Where's it going?

Sent from my Z30 using Tapatalk

 

There's a calculator here that I used to figure out the volumes:

Car Audio - BANDPASS SUBWOOFER Box Design Using Gain (Fourth-Order)

Here's a pic of it. Yes, it's ugly, I need to sand and paint it. Next to it is one of those 4" cubes from Cambridge Sound Works. I used that as a 'baseline' to be sure that my efficiency, output and distortion levels were within reason.

This type of alignment is similar to this Polk subwoofer from the 90s, except the Polk is isobaric, and mine isn't. Push-pull mounting really nukes second harmonic distortion, to a level that's lower than what you'd see with a Scanspeak or Dynaudio woofer.

 

I've come up with something new and thought people might be interested in seeing the results.

Careful with this, as some of the naysayers love pointing out how some of your ideas are re-packaging of already-existing concepts. I only care about new applications for old tech, so carry on

 

Oh I steal stuff shamelessly. All of my best stuff was either a copy or a variation on something built by Kef, Danley, Dunlavy or Geddes.

 

Gotta love your weird constructions

Interested to see where you going with this...

 

Looks interesting buddy !

 

Here's the predicted response of the enclosure. Using hornresp to do the sims, as usual. You're not going to get more than one to two octaves of output out of one of these.

Here's the measured response of the loudspeaker (green) and the Cambridge speaker for comparison (red)

The line at the bottom is the measurement of total harmonic distortion. You can see the bandpass is lower than the Cambridge. This is particularly noticeable in the measurement of 2nd harmonic distortion, which is reduced by push-pull mounting. (I'll post that measurement shortly.)

The woofers are Dayton ND91. Similar to the ones pictured here. (not mine.)

If you wanted to generate a lot of SPL from 200-400hz, there are a lot of 4" and 5" woofers that will do the job. Main reason I went with the Daytons is that a pair of them will move more air than most five or six inch woofers. And you can only do "push pull" with a pair of woofers. The Daytons also have a shorting cap on the voice coil, which reduces distortion further.

The CSW enclosure fares pretty well, particularly since the amp, sub and two speakers retails for under $100. Note how flat the response is.

 

Here's my favorite part of the design. I stuffed a cylinder of closed cell foam into the port. This measurement shows the response of the loudspeaker with and WITHOUT the foam 'bung' in the port.

This is another 'trick' I stole from Kef. They've been using this technology in their speakers for a few years now. Here's a description of the bungs from Kef's documentation:

A pair of bungs is supplied with the X3OOA system for you to fine-tune the bass output. When the speakers are positioned close to the wall, it is recommended that the bungs be placed into the rear port if you find the bass output to be overwhelming or boomy. The bung can be separated into outer and inner for more bass reduction, you may put the full bung bungs. The use of the outer bung only will give less into port, bass reduction. It is recommended that you try the different bung configuration to achieve the most optimal sound output for your environment.

"I need foam for my bunghole"

Here's why the bungs are important. All boxes with a port have multiple resonances. That goofy program WinISD won't show you this, but McBean's Hornresp will. All of the programs will show you the low frequency resonance, which is a Helmholtz resonance. That's the resonance we're all familiar with, and it's the one we tune with ports when we build a vented box or a bandpass box. But there are a bunch of resonances that are higher in frequency. These are standing waves. And the standing waves are actually louder than the Helmholtz resonance!

Standing waves are very powerful. The force of standing waves destroyed the Tacoma Narrows bridge. (I was born there, it's a big ol' bridge, trust me.)

The closed cell foam 'bung' reduces output by about three decibels; but it virtually eliminates the resonances caused by the standing waves. The peak at 800hz is reduced by about 50%, and the peak at 1500hz is reduced by about 100%.

 

In... just to see where this leads.

 

How does all this increase depth in the soundstage?

 

You can make the apparent source wherever you'd like.

For instance, a few years back I had midranges up on my dash. And I used the tiniest midranges possible, because I wanted to push them as far back as I possibly could, so that the soundstage would be wide and deep.

Using a bandpass enclosure for midrange is hardly conventional; in fact I've never seen anyone do it. But it appears that it's possible.

In a bandpass box, the apparent source of the sound isn't the loudspeaker box; it's the mouth of the port.

The Kef 105/3 is an excellent example of this. There are two woofers, buried about a foot into the cabinet. But the source of the sound isn't the woofers; it's the port.

This is because the wavelength are much longer than the dimensions of the port. Five hundred hertz is 27" long. So the wave can't even fit in the port. The wave forms at the mouth.

All of this allows for some seriously bizarre loudspeaker enclosures:

1) You could stick your midbasses behind the dash. As long as the exit of the port is against the firewall, and as wide as possible, the sound will emanate as if there was a speaker located where the port exit is.
2) You could get one of those 'thin' subwoofers and use it as a midbass. Put it under the seat, in the center console, wherever. The important thing is where you locate the exit.

Cars with a 'cab forward' design are good candidates for putting mids up on the dash, firing AT the windshield instead of away from it. Plenty of people have bounced tweeters off the windshield, but the process is a bit inexact. These enclosures are very specific though. They'll give you about one to two octaves of sound, and you can put it where you want it.

 

just to clarify---Depth is a measurement of the distance from the FRONT of the Stage to the rear of the stage.

Soundstage relative to listening position is how far away the soundstage appears to originate from the listening position

 

LMAO...that Beavis gif is killing me.

"2) You could get one of those 'thin' subwoofers and use it as a midbass. Put it under the seat, in the center console, wherever. The important thing is where you locate the exit."

I looked into doing this with my midbass. I was going to run them directly behind the drivers seat in the rear passenger footwell and use a port to route the output under the car and have it exit at the firewall. But...logic got the best of me and I went for the easy way out, plus I was worried about resonances from having that long of a port, but I like the idea of using bungs to take care of that. I can't wait until I retire this car as a daily driver. :evil:

 

you know, if this works, you're going to have to name it "Gertie"!

Tailing along

 

Here's a comparison of the frequency response of the sealed 3" woofer, and the bandpass box. The bandpass box has a foam 'bung' in the port to reduce the high frequency standing waves.

This measurements shows the second harmonic distortion; note that push-pull really nukes those harmonics.

To give you an idea of just how good this performance is, take a look at John Krutke's measurement of the Scanspeak 15W8530k00. The Scanspeak is $212; the Dayton is $29. Mounted push-pull, the second harmonic distortion of the Dayton is lower. Push pull is a no-brainer.


One other note - the drivers in my bandpass box are wired in series. That's why the output level is about five decibels lower than you'd expect. The 'raw' drivers are four ohm, but once you load them in a bandpass box, the impedance goes up about 50%. Wire them in series and you end up with a load that's about twelve ohms. I'm going to have two cabinets per side, wired in parallel, for a total load of about six ohms.

 

Subd

 

in to see where this ends up.....

 

Subd..waiting for more Bung

 

Not much left to do, except paint them

Here's a summary of my findings:

It's well known that you can 'relocate' sound with bandpass subwoofers. For instance, there was a commercial sub in the 90s that put the enclosure in the trunk and connected it to the cabin via a tube.

It wasn't clear to me how high in frequency you can go. Can you do a bandpass that goes to 250hz? To 500hz? To 2000hz? The answer was unknown.

It appears that the answer is YES.

The limiting factor is the standing wave. A port has multiple resonances. Generally the lowest resonance is the Helmholtz resonance. That's the resonance that we're 'tuning' when we adjust the size of the port. But there are resonances higher in frequency : the standing waves.

This simulation, and this measurement, show both resonances. The helmholtz occurs at 267hz, and the first standing wave at 986hz.

Normally, that standing wave might make this box semi-unlistenable. It's like having a distortion that's actually louder than the fundamental. Not good.

A little piece of foam - about 2" x 1.5" - is placed into the port. Similar to what Kef does in their LS50.

That little piece of foam lowers the efficiency a bit, but it knocks down the standing wave by nearly 50%. I tried using larger pieces of foam, and it IS possible to reduce the peak entirely. I settled on a 2" piece as it's a compromise between efficiency and smoothness.



And that's about all there is to it! The rest is just standard bandpass subwoofer theory, and the calculator at carstereo.com crunches the numbers nicely. (Car Audio - BANDPASS SUBWOOFER Box Design Using Gain (Fourth-Order))


Note that you can only get about 1-2 octaves out of the device. To go high in frequency, you'll need a driver with a high FS and a low QTS.

With something like a Faital 3FE20 or a Fostex I could reach 800hz, but I'd also have a higher F3.

 

Oh, a couple more things:

1) The foam bung should work on *all* types of boxes with a port. Bandapss, vented, etc. It should even work in horns; in fact I tried it using the reticulated foam that Geddes puts in his waveguides. I switched to closed cell foam over reticulated foam because the effect of the reticulated foam was too subtle. (Reticulated foam would probably work fine if you filled up the entire volume of the port. Basically you just need less when you use closed cell foam.)

2) One concern I had was distortion. That's why I did a bunch of distortion measurements, both THD and of the individual harmonics. I used fiberglass to stuff the ports on another project, and the distortion went up significantly. This did NOT occur with closed cell foam. (Check out the measurements I posted in this thread.)

So it looks like the foam has to restrict the air to a certain degree, but if you get carried away, distortion will go up. So reticulated foam doesn't increase distortion, and closed cell foam doesn't increase distortion, as long as you don't use a lot. In my experiments, fiberglass DID increase distortion, so I'd stay away from that.


One could use the foam to make a box that lives in a bit of a grey area between a vented alignment and a sealed alignment. It wouldn't be as efficient as a vented box, or as inefficient as a sealed box. It wouldn't roll off as quickly as a vented box, or as slowly as a sealed box. It would live in the middle. And, obviously, you could pull out the foam if you wanted to blast out the SPL.

 

Thanks Patrick for the detailed explanation, so theoretically if I could manage 2 6.5" (or may be even smaller) midbass speakers in BP Isobaric configuration, I could easily fit these in my kicker panels considering here in India we have RHD cars which is almost always Manual making it almost impossible to place the midbass in the the kicker panel area...until now .
However I still have one question that is..wont the BP box by itself change the phase causing it to lag the tweets and mids. Can you share your experience on how well it gelled with the other components.

Cheers,
Venki

 

You just made my day! I wanted to discuss the topic of phase, but I figured that this thread was already a bit confusing and I didn't want to make it worse.

The short answer is that the enclosure basically behaves like a driver with an electronic bandpass filter. So, the phase isn't flat, but it's not supposed to be. Electrical, electronic, and physical bandpass filters change the phase of the driver. Filtering your woofer physically, like a bandpass does, will change the phase. But so will an electronic filter.

In my situation, this is actually a feature, not a defect. For instance, a passive bandpass filter would cost me about $15. Which is quite a bit, for a $29 driver. An active filter is even worse; I'd have to spend $80 for the active crossover and about $100-$200 for two more amp channels.

But I'm not going with a bandpass enclosure because I'm cheap, I'm going with a bandpass enclosure because it lets me hide the woofer enclosure. The fact that I can run without an electronic or passive crossover is icing on the cake. Note that a bandpass box reduces distortion also, while the other crossover types do not.

Okay, that's the simple answer, here's the complex answer, hope this makes sense:

Here's the predicted frequency response of two Dayton ND91s in a 0.18 liter sealed box, versus the same two drivers in a 0.36 liter box. Note that this is a TINY amount of air; 0.36 liters of air is about the size of a golfball. To be specific, it's a cube of air that measures 2.8" x 2.8" x 2.8".

The reason that my box looks a lot bigger is that I built it to take some serious punishment, I am running close to a thousand watts to an array of drivers.

In the predicted response, we see that the bandpass box has narrower bandwidth, but a lower F3, and a little bit more output.

Here's the *predicted* phase response of the sealed box and the bandpass box. The phase of the sealed box has some rotation, but the rotation is slow and gradual. The phase of the bandpass box rotates much quicker. This is definitely a valid complaint about bandpass boxes. I personally think that one of the reasons that bandpass boxes sound 'slow' is because they're phase response can get pretty ugly.

The trick to make this work is to keep the bandwidth to a minimum. For instance, the phase of the bandpass box is actually superior to the sealed box, if you restrict it to just 200hz to 400hz. Basically, the bandpass has flat phase response in it's passband.

Here's the *measured* response of the sealed box (Cambridge) and my bandpass box. In the measured response, we see that the phase looks a lot better than predicted, for both box types. The phase of the sealed box is basically ruler flat, from 300hz and up. Sealed boxes have really good phase response. The bandpass box is NOT flat; but it's rotation is less than 90 degrees over the span of two octaves.

A big factor here is going to be the frequency response. Basically the steeper the rolloff, the more the phase is going to rotate. Mating up another driver when the phase is only rotating ninety degrees is a piece of cake; but if you add additional filters into the mix it's going to get trickier.

If you went the opposite way, and flattened out the response with an EQ, the phase would get even flatter.

Here's one more thing to ponder. I have personally been thinking about this next thing for YEARS and one day I hope to wrap my brain around it:

If you look at the phase response of the bandpass woofer, you'll notice that in the two octaves from 400hz to 1600hz, the phase is nearly flat. But the response is falling at the same time. If you're clever, you could put another driver one octave above your first driver, and the two drivers would behave as if they were one. IE, you would have two drivers, and each driver is covering about one octave, but their phase response would look like the phase response of that sealed box. You would have two bandpass boxes that measured and sounded like a sealed box, but with lower distortion and higher output than a sealed box. I'll bet Jason knows where I'm going with this idea...

 

when you can mesh this bandpass re-location with the Unity design and get 250 hz out of ID full bodies using a PVC connecting pipe and hide your bandpass boxes under the center console, that would be pretty cool...

what, 2 ND91's per side, about 25 bucks in PVC plumbing, and getting 105 db @ 1watt from 250 hz to 18K, maybe it's a pipe dream?

 

That was pretty punny....:laugh:

 

yeah i have been talking about doing something like this for a while pat.

the big question is, how does it sound in your car. not measurements, but application?

 

Forgive my newbness, but just a quick question. "Aligning" the the drivers using a time delay and aligning them physically (within reason, I am not talking one driver behind you and one in front, just drivers within say 2ft of each other or so) would amount to the same thing, right? So you could acheive the handing off of the phase with the drivers not mounted perfectly physically?

I am thinking along the lines of having the tubes from the bandpass box poking up from the dashboard, and the tweeters mounted near them and then using digital time alignment to get the phase handoff right?

Also, I love my Vandersteen 2ci, I have yet to listen to a speaker that would make me want to switch.

 

Unless I am missing something (and I may be), it would be theoretically "more perfect" to have the midrange and the tweeter very close together, and use DSP delay to line them up.

The driver overlap on speakers like Dynaudio and Vandersteen and Dunlavy is very very high. For instance, in this Stereophile measurement of the Dynaudio Evidence Temptation, the midrange is only 10dB louder than the woofer at 500hz. And at 70hz, the midrange is only 10db quieter than the woofer.

And this speaker is nearly eight feet tall!

The net effect of this is that sound from the Dynaudio just sound BIG.

The big speakers from Wilson and Focal don't do this. Because they use high order slopes, there isn't a hint of midrange coming from the woofer. At five hundred hertz, the output of the Focal woofer will be virtually nil, as much as sixty decibels down. When you hear midrange on a Focal, you're hearing the midrange driver only. On a Dynaudio or a Dunlavy, you're hearing a combination of multiple drivers, due to the shallow slopes.


Again, a lot of this is going to depend on personal taster. Personally, I was floored by the Dynaudio. It had dynamics like my Gedlee Summa, but the soundstage was just huge, like twenty feet wide and as tall as the room. But the thing that really grabbed me was that it had the articulation and intelligibility I've come to love from Synergy horns.


All of this comes at a cost : you have to sit in the sweet spot. If you stand up or crouch down, the imaging is gone. (They still sound pretty good, they just don't image as well.)

So it's basically a choice of whether you can live with a speaker that only images like a champ in one sweet spot, or a speaker that can't quite image as well, but can do it in a wide area (like the Focal and the Wilsons.)

 

I find that people who are really picky about phase coherence love Vandys and yes not too many speakers image like them. I have heard the 2ce pictured above and they had incredible imaging. My issue was with tonality. I found them a bit dark and laid back in the mid range even more so than the dyns.

 

That is the other reason I really like them. I find most speakers far to bright for my likeing, and a lot substitute brightness for detail. They seem detailed, but it is all the tweeters, the Vandies are very detailed with no brightness at all. I have been fighting with my car system to get the brightness down and detail up. Even the HAT Mirus I put in my Miata recently (second car, just a simple set up for that one) I end up with the trebble on the head unit turned down to -8 out of 10 (whatever that works out too, need to measure sometime).

 

In my last post, I said I'd try to explain how the bandpass boxes change the phase situation.

Tom Danley doesn't post to this forum, but he does post to diyaudio, so I posted the info over there, in the event that he gives us a few hints. Here's the post:

Understanding Synergy Horn Phase - diyAudio

 

So theory and measurements aside...how does it sound in the car?

 

Dunno. I have two Synergy Horns that are about 40% complete, a horn loaded sub that's about 75% complete, a bandpass midbass that's about 75% finished and one more that's about 25% complete.

The plan is to demo this at the show in Riverside in May.

 

Awesome....I plan to be there.
Your location saying Manhattan....do you mean Beach?

 

The location is a goof. I actually live in San Diego.