Small Tapped Horn For a Car

**Patrick Bateman** · 09-18-2009

A few months ago I published some plans for a pair of tapped horns. Thought I'd start this thread to document the second of the pair.

First, a little background. A tapped horn is an invention of Tom Danley, who also invented the Unity horn, which is also in my car. The tapped horn is something like a bandpass box, but with some enhancements:

Like a bandpass box, a tapped horn has a limited bandwidth
Like a bandpass box, a tapped horn is one of three enclosure types which physically filter 2nd and 3rd harmonic distortion
Like a vented or a bandpass box, a tapped horn has much lower excursion than a sealed box
Because it has lower excursion, a tapped horn generally has less distortion , higher power handling, and higher maximum SPL than a sealed box

Long story short, a tapped horn is complex, but worth the trouble if you're a fan of low distortion.

Here's a pic of the first of the pair of tapped horns; it's the second I've built. It's the one at the bottom.

Here are the plans.

One cool thing about tapped horns is that they're tolerant of a lot of different drivers. You can literally pop the woofer out, put another one in, and if the params are in the ballpark, it's going to work pretty darn well. The sound of the sub is dominated by the dimensions of the horn, not the woofer.

Not to say you can throw any ol' woofer in here, but as long as the parameters are *fairly* similar you'll be fine.

Bandpass boxes are *very* particular about woofers, if someone wants to know why I'll explain. Suffice it to say that one of the reasons that bandpass boxes often sound like shit is that they're not forgiving of errors.

Anyways, here are the plans for the SECOND tapped horn, which uses a pair of MCM 55-2421s instead of the P-Audio SN-12MB. The second tapped horn is 9.5" tall, instead of 12.5" tall. In other words, it's 25% smaller.

But don't build the second one, because I have a better one in store. Stay tuned.

**thehatedguy** · 09-18-2009

How low in the car are these capable of playing?

I haven't decided on what to do for subs in my car, and I loved the TH I had in the house.

**fish** · 09-18-2009

What are the "ballpark" parameters for a tapped horn?

**Patrick Bateman** · 09-19-2009

Quote:

Originally Posted by thehatedguy

How low in the car are these capable of playing?

I haven't decided on what to do for subs in my car, and I loved the TH I had in the house.

The simple answer to that question is that I have built a tapped horn that plays down to 30hz in the car, using three eight inch subwoofers in a box that's less than three cubic feet.

Horns are almost the ideal enclosure. They have lower distortion than sealed or vented boxes, lower excursion, and power handling which is downright ridiculous. I have personally built a clone of Bill Fitzmaurice's Autotuba, loaded it with a woofer that's rated for 120 watts, and fed it EIGHT HUNDRED. It took it without breaking a sweat.

Here's a pic of the guts of a similar horn:

Everyone on the board has seen "woofer porn" where the cone is practically jumping out of the basket. Last night I powered up the new tapped horn, set it to a ridiculous volume, and the cone of the woofers was barely moving. Hell, the trunk lid was moving almost as much as the woofers were. (Gotta fix that!)

But here's the problem with horns, they're too big. You can reduce their size, but the frequency response goes to hell. To demonstrate how awful the response of a small horn is, here's a simulation of a one cubic foot horn using an MCM 55-2421 and powered with one watt.

Here's the response of the exact same woofer, in a tapped horn which is a similar size.

So what do these simulations show us?

The frequency response of both are atrocious
The box size for both is rather large. A cubic foot for an eight is a lot.
The boxes aren't easy to build. Take a look at that pic of the autotuba, it's way easier to build a sealed box
For all their problems, the excursion on the two boxes is very low, and the efficiency is very high. The efficiency of the tapped horn isn't as high as some people think it is, but the excursion of the tapped horn is a lot lower. Which is very important for car audio.

So all in all, a horn in a car just seems like an incredibly bad idea.

But here's the key - I believe the sims exaggerate the frequency response problems. In this thread I am going to demonstrate that horns measure better than the sims would have you believe.

In particular, tapped horns can be made to measure very flat, with a few tricks. In this thread I'll reveal what those tricks are.

Once you deal with the frequency response problem, you're left with a sub box which has astonishingly low excursion, and high efficiency. And those two things translate to higher power handling than anything else you can build, and lower distortion.

60ndown · 09-19-2009

a subwoofer imo, only needs to work from 25-60hz (maybe even 30-50 only) if the fronts are strong.

post up plans 'paddie' of a small taped horn that can generate 120db @ 30 hz in a van, and ill build it

g0a · 09-19-2009

How does a TH allow a woofer that handles 120 watts to handle 800 ?
Are you talking about the physical excursion limits and not the eletrical limits the coil can handle?

60ndown · 09-19-2009

Quote:

Originally Posted by g0a

How does a TH allow a woofer that handles 120 watts to handle 800 ?
Are you talking about the physical excursion limits and not the eletrical limits the coil can handle?

for a short time

**thehatedguy** · 09-19-2009

With dynamic music, the speaker would only see bursts of that much power for short short times.

Tapped horns are very cool. But you need the right speaker to go in them or the response goes all to hell.

The jump from 30 hertz to 20 is a pretty large one considering the additional lengths of the waves we are talking about.

I think looking at horns in 1/8th space would give you a rough idea on how they'll function in a car. But even in a car, you aren't really in 1/8th space since the car for purposes of bass is considered a constant pressure environment meaning there is no wave propagation happening at those frequencies. There "may be" some modal issues in the car though. I know I dismissed a multisub setup in the past ala Dr. Geddes's home approach, but there might be something to it in the car. I suspect it would be more important for bass 50-80 hertz range rather than subbass. And we could use this to our advantage because our ears localize bass so poorly to begin with.

I need to read more on eigenmodes in small rooms first.

But I suspect that if we had a small sub say a 10 in the front or near the front of the car and another one in the rear, we could get the impact and anchoring in the front from the one in the front and feel the rumble from the one in the rear. Get them aligned via time alignment, and I think it would be very awesome...and exponentially better than just 2 subs.

I did have this happen at IASCA Finals back in 04. Had a 10 in front and it wasn't doing the rumble very well. So we brought the 3 12s in the rear in, and I had the impact from the 10 in front, bass was anchored SOLIDLY in the front and the 12s in the back provided the rumble. Was pretty cool.

**Patrick Bateman** · 09-19-2009

Quote:

Originally Posted by g0a

How does a TH allow a woofer that handles 120 watts to handle 800 ?
Are you talking about the physical excursion limits and not the eletrical limits the coil can handle?

I could probably talk about SPL all day, so I'll do my best to keep this brief.

In the 90s the primary reason I couldn't get the SPL that I wanted was that I'd run out of amplifier power. So I learned how to build bandpass boxes to maximize efficiency, and made some good ones. My studies of bandpass boxes were what led me to learn about waveguides, since Geddes is the expert on both.

In the past ten years, amplifiers have become cheap and ubiquitous. Now the main thing that limits how much SPL I can generate is the loudspeaker, the alternator, and resonances in the car itself.

So that's one of the reasons this tapped horn exists. Basically the box with the lowest amount of excursion will win any SPL comparison.

After solving the excursion problem, we have to solve the thermal problem. Basically, how can we increase the thermal power handling of a loudspeaker?

In the tapped horn, there are two innovations which increase our power handling. First, let's look at the impedance of a woofer in a sealed box.

Of course we all know that the impedance of a woofer isn't flat. So a woofer that's "4 ohm" actually has a peak at resonance, and then falls below that. You can see that in the pic.

What that means for power handling is that the woofer will increasingly draw more current from the amplifier as the frequency goes lower. So at 60hz a sealed box is barely drawing any current, because of the high impedance, but at 30 hz, it's pulling nearly all the power it can get from the amplifier.

And of course the more current you draw from the amp, the hotter the voice coil gets. Which lowers our SPL, reduces our dynamics, and kills the woofer over time.

Here's the impedance curve of a commercial tapped horn, the TH-412. Here's what I see:

In the span of three and a half octaves, there are five impedance peaks
Because of those impedance peaks, the woofers in this sub are rarely seeing anything close to what the amplifier is producing
For instance, if our amplifier puts out ten thousand watts into a two ohm load, the only time this box will see all 10,000 watts is 29hz and at 58hz
At the impedance peaks, the woofers in the sub will only see a small fraction of that 10,000 watts... and there are FIVE peaks.
I believe that the "secret sauce" in the tapped horns that have been built in the past year is that using multiple woofers smooths out the response, and broadens those impedance peaks
Tapped horns are unique in this respect. If you put three woofers into a sealed box, the impedance curve will not look any different than if they have their own chambers. But the impedance curve of a tapped horn changes because the woofers are interacting with each other inside the horn

Anyways, you probably weren't expecting me to write a book for an answer. I hope this post demonstrates why feeding a 125 watt sub with 800 watts isn't such a bad idea... And how we can generate ridiculously high SPL with a $35 subwoofer in a 3cf box.

Stay tuned for more info over the next few weeks.

Oliver · 09-19-2009

For subwoofer frequencies , 3 of those peaks are null and void.

If you cross it from 80Hz down.

It will have to deal with some serious amperage from about 25Hz to 39Hz though.

grecon144 · 09-20-2009

Hello Patrick,
I actually just ordered two 55-2421 today for a car sub project and now have read with interest about your dual 8 TH. Thanks for sharing your project here. How do you think it might compare to a BFM dual8 TAT or the JBell dual T18?
Thanks

Brian Steele · 09-21-2009

Interesting thread so far.

Concerning THs, the two problems I've found theoretically with THs in car are

(1) apparently a simple transfer function calculation to predict the effect of cabin gain doesn't work as well as it does for simple "direct radiator" alignments.

(2) while efficiency goes up, peak SPL is still very dependent on driver Vd. Don't expect miracles.

(3) while it might a be an efficient use of a driver, it may not be an efficient use of the volume occupied by the box. 3 cu.ft. box for a TH on an 8" driver? You can fit three 12's in that same volume. It will be easier to build and go down to single digit frequencies at significant SPL levels in-car.

As for that 66Hz ~ 79 Hz "notch" - yes, I've noticed that too. Most annoying...

**thehatedguy** · 09-21-2009

What about dual reflex ported enclosures? Group delay go all to hell? They seem to be one of the best ways to eek out the most performance for a given foot print.

**Patrick Bateman** · 09-21-2009

Quote:

Originally Posted by Brian Steele

Interesting thread so far.

Concerning THs, the two problems I've found theoretically with THs in car are

(1) apparently a simple transfer function calculation to predict the effect of cabin gain doesn't work as well as it does for simple "direct radiator" alignments.

(2) while efficiency goes up, peak SPL is still very dependent on driver Vd. Don't expect miracles.

(3) while it might a be an efficient use of a driver, it may not be an efficient use of the volume occupied by the box. 3 cu.ft. box for a TH on an 8" driver? You can fit three 12's in that same volume. It will be easier to build and go down to single digit frequencies at significant SPL levels in-car.

As for that 66Hz ~ 79 Hz "notch" - yes, I've noticed that too. Most annoying...

I really wanted to "hold off" on showing the results that I am getting with the new sub that I've built. I think this new tapped horn has fixed many of the problems with the tapped horn technology. The new design smooths the horns response, and increases it's bandwidth.

But I couldn't resist posting the results of my work, since you noted there are notches in the tapped horns response.

I would like to demonstrate that those notches do not exist "in the real world..."

This is the predicted response of the tapped horn that I built. I measured the actual finished horn, down to the quarter inch, so these results SHOULD match our results "in the real world."

Note that big fat 12db peak at 180hz, preceded by a six DB dip.

This is the measured response of the tapped horn above, measured outside. Pretty darn smooth huh?

Admittedly, I used smoothing on the graph, since a lot of sub manufacturers do that too...

Here's the same response as above, without the smoothing. IMHO, this is still excellent. It's +/- three db over nearly three octaves.

The peak and the dip in the simulation is there in the real world, but it's much more benign than the simulation predicts. The 12db peak at 180hz is a 3db peak at 200hz "in the real world."

Combine that with the cabin gain of the average car, and it will cover close to four octaves...

For the sake of comparison, here's the response of the NEW tapped horn (purple line), the LAST tapped horn (dark red line), and a front-loaded horn I built in 2007 (red line).

You can read about those subs here:

The Smallest Tapped Horn - diyAudio (dark red line)

Free Horn Sub Design (sorta.) - Page 2 - CARSOUND.COM Forum (red line)

PLEASE NOTE THE LEVELS AREN'T RELATIVE.

The last two subs were measured at the same distance, but I had to measure the new sub at a closer range.

AND

The new sub isn't even finished. The response will improve once I make a few tweaks to it.

The new sub is about 90% finished - I wanted to demonstrate how flat we can make a tapped horn.

Also, big kudos to Tom Danley. It wasn't MY idea to use multiple subs. I just noticed that he's been doing that lately, and had a hunch that it would smooth the response. (and obviously, it DOES.)

savagebee · 09-21-2009

this looks like it has been a really good experiment, but Im still thinking more drivers in a smaller enclosure may be a better bet. I know youre not done yet, so this may not be the case, but it seems like a ton of work for no gain, what about getting a graph of the same driver in a proper qtc enclosure so we could compare the two?

or maybe I missed it and its already here

thanks you for the work youve been doing and posting everything up here, I love the experimentation

Brian Steele · 09-22-2009

Quote:

Originally Posted by Patrick Bateman

I would like to demonstrate that those notches do not exist "in the real world..."

I was referring to the notch caused by the car's cabin dimensions, not the notches in the TH's raw response. An easy way to deal with those TH notches, well, it worked for my tapped-pipe, is to simply stuff part of the enclosure, then filter the rest with a LP filter. The stuffing is basically transparent at low frequencies, but increasingly attenuates at higher frequencies. The effective path length is also increased a bit - usually a good thing

.

Quote:

Originally Posted by Patrick Bateman

Also, big kudos to Tom Danley. It wasn't MY idea to use multiple subs. I just noticed that he's been doing that lately, and had a hunch that it would smooth the response. (and obviously, it DOES.)

I suspect other factors behind the multiple small drivers are to keep total Vd high without having to settle for the lower Fs of a larger driver, and to improve power handling. Using multiple drivers also provides a bit more flexibility in the TH design as far as dimensions are concerned. The down side of course is increased cost.

Brian Steele · 09-22-2009

Quote:

Originally Posted by thehatedguy

What about dual reflex ported enclosures? Group delay go all to hell? They seem to be one of the best ways to eek out the most performance for a given foot print.

They are more complex to construct than a simple sealed or vented alignment (particularly if you want to minimize the use of available space in the car by building a molded enclosure), and they usually require more internal volume to implement as well. For the novelty factor, they're right up there with THs and other "exotic" alignments, but from the perspective of wanting maximum benefits while giving up minimum space in the process, they're worse than sealed, vented and 4th order BP alignments.

Electrodynamic · 09-22-2009

So you're only gaining 3 dB by placing your tapped horn in a car vs. outside a vehicle? Shouldn't your narrow-bandwidth alignment be +10 dB over a typical broad-band car audio alignment to support your efficiency / horn alignment in a car audio environment?

**Patrick Bateman** · 09-23-2009

Quote:

Originally Posted by Brian Steele

They are more complex to construct than a simple sealed or vented alignment (particularly if you want to minimize the use of available space in the car by building a molded enclosure), and they usually require more internal volume to implement as well. For the novelty factor, they're right up there with THs and other "exotic" alignments, but from the perspective of wanting maximum benefits while giving up minimum space in the process, they're worse than sealed, vented and 4th order BP alignments.

I built a ton of dual reflex boxes back in the 90s, and could never get them to sound "correct." The measurements looked good, but the boxes sound wonky.

Tapped horns are the opposite - I've found that the frequency response measurements are a bit frightening, but the sound is remarkable.

Did you notice that Bose stopped selling their funky triple chamber bandpass boxes? And now they're selling transmission lines?

I have a feeling they came to the same conclusion that you did, which is that double and triple reflex bandpass boxes sound weird.

**Patrick Bateman** · 09-23-2009

Quote:

Originally Posted by grecon144

Hello Patrick,
I actually just ordered two 55-2421 today for a car sub project and now have read with interest about your dual 8 TH. Thanks for sharing your project here. How do you think it might compare to a BFM dual8 TAT or the JBell dual T18?
Thanks

What is a "BFM dual8 TAT"?

digital · 09-23-2009

Quote:

Originally Posted by Patrick Bateman

What is a "BFM dual8 TAT"?

Bill Fitz Maurice's Tapped AutoTuba perhaps?

grecon144 · 09-23-2009

Hello Patrick,
I was starting a car subwoofer project using two 55-2421.
I had intended on building a dual 8 autotuba referred to as the
Tall Auto Tuba(TAT,20x22x17) or pehaps the jbell dual tuba18, when I ran across your dual 8 TH design.
As your TH appears smaller, I was trying to compare other differences in these designs.
Any thoughts on this would be appreciated.
Thanks,
grecon144

**Patrick Bateman** · 09-23-2009

Quote:

Originally Posted by grecon144

Hello Patrick,
I was starting a car subwoofer project using two 55-2421.
I had intended on building a dual 8 autotuba referred to as the
Tall Auto Tuba(TAT,20x22x17) or pehaps the jbell dual tuba18, when I ran across your dual 8 TH design.
As your TH appears smaller, I was trying to compare other differences in these designs.
Any thoughts on this would be appreciated.
Thanks,
grecon144

I have a clone of an Autotuba sitting in my garage. It's not *exactly* the same. I basically found a pic of one on google images, and figured out the dimensions, and built it.

If you can wait a day or so, I'll finish of the triple-8 tapped horn, and pit them head-to head.

That will show you the efficiency of both, frequency response, etc...

The response graphs that I've posted are of an unfinished version. I haven't glued on the side of the box. When I did that measurement I weighed the box down with 200lbs of sod on the unglued panel.

Even with 200lbs of weight, it will still leak a bit. It has to be 100% airtight to go down to 40hz.

In the car we should get 30hz.

**Patrick Bateman** · 09-23-2009

I still need to post the plans for the new sub. In the meantime, if anyone has some Image Dynamics ID8s lying around, they make an awesome match for this sub. With a pair of these you lose a little bit of efficiency over the MCM, but the simulated response is flatter and they can gobble up 600 watts without breaking a sweat, because they have double the excursion.

With a high pass filter, you could probably get away with 1000 watts at 4 ohms.

I've attached a pic.

**thehatedguy** · 09-23-2009

I have a few.