Subwoofer Driver choice?

[efzauner]

Hi everyone

I am trying to decide what drivers to use for a subwoofer for audio and HT. I would like your thoughts on the design process. My HT room is 16x12x8ft. That means a room gain frequency of about 565/16=35 Hz. I plan on using 2 or 4 subs driven by a Pro Audio amp (Mackie, Beringer, Crown etc) so I do not need a single ultra sub in the range of JL12W7 or other really high Xmax sub. I think that several subs working at medium levels in each room corner would be bether than one big sub running at high power with associated power compression and distortion. My mains are LaScalas with Heresy surrounds and center.

With the 35Hz room gain +3dB frequency should I design a sub that has a -3dB point also at 35Hz and try to minimize any low frequency hump? Or should I simply try to make a sub that has the lowest -3db point?

Attached are 2 WinISD design trial with 4 different subs. The first are sealed designs and the second are vented Designs. 2 of the subs are commonly used for home DIY subs, the other 2 are car audio subs. These are subs in the $100 to $200 range.

If I go with a closed design, the Madisound driver gives great low end response, but with a large cabinet. The others have -3dB points in the 35-40Hz range. The Infinity car sub is actually the cheapest of the lot and no slouch!

If I go with a vented design, the Madisound seems out of place, seeming to prefer a sealed design. The others all dig quite low with -3dB points at 18-22Hz.

I hear about many DIY going for the Dayton Titanic in a sealed cabinet but I dont understand why the performance is so great. I also see that most commercial subs are vented designs and they all seem to go very low.

All else being equal (sensitivity, xmax, max power etc) should I go with a sealed Madsound or a vented design with one of the other subs. Should I go for a vented design and pick the sub that gives the highest output given its Xmax, sensitivity and max power?

[efzauner]

Here are the Vented designs

[michael hurd]

efzauner: did you model those curves with a voice coil temperature rise??? I model with at least 100 degree C temp rise, take a look at the excursion and the transfer function curves.

As the motor heats up and the motor strength dwindles, the woofer " likes " a bigger box.

[michael hurd]

I would not put a subwoofer in each corner, put them all ( however many you are using ) in a common box, in one corner. The infinity 1230w ( now replaced with the 1240 ) did well in a 2.77 cuft sonotube, this was in a car, tuned to 30 hz. It actually modelled well in a big (over 4 cu ft box ) for HT use, four of them should have near 120 db @ 20 hz from the simulation, staying under xmax.

[efzauner]

efzauner: did you model those curves with a voice coil temperature rise??? I model with at least 100 degree C temp rise, take a look at the excursion and the transfer function curves.

As the motor heats up and the motor strength dwindles, the woofer " likes " a bigger box.

No, can you do that with WinISD? did not think so. WinISD beta does not even give you excursions, which would be important to know in a vented design. Does the heat in the winding simply raise resistance of the coil, dependant on the design and cooling, thus draw less power?

But is my though process correct? I was expecting much better low end performance from some of the subs that people rave about on sub DIY sites. The infinity reference series is low end in comparison and seems to perform much better.

Why do you suggest all in one place? I thought it woudl be easier to controll room modes with several subs, plus 2 tall cub cabinets would make great stands for the rear Heresy's!

The basic question, do you go all out for low end performance with a vented design or do you try to tune it to match the room effect?

[michael hurd]

I have Win ISD pro, it is free. To model excursion, you have to know Sd and xmax. In WinISD pro, yes you can model with a voice coil temperature rise, I do this when I model.

The voice coil temperature has a direct relationship to Re, as Re goes up, motor strength falls, yes. With a vented design, I try to wring most of the theoretical performance out of the driver, watching the excursion and transfer function curves closely, as well as power input. I know that the tubes that I built worked great at 2 previous residences ( both open floorplans ) are a little too much for a small room now.

Now that I have built a few, I would try to build to suit a room slightly larger IE: tune lower. You can always shorten the port and raise the tuning frequency. Having stereo subs versus co-location... that is a topic unto itself. Myself, with a low crossover point, I would be tempted to try stereo subs, but with a common 80 hz crossover ( like in most HT receivers that is non-adjustable ) I would co-locate.

[Jay481985]

use the alpha version, beta sucks.

[DrWho]

For starters, what's the budget? And do you already own the proamp?

Ultimately,
you want to use as large of an enclosure as possible - choosing a
target cabinet size and shape is always the first place to start.

Next,
you want to decide how low and how loud you want it to play. The lower
you go, the more expensive it will be for the same SPL. I generally
target 120dB down to 20Hz.

Once you have these two criteria
determined, it's just a matter of finding a driver on the market that
meets your expectations in your cabinet. You probably won't find a
perfect match, but that's when you go back and tailor the cabinet back
to the driver you end up with - but you still start with the cabinet.
When the driver manufacturers design subs - they don't just randomly
throw parts together....they optimize the driver to work within a
specified design. So going by the manufacturer recommendations is a
good tool to make sure you're not doing something crazy.

As far
as room gain - don't count on it. Yes it exists, but nowhere near to
the level that the ideal world predictions show. Also, there are some
psychoacoustic effects involved too: namely that any instrument that
you put into the room would also experience the effects of the room
gain - so if your sub rolls off to compensate for the gain, then it's
not going to sound like the instrument is in your room. And then couple
that with the Fletcher-Munson equal loudness curves and there is even a
larger case for having a little boost at the bottom of the response:
http://sengpielaudio.com/Acoustics226-2003.pdf
Also,
room gain doesn't start to happen until you start cranking the
volume...usually it's gotta be over a good 90dB before you start
noticing it - which in a way is a good thing when you consider the
nonlinear behavior of the system...

As far as building multiple
subs....I fully endorse the idea. The biggest difference is going to be
a change in the modal distribution of the room:
http://www.harman.com/wp/pdf/multsubs.pdf
And
then you have the nonlinear behavior of each driver. Increasing the
overall surface area of the system is going to displace the same amount
of air, but minimize the cone excursion in the process. Ultimately,
every form of distortion increases with excursion. It doesn't guarantee
that two drivers are better than one, but it requires much less
precision with the motor design for the two "weaker" drivers. But I
don't think general rules can really be applied here because there are
bound to be exceptions.

As far as the final cabinet design, I would also recommend using winISD Pro Alpha:
http://www.linearteam.dk/default.aspx?pageid=winisdpro
Now only does it allow you to model more behavior, but it's also more accurate too.

One
of the most important things most DIY builders never consider is the
influence of the amplifier on the subwoofer system (which is why I
asked if you already purchased the amps). You absolutely must use some
form of EQ - especially for ported systems. You need to control the
cone excursion so that you minimize the nonlinear behaviors. One of my
most recent projects involved a comparison of two subwoofers with 1000W
of proamp behind them versus a single driver with a customized 250W
subwoofer amplifier - the single driver with the smaller amp sounded
infinitely better because it had the benefit of EQ and better control
over the cone excursion. Though it's tempting to shoot for high SPL's
and low extension, it's good to exercise some restraint and back off
just a notch.

That's not to say that a proamp can't be
used....but you're going to want to put some kind of processing ahead
of the amp. But usually you can get about the same amount of power for
the same price and get the processing built into the dedicated
subwoofer amps.

And finally - you want to predict non-linear
behavior as best as you can. The temperature of the voice coil is but
one of many factors that are going to change. Another big one is port
compression - which both reduces the output of the port while also
lowering its tuning point. At higher excursions the effective Sd of the
driver decreases, the Bl decreases, the stiffness of the suspension
usually increases, etc etc...a good design isn't necessarily going to
account for all of this behavior (it's impossible really), but you can
choose alignments that try to keep them from blowing up. For example,
you'll notice that some drivers will have huge peaks resulting from VC
temp rise while others won't be affected so much.

You can't
accurately predict nonlinear behavior and you're not going to be able
to compensate for it anyway - but it's important to keep in mind. Once
you start playing around with variables you'll see what I'm talking
about. Each one has a distinct influence on the frequency response. If
you'd like to learn more, here is a great resource:
http://www.klippel.de/pubs/default.asp
And
if you don't want to get caught up in the math, just browse through the
charts and note the behaviors. He also mentions the general trends
which you can manually enter into WinISD.

Ok, enough theory...one driver you might consider for its insane bang for the buck is the Assassin 12" from Ascendant Audio:
http://www.ascendantaudio.com/assassin12specs.html
I
just got finished building 16 subs using these drivers and have been
more than impressed. If you put a pair in each enclosure, then I would
recommend 8 cubic feet tuned to 20Hz with a +3dB peaking 2nd order
highpass at 24Hz to control the cone excursion. A pair of these 4"
flared ports at the full 17" length should be fairly adequate:

http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=268-352

Another alternative would be a pair of the Dayton Titanic MKIII 12" drivers in the same cabinet with the same ports:

http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=295-404

The only difference is that you'd only need a normal highpass filter (Q
of 3 at 18Hz) to achieve the same response as the Dayton - you just get
1dB more output for $60 more per driver. The goal here would be more
linear behavior. But this also gives you the opportunity to make the
cabinets smaller...For example, you could move to a 6 cubic foot
enclosure, keep the same 20Hz tuning but go with a +3dB peaking 2nd
order highpass at 24Hz. The only downside is your dual 4" flared ports
would need to be over 26" in length [] Nevertheless, it gets you the
same peak output as the Assassins in this configuration.

Port velocity for all three configs is under 30m/s at peak output.
Also, WinISD predicts the half-space response of the system....if you
put the sub in a corner, then it will effectively be in 1/8th space. To
accomodate, you can drop in a 6dB static gain to get an idea of the
true 1m output in the room. I tend to ignore it and trust that I'll
lost about 6dB between the listening position and the position of the
subs in the corner. But if you wanted to compare against the specs of
other manufacturers, then you'll want to add 6dB. The systems I
proposed would have F3's around 20Hz and about 123dB peak output. Build
two of each enclosure and you're looking at an ideal 129dB [] (that's
Klipsch THX Ultra2 performance...)



However, ignore the EQ and you're looking at peak clean SPLs on the order of 120dB []

Once you finalize the design of your system, then it's just a matter
of finding an amp that does what you need it to do. For example, the
Dayton 1000W amps work perfectly with the Dayton drivers:

http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=300-808

http://www.partsexpress.com/pe/pshowdetl.cfm?&Partnumber=300-810

The amps from Rythmik Audio work real well with the Assassin drivers:

http://www.rythmikaudio.com/nonservo_product.htm

You would want to go with the 250W version with the $5 addon filter, or
the 350W special addition if you wanted a little headroom. You would
want one amp per driver, but it's still cheaper than going with the
Dayton amps. However, they probably won't be a good match to the Dayton
drivers.

Anyways, I hope this helps. Everyone has their own opinion on how
things should be done so I don't propose this as the only way to go
about subwoofer design. It's just what works for me. Hopefully it will
spur some discussion though and then you can choose what compromises
you want to make.

[DrWho]

Man that ended up as one long post - yikes []

Anyways, I wanted to mention to not forget about TC Sounds:
http://www.tcsounds.com/tcdrivers.htm

Though they've got some insane products out there, their TC driver lineup is more on the normal side and has products at each price point. For example, the db-500 is a "copy" of the Assassin 12" - which both are arguably copies of the old Adire Audio Shiva (which was very highly acclaimed - just more expensive with less performance). Generally speaking, you'll be able to go with smaller cabinets with TC Sounds, but you're going to need bigger amplifiers too.

[efzauner]

Wow, thanks Mike. That really covers the basics.
So I summarize:

Go for vented, determine how big a box I can have, then fit a driver that works best with that box and has the best overall SPL/low end. Use an amp designed for Subs because they have built in HP filters to prevent overexursion below the vent tune frequency, and of course take into account increased R due to heating. So my thought process so far was in line. I was just expecting a much bigger differenc in simulated low end performance between a very expensive JL driver and a cheap one like the Infinity. I was also expecting the same kind of thing with the dayton sub. So bottom line, I need to sweat the details.

Thanks a lot for the input. I will let you know what I do. BTW the HT room is not built yet. I am currently using just 3 of the Heresies in LFC and some cheap surrounds in the playroom wich is the same dimesions. Powered by a 40wpc pioneer digial 5.1 amp. And that is too loud already. I think that 2 12 inch subs off a 300wpc amp will be aplenty!

[DrWho]

The difference between say a JL and the three drivers you modelled (minus the dayton) is the excursion capabilities of the drivers. More excursion = more SPL. So though the one goes extremely low, it's not going to go extremely loud (in fact, i would expect under 100dB which is too little).

You really want to be using less than 1/4 the power handling and 1/2 the cone excursion at normal playback levels (and even less is better).

But ya, your summary pretty much nails it. Just one minor thing though...it isn't required that you use a dedicated subwoofer amp. They just tend to come with all the necessary features. If you went with a pro amp (which generally speaking should have more output, less distortion, and all for a cheaper price), then you will want to get some kind of processing to put inbetween the receiver and the amp. Something like say a BFD might be a very feasible approach. Just make sure you find something that can do processing down to 10Hz or so.

Here's a link that talks about the BFD:
http://bfdguide.ws/

Though EQ is nice - it won't be able to fix time-domain issues that show up in the frequency response (so even though you can make the graph measure flat, you're still not fixing the time domain which is going to be what your ears are actually hearing). Anyways, that's a totally different topic and up to a lot of debate. []

[sfogg]

"My HT room is 16x12x8ft. That means a room gain frequency of about 565/16=35 Hz. "

If you have a solid well sealed room trying to take advantage of room gain can be very effective. But remember, you need a second order rolloff matched to where room gain kicks in for it to work. That means a sealed subwoofer design.

I do this in my theater and it is very effective. So much so that the sealed subs with a designed f3 of 33hz (Q 0.5) go deeper *in the room* then a ported design with a f3 of 16hz.

Shawn

[sfogg]

Mike,

"Though EQ is nice - it won't be able to fix time-domain
issues that show up in the frequency response (so even though you can
make the graph measure flat, you're still not fixing the time domain
which is going to be what your ears are actually hearing). Anyways,
that's a totally different topic and up to a lot of debate."

Not so. You just need to target room resonances very carefully and need very precise EQ abilities and you can help reduces time domain issues from room resonances. Put less energy into the room at a resonance point and it doesn't resonate for as long as it did before.

Read the papers for both Lexicon and Meridian's room EQ solutions and they both take the same approach to handling time domain issues. Flat frequency response is not eithers goal but reducing unnaturally long decay time is.

Shawn

[DrWho]

on the topic of room gain...

I know we've had some discussion on this before, but isn't measuring the final response with an RTA and pink noise a bit misleading since it gives the room a lot of time to get excited and build on itself? With transient signals (as is typical with music and movies), the reflections causing the gains aren't going to be arriving at the same time and you're not going to realize the compounding effect....if anything, you're just going to have a room that holds onto the notes a bit longer. Granted, as the signals get longer the gains will start to increase - but it's going to be a ramping effect and not as dynamic as an "instant" build up. Also, the few rooms I've measured indicate that the gains don't start happening until you start cranking the SPL - probably has something to do with the walls' ability to damp lower level signals?

Wouldn't it be better to design a system to be flat to 10Hz or whatever the desired cutoff is and then use EQ to back off any extra room gain? In the process you're going to be reducing the distortion of the system - something you can't do when relying on room gain in the first place.

[DrWho]

Not so. You just need to target room resonances very carefully and need very precise EQ abilities and you can help reduces time domain issues from room resonances. Put less energy into the room at a resonance point and it doesn't resonate for as long as it did before.

Read the papers for both Lexicon and Meridian's room EQ solutions and they both take the same approach to handling time domain issues. Flat frequency response is not eithers goal but reducing unnaturally long decay time is.

Using EQ does not change the behavior of the decay at all....it just makes the tail hit the noise floor sooner. In systems with very low noise floors, you're not going to be able to tuck it away without destroying the intial direct signal. At least that's what I'm reading from the guys designing the new time based measuring tools...

I'll have to look up the papers though...maybe they're doing something besides just straight up EQ?

Links would be appreciated, but I might be able to find them.

I just wanted to add that I'm not saying that EQ can't subjectively improve the situation - it's just not a true objective solution. It's a band-aid. I would love to sit down and A/B the differences between proper acoustical treatment and EQ in the same room. I'm sure there is a point of diminishing returns, which needs to be factored into the bigger picture.

[CONVERGENCE]

Theaters don't RTA subs. The Fr range is usually Any where from 20 to 200 Hz . They measure the SPL with

a meter and set at between 80 and 90 DB. It should match the rest of the system So do an SPL check with out the subs . Of course you should RTA your system if your main speakers and surrounds separately and average it out. Theaters use a multiplexer to do this .The EQ is 1/3 octave 27 band.

But all this is professional . All you need to adjust is the SPL level of your subs with a meter.

[sfogg]

"I know we've had some discussion on this before, but isn't measuring the final response with an RTA and pink noise a bit misleading since it gives the room a lot of time to get excited and build on itself?"

Room gain kicks in on the initial cycle of the wave. That is basically why you get the room gain... the reflections within the room are inphase with what is basically the same cycle of the wave. The more inphase the wave gets the more room gain gain you get... that is why it kicks in below the longest resonant point in the room and increases as you go lower in frequency as the wave gets longer and longer and therefor couples more with itself.

"- but it's going to be a ramping effect and not as dynamic as an "instant" build up."

That is not my experience at all. In fact I'd say my sealed subs that utilize room gain sound more dynamic then the ported sub. Remember... at the top end of a subs bandwidth you aren't using room gain, just down deep. To really take advantage of room gain you need subs with a classic second order rolloff... so a Q around 0.5... that results in a punch/dynamic sounding subwoofer.

"...if anything, you're just going to have a room that holds onto the notes a bit longer. "

Nope, room gain is below the rooms resonance points.

"Also, the few rooms I've measured indicate that the gains don't start happening until you start cranking the SPL - probably has something to do with the walls' ability to damp lower level signals?"

Possibly, and also leaking bass out reduces room gain. That is why I asked about how solid his room is. If it is a very lossy room then room gain isn't as effective.

"Wouldn't it be better to design a system to be flat to 10Hz or whatever the desired cutoff is and then use EQ to back off any extra room gain?"

If you are worried about the time it takes for room gain to build up why wouldn't you be concerned about the exact same thing here? If you think this is an issue then in this case your initial response would be cut by the EQ until room gain kicks in and be no different then a system that utilizes room gain.

"? In the process you're going to be reducing the distortion of the system - something you can't do when relying on room gain in the first place."

How do you figure? If you EQ out some of the output of the driver to account for room gain how is that different from having the woofer play at a lower level without EQ and ending up at the same SPL?

Shawn

[CONVERGENCE]

I made a mistake here Linearity is important and subs should be RTA for Fr 15 to 125 HZ . Then SPL can be adjusted.

Sorry for the commotion.

[sfogg]

"Using EQ does not change the behavior of the decay at all....it just makes the tail hit the noise floor sooner"

Sorry, but it does. I've seen the before and after waterfalls. Again, the Lexicon and Meridian systems are both built specifically to target frequencies in the room with unnaturally long decay times. They do not try and get 'flat' response, in fact the designer of the Lexicon setup said just trying for 'flat' response in an EQ system like this is the wrong way to go.

"..maybe they're doing something besides just straight up EQ?"

Nope, straight up parametrics... only thing special about them is how narrow a bandwidth they can be set to and that the measurement/setup stage is looking at decay instead of simple frequency.

"Links would be appreciated, but I might be able to find them."

The best would be some of the posts by the designer of the Lexicon system. Go here:

http://forums.smr-forums.com

and enter as a guest.

Then check out these threads as a start. Look for posts by Jim Muller

http://forums.smr-forums.com:8080/read?27290,5e#27290

http://forums.smr-forums.com:8080/read?28557,5e#28557

http://forums.smr-forums.com:8080/read?27779,5e#27779

http://forums.smr-forums.com:8080/read?30128,5e#30128

http://forums.smr-forums.com:8080/read?32878,5e#32878

They are long threads so you are better off just looking for Jim's posts. He has many others in the forum if you do a search on his name they will come up. You will like them, lots of good information though be prepared to spend literally hours if you wanted to read them all. In with them is a room modes program he wrote that you would likely enjoy playing with too, have a fast computer standing by.

The start of the users guide talks about Lexicon's thinking for their room EQ.

http://www.lexicon.com/downloads/mc12eq/MC-12_EQ_rev1.pdf

But Jim gives much more details in the forum posts themselves.

I think Meridian has a 'white paper' or two of information on their system as well. But I'm far more familiar with the Lexicon setup.

Shawn

[bodcaw boy]

guys,

is room gain more associated with the "air load" the driver has to couple to?

have a blessed day,

roy delgado