[Paul Spencer]
Contents
This wiki includes:
- Transitioning from convention to horn loaded design: What are we trying to achieve?
- Bass horn design goals
- Mark Seaton on horns and the Lab sub
Transitioning from conventional speaker design to horn loaded design: What are we trying to achieve?
It's not hard to know when you have achieved your design goals when
simulating sealed or vented designs, but what about bass horns?
Consider the Lab Horn designed by an authority on horn design,
Tom Danley. It's possible to design a horn that size which is
flatter in frequency response, but Danley chose not to. Why?
How could it be made more flat? A smaller throat and a larger rear
chamber volume would tame the peak at the bottom end for a start.
I think there are 3 reasons why he didn't do this:
1. A smaller throat would have a compression ratio that is too high.
2. A larger rear chamber means less excursion control below fc.
This means less power handling and higher distortion. The air spring
of the sealed box in the rear chamber is more linear in its behaviour
than the driver suspension, hence limiting its volume lowers distortion.
It's easier to then get it as flat as you need, than it is to lower
distortion.
3. The Labhorn was designed to be used with 4 to 6 in one stack.
In this stack the total of the Labhorns will behave like they're one big horn, every Labhorn is just a small building block that will, once together form a horn, with a flat frequency response down to 30 Hz. All this with a sensitivity in the order of
108 -110 dB/W/m for a stack.
Not much you say? See how many 18" (or other) subwoofers you'll need to get the same sensitivity flat down to 30 Hz. Accordingly each Labhorn will handle 1000-1600 Wrms in this setup.
This technic where multiple, relatively small horn cabinets, form a single monster horn when stacked together, is used widly in many PA designs. It's the use of an ability horns share.
Bass horn design goals
Your goals may differ to mine, but here are a few suggested goals:
* limit cone excursion below fc:
rear chamber volume should be small; below the peak, cone excursion
should be flat or reduce as you go down in frequency
* response flat enough to allow eq within the intended bandwidth:
avoid sharp high Q peaks and dips - they are acceptable if you can eq them
* the driver should be horn loaded over the range you intend to use it
* low distortion and max SPL:
These 2 goals go together. One way in which low distortion is achieved
is by operating the horn well below its max SPL. This applies to all speakers,
however, with horns, they are likely to have a lot more headroom.
Mark Seaton on horns and the Lab sub
This section is based on online forum posts by Mark Seaton.
How a horn works
In the ideal, a horn provides a better coupling between the driver(s)
and the air. Looking at actual power into a speaker vs. acoustic power
out, we see that for a normal speaker, like a ported box, the efficiency
is quite low, on the order of 1%. This means that 99 of 100W is being
lost to heat. In the case of a horn, we can ideally reach about 50%
efficiency while still having ~flat response. As you can see, we just
took 49 watts of that 100 and made it sound power rather than just heating
up the driver more. Now, most of us understand that making bass is about
"moving air" or total volume displacement.
So how is it that two 12" woofers can produce more output than 4 18" drivers?
The reason is because of the mechanical "matching" of the air to the driver.
This makes more sense when we consider this ideal case of reaching ~50% efficiency.
At this efficiency, what effectively happens is we couple the motion of the air at
the throat(in front of the drivers) to that of the horn mouth. This means that 1/2"
of air movement at the throat makes this large horn mouth move the same 1/2"
(22" x 45" in the case of the LAB horn).
Compression loading:
Figure out how large a driver that is! Now we have this other tool we can use which
is called
compression loading. This means that the area of the throat is smaller
than that of the driver. In the case of the LAB horn, the throat is 1/2 the size
of the two 12" drivers, so it is ~equal to the area of a single 12" driver. This
means that the 2 12" drivers only have to move 1/4" each to produce 1/2" of movement
in the throat. Since motion in the throat is ~equal to motion at the mouth,
you quickly see that even a pair of 12" drivers produce a lot of air motion. Think
of the horn as a mechanical "lever" between the driver an the air. As with any lever,
if you are moving a lot with a little motion, it takes more force than normal, and thus
the hefty motor on the LAB 12 drivers.
Benefits of horn loading:
The loading of a horn not only makes for a more efficient system, but also makes for a
more accurate system over its range of loading. A typical sub will by nature smear the
sound with respect to time. Effectively the system delays lower frequencies more than
higher frequencies. As we approach 50% efficiency, this phenomenon changes, and more of
the frequencies eminate with at the same time. If we were taking measurements, this can
be seen most directly in a measurement of group delay. Looking at a measurement of phase
can tell us the same thing, yet is less intuitive as flat is only "flat" if the response
is near zero.
Placement: stacking vertically in a corner:
As for two stacked in a corner, the idea is to essentially couple to the height of the
room. By placing the mouth firing *into* the corner, spaced 28" from the corner, you
effectively extend the horn by a large amount. It is a somewhat difficult situation to
model, but I would expect SERIOUS output to ~28-25Hz. Given some room gain and other
effects, the roll off could be somewhat gradual, but I can't say for sure. One nice
thing is that even after the horn begins to unload it is still a sealed box and allows
for a good bit of EQ. I would bet money that two stacked in a corner could be EQ'd to
still produce some pretty serious 20Hz output. You have the benefit of the drivers barely
moving from 30Hz up, so you have some extra excursion to play with. Adire's DPL-12 might
give a little more in this application as well.
Efficiency of the Lab sub: 2 stacked >>> 107db 1W 1m ~ 35% efficient
Unfortunately the doubling of efficiency per doubling of drive units is really
an approximation that holds reasonably well to about 30-35% With 2x the number
of boxes you can certainly produce twice the acoustic power, but the efficiency
does not go up. That said, it IS possible to achieve more than 112dB reference
sensitivity by confining the radiation angle. Remember that the 112dB sensitivity
number is referenced to full, 4pi radiation. Also realize that you quickly get to
a point where the source is large relative to the frequencies being produced, so
there is directivity to consider, even for a subwoofer when you use enough!
