred posts telling me that the thing wouldn't work because of this theoretical point or that, or because someone's simulation program said it wouldn't.
But, you see, I had a secret weapon. I had a real-live, functioning prototype. I knew the thing worked, no matter what people said. And reality beats theory seven days a week in my book.
The same applies here.
One keyword you appear to have missed in my earlier post is 'linear' Xmax. I do not feel that the Xmax specified in the ratings is acceptable for hi-fi reproduction. I automatically cut it in half--frequently more. I have noticed that drivers near their so-called linear limits sound pretty loose. So I use them at considerably less than their rated Xmax. Now, if you feel that you want to use all the rated excursion, that's fine, but I don't want to. The first generation Titanics had (quoting from memory) a 1" Xmax p-p. Once I derate that, I'm pretty near the limit.
You're also making the implicit but invalid assumption that there's a brick wall HP filter at 19.99Hz. In your estimates, a 12 or 15Hz subharmonic/beat frequency/synthesizer tone/sub-component of a drum strike (Fourier analysis, here) doesn't come along and upset the apple cart. Give that excursion increases by a factor of four per octave, we're well past your .2" figure. Even assuming that your eq is letting go and the response is beginning to taper off, you'll bottom your cones. Perhaps that's acceptable in your book. It's not in mine.
Again, too much theory. Not enough reality.
I once participated in a thread wherein some of the posters felt that recording folks were automatically high-end oriented and good critical listeners. My experience suggests otherwise. I only know one recordist who wears both hats comfortably. The rest? They don't. You may...or you may not. I honestly don't know. But making a recording (or even dozens of them) doesn't automatically qualify anyone for anything except the ability to plug in a mic. (As evidence, I offer nearly the entire catalog of any given record label--and those guys do it for a living, too.) Again, just my point of view.
Somehow you are not understanding the problems with this class of drivers. I'm sure the failure to get my point across is mine, and for that I apologise. Rather than have this turn into a cat fight, I'll sign off and let you theorize to your heart's content.
Besides, for all I know the current generation of long throw drivers is better for hi-fi use. Perhaps it's all moot. Wouldn't that be neat?

Grey
 
Disabled Account
Joined 2005
GRollins said:
Once upon a time, I cobbled together a little circuit I chose to call the Aleph-X. I posted the thing here. The first page of the thread was filled with the usual 'Wow, cool, tell-me-more' kind of comments that any thread with a schematic seems to accumulate. The next ten or twenty pages were filled with a few hundred posts telling me that the thing wouldn't work because of this theoretical point or that, or because someone's simulation program said it wouldn't.
But, you see, I had a secret weapon. I had a real-live, functioning prototype. I knew the thing worked, no matter what people said. And reality beats theory seven days a week in my book.


the same thing here. I was reading with amazement in the old JLH thread where so many "experts", self-claimed or otherwise, insisted that the MOSFET version of it never worked, because the JLH was designed on "current drive" not "voltage drive", blahblahblah.

Never mind that I had one signing in my base for 10 years, and the JLh is nothing but the Williamson tube (voltage drive) implemented with BJTs (current drive).

We have way too many "experts" here who are great at claiming "theoretical impossibilities" vs. having an open mind.
 

PRR

Member
Joined 2003
Paid Member
> We have way too many "experts" here who are great at claiming "theoretical impossibilities" vs. having an open mind.

Sorry. I thought there was a question about woofer power and performance demands, and a $1,600+ system that does not meet the builder's goals, but "derate the heck off the specs" and "no foam" is clearly more important than understanding the reasons "why". Bye.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.