The following is the text of an article that appeared in the Two:1994
issue of Speaker Builder Magazine.  Enough folks have asked for this
prior to publication that I pre-printed it once to issue #400 of The
Bottom Line bass mailing list (for subscription information send e-mail
with the line "info bass-digest" help in the body to:
majordomo@magpie.com).  Please do not reprint it without my permission.
All rights to this article and the design described therein are
reserved.

If you'll send me a self-addressed double-stamped envelope (or four
International Reply Coupons if you're overseas plus an envelope), I'll
be glad to send you a copy of the article, including the figures and
photographs.

My current address:

	Len Moskowitz
        Core Sound LLC
	405 Cedar Lane, #1
	Teaneck, NJ 07666 USA

You can subscribe to Speaker Builder by contacting:

	Audio Amateur Publications
	Speaker Builder Magazine
	P.O. Box 576
	Peterborough, NH 03458  
	603-924-9464

As of April 2004, the drivers and crossover described in this article
are still available from Madisound Speaker Components
(www.madisound.com).

On February 24, 2005 we added a section at the end of this article about
parts availability.

                   A Compact Bass Guitar Speaker Bottom
		  ((c) Copyright 1993, All Rights Reserved)
                                    
                            Leonard Moskowitz
                            405 Cedar Lane, #1
                          Teaneck, NJ 07666 USA
                              201-801-0812
                                    

I've been playing bass guitar since I was 11 years old.  That makes it
more than 25 years that I've been thumping away, holding down the bottom
end of the rhythm section.  Looking back I see that the world has
changed a lot during that period, but one thing has remained constant:
bass guitar speaker bottoms are still huge, still weigh a ton, and are
still a pain in the rear to haul around!

This realization struck me as good motivation to design a speaker bottom
that was small and light weight.  But that wasn't all I wanted!

Most bass guitars have four strings, with the lowest being E1 at 41.2
Hertz.  Two recent developments are the five and six string basses, both
having a low B string ringing out at 30.9 Hertz.  My new bottom had to
reproduce that low B cleanly.

In the old days, the star bass players like Carol Kaye and James
Jamerson used dull sounding flatwound strings.  They played through amps
like the Ampeg B-15 Portaflex that had one fifteen inch speaker in a
too-small sealed box, giving a pronounced peak in the mid-bass.  This
made bass guitars sound boomy and thumpy, and they lacked any
significant frequency content above roughly 1 kiloHertz.  Since the
seventies though, innovative bass players including Larry Graham, Sting,
Geddy Lee, and Jaco Pastorius popularized techniques like slapping and
popping, and sometimes using picks to increase definition and clarity.
These pioneers favored roundwound strings that rang out brilliantly like
a Steinway piano's low registers. My new speaker bottom had to reproduce
that high end too.

Since I had to compete with guitar players playing stacks of Marshalls,
this new bottom had to be reasonably efficient and loud.  And finally,
it had to be able to handle the roughly 200 Watts of power that a
standard bass amplifier like the Gallien-Krueger 400RB puts out.  Oh,
almost forgot: it couldn't cost an arm and a leg -- just an arm.

"Well," I thought, "that should be a challenge!"  And as you'll see, it
was.

COMMERCIAL OFFERINGS

Couldn't I buy a commercially made bottom that did what I wanted?
Unfortunately, the answer was "no."

Perhaps the hottest ticket in small bass bottoms these days is SWR's
Goliath Junior, a 3.4 cubic foot box with 2 10-inch drivers and a
deep-throuated horn tweeter.  Eden's D210T and Peavey's 210TX have
almost identical configurations and rival the SWR for popularity.
Another popular manufacturer, Hartke, sells a bottom with two of their
aluminum-coned 10-inch drivers and no tweeter.  Eden also offers another
slightly more compact model, the D110T with one 10-inch driver and a
horn tweeter in a 2.6 cubic foot box.

All of these products have their F3 down in the 60 to 80 Hertz region.
They produce useful output down to the low E, but if you want flatness
you have to equalize using electronics in either the bass guitar's
on-board pre-amp or in the amplifier. If you want to reach that low B at
performance sound pressure levels, expect to exceed the drivers' linear
excursion limits and to encounter serious distortion.

Why don't these bottoms go lower?  I've deduced that it's because their
designers have chosen to trade off low frequency extension for
efficiency and power handling.  To get efficiency, the designers chose
drivers with powerful (and heavy) magnet structures. To get power
handling they specified large/long voice coils and stiff suspensions.
Both of these tend to raise a driver's Fs.  A 10-inch driver with a
sensitivity rated around 95 dB (for 2.83 Volts input measured at 1
meter) and power handling above 200 Watts will usually have an Fs well
above 40 Hertz; an Fs in the 60's is common and some well regarded
drivers are in the 90's.  When I tried to find drivers that might offer
reasonable response down into the 30's in a small box, I found that
efficiencies were down in the low 90's or even the high 80's.

What about the high frequency end of things?  Hartke bottoms don't use a
midrange or tweeter; they roll off severely above roughly 4 kiloHertz
but still reproduce the bass guitar's overtones fairly well.  SWR and
Eden use piezo horn tweeters and spec their systems up to 12 and 13
kiloHertz respectively.  In my own critical listening I've decided that
a bass bottom must reproduce frequencies up to roughly five kiloHertz;
seven kiloHertz is marginally better.  Anything less than that and sound
quality suffers.  There's almost nothing coming out of a bass above that
so there's little reason to go higher.

PICKING A WOOFER

I'd concluded that the manufacturers had the right idea in using a
two-way design.  A two-way design keeps the complexity and component
cost down.  Also, professional usage tends to break things, so the
simpler the better.

The first thing I did was decide how big a box I was willing to lug
around to rehearsals and to gigs.  That ended up being around 1.6 cubic
feet, quite a bit smaller than even Eden's D110T.  This prompted me to
attempt to use a 10-inch woofer; no 8-inch woofer would be efficient
enough at 30 Hertz and 12-inch drivers would probably need a larger
volume than the 1.6 cubic feet.  So I did a survey of likely 10-inch
drivers using Warren Merkel's Perfect Box speaker system CAD software.
(Perfect Box is shareware and available for download from Madisound's
BBS and many InterNet sites.  If you use Perfect Box, don't forget to
send some money to Warren.)  I excluded drivers with foam surrounds
based on my bad experiences with their very limited longevity.

I discovered that the 10-inch drivers specifically designed for musical
instruments rolled off at way too high a frequency.  That drove me to
broaden my search into auto sound and home audio drivers.

At first, a few auto sound drivers looked promising.  One in particular,
Stillwater Designs' Solobaric S-10, gave very good low end response in
very small boxes along with outstanding power handling, but at the
expense of efficiency (down below 87 dB), impaired high end response,
and very high cost (over $200 retail per driver).  Other auto sound
drivers were rejected for similar reasons.

Searching through the home audio offerings, I came across the Peerless
1759, a member of their premium CC line.  The 1759 has an extra thick
polypropylene cone, a rubber surround, good high frequency response up
to around 1 kiloHertz (see figure 1), acceptable power handling (up to
220 Watts maximum), a low Fs (22.4 Hertz), reasonable efficiency (91.4
dB for 2.83 Volts input at 1 meter), and acceptable cost (commonly
around $65).  Also in its favor is that Peerless has a good reputation
for consistency and quality in the audio community.  The 1759 is widely
available from a number of distributors.  I ordered mine from Madisound.

Perfect Box showed that with the 1759 in a 1.6 cubic foot fourth order
(ported) box tuned to 31 Hertz, I could expect an F3 of 35 Hertz and an
F10 of 24 Hertz (see figure 2).  At its 220 Watt maximum power level it
would put out a healthy 112 dB SPL --not ear splitting but still
seriously loud.  And it would handle close to its maximum power down to
below 30 Hertz.  Pretty good so far.

EFFICIENCY

How did Peerless manage to give us a 10-inch driver that goes
low and still is reasonably efficient?  One part of the answer is that
the 1759 has a 4 Ohm voice coil.  Based on how we measure efficiency,
the 4 Ohm voice coil is an advantage.  To understand this let's take a
little diversion into how efficiency is specified.

Ideally we'd like to be able to characterize efficiency with a single
number that could be used to compare drivers.  The test procedures that
produce the number should be standardized.

One way to standardize is to specify that the driver be stimulated with
a known power (say 1 Watt), and then to measure the sound pressure level
at a fixed distance (say 1 meter).  If we could do this easily we'd have
a convenient measure of comparison.

Unfortunately, to stimulate a driver with 1 Watt, we need to know what
the driver's impedance is.  In the real world this isn't easy: at
different frequencies the impedance varies.  As the impedance varies, if
we want to maintain the stimulus at 1 Watt, the input voltage must vary.
This gets complicated!  It's easier to supply a stimulus that's a fixed
voltage.  Based on the assumption that most drivers are nominally rated
at 8 Ohms, and that to produce 1 Watt into 8 Ohms requires 2.83 Volts,
the standard practice became to use 2.83 Volts as the stimulus
regardless of the driver's actual impedance.

As a driver's impedance drops, that 2.83 Volt input produces more than 1
Watt.  In fact, for a driver impedance of 4 Ohms, 2.83 Volts produces 2
Watts.  So, if everything else is held constant and all we do is reduce
a voice coil's impedance from 8 Ohms to 4 Ohms, we would add 3 dB to its
so-called efficiency rating.

The upshot is that when we reduce the voice coil impedance we don't
really increase the driver's efficiency at all, despite the higher
"efficiency" number.  All we do is make it possible to push more current
through it for a given input voltage.  But as long as our amplifier can
provide enough current, the result is the same: we get a higher sound
pressure level for a given input voltage.  In most cases, modern
amplifiers can source the current needed to adequately drive a 4 Ohm
driver, so it makes sense to use one when we seek higher sound pressure
levels.  And I wanted high sound pressure levels.

CHOOSING A MIDRANGE

The 1759 woofer starts trailing off at around 1 kiloHertz (see figure
1).  I needed to find a midrange that could pick up there and carry the
ball up to at least 5 to 7 kiloHertz.  It had to be comparably efficient
to the 1759.  It had to be rugged and handle at least 100 Watts.  Its
directivity pattern couldn't fall off too severely out to 30 degrees off
axis, and it couldn't cost too much.  Its resonance had to be at least
an octave below my intended 1 kiloHertz crossover point.

Now unlike the pro-audio designers, no self respecting home audio
speaker designer would use a piezo midrange (or tweeter) because of its
unfavorable frequency response and directivity.  But there are good
reasons for its choice by many professional sound designers.  Piezo
drivers have a reputation for being nearly indestructible.  They don't
require much in the way of crossovers -- a simple capacitor is often
enough.  They're also cheap.  Those considerations didn't enter into my
choice.  As a rule I think that piezos sound terrible and I wouldn't
think of using one in my new bottom.

In contrast, I've had particularly good experiences with treated fabric
domes.  A well designed treated fabric dome midrange can sound very good
indeed.  But domes are typically not very efficient and it takes special
attention to design a dome driver that can handle high power.

Dome efficiency is typically in the high 80's.  Their efficiency can be
tweaked up a bit into the low 90's by using a short horn aperture to
shape their directivity pattern.  Dynaudio did this with their well
received D-28 tweeter and D-52 midrange.  Power handling of more than
200 Watts can be achieved if adequate operating margins are allowed for
in the design.  My conclusion was that a well selected midrange dome
would live a long healthy life in my new bass bottom.

Peerless had an appropriate driver, the 1646.  The 1646 uses a 2-inch
treated fabric dome loaded in a short horn aperture resulting in a 93 dB
sensitivity rating and acceptable directivity.  Its resonance is down at
454 Hertz and it's rated to handle 200 Watts maximum.  Its frequency
response has a slightly rising character (see figure 3) and is down
roughly 3 dB at 7 kiloHertz on-axis.  At 30 degrees off-axis its
response at 5 kiloHertz is down around 4 dB, rolling off severely from
there.  I bought one from Madisound.  It sounded good.  I'd found a
winner.

CROSSOVER

Then came the problem of designing an appropriate crossover for the two
Peerless drivers.  At first I considered bi-amping using an active 24
dB/octave Linkwitz-Riley crossover, as I do with my Marchand crossovers
in my home audio system.  I discarded that idea because it would've
increased my luggage load with an additional piece of electronics.
Another alternative was to not provide a crossover at all but instead
provide two direct inputs to the two drivers.  Many popular bass
amplifiers, including the Gallien-Krueger 800RB, SWR 400, and Hartke
7000 provide built-in active crossovers and bi-amplification so that was
a workable design decision.  But then I wouldn't be able to control the
crossover points nor the steepness of the roll-off.  That could spell
reliability and sound quality problems.  In the end I decided to use a
conventional passive two-way crossover.

In order to reduce interaction between the two drivers in the audibly
critical crossover region I needed a high order crossover with its
attendant steep slopes.  To keep the midrange driver away from its
resonance region also dictated that I use steep slopes.  In the end I
chose an 800 Hertz crossover point and 24 dB/octave acoustic slopes for
both sections of the crossover.

I didn't have adequate resources to design it so I asked Larry Hitch at
Madisound to design it for me.  Larry fired up LEAP, loaded the models
for the two Peerless drivers and came up with the circuit shown in
figure 4.  The crossover consists of an RC Zobel impedance compensation
network for the woofer, a resistor to lower the output of the midrange,
and various LC networks to provide the high pass and low pass functions.
The coils are all air core types.  The capacitors are a mix of film and
non-polarized electrolytics.  The resistor is a sandblock type.  I added
an L-pad to the midrange for level control.

The whole system's projected acoustic response at crossover is shown in
figure 5.  (This plot isn't accurate for projecting low frequency cutoff
in the box I used.)  Overall efficiency is roughly 90 dB.  The project
continued to look very promising.

CONSTRUCTION AND MATERIALS

I made the box from three-quarter inch MDF using six panels and two
small trim pieces.  The dimensions of the pieces are shown in figure 6.
For lightness, I originally contemplated vacuum pressing the cabinet
panels from foam/epoxy glass as I'd done for the shell of an uncompleted
previous project, but decided that the focus of this project was the
electronics and the cabinet's form factor -- not light weight.  I
decided to accept MDF's weight penalty in exchange for ease of
fabrication.

I assembled the cabinet using slow setting (3 hour) epoxy glue, holding
the box together in the interim using three Pony #94 framing clamps.
These clamps, made from long threaded rods and filled-nylon corner
pieces, are indispensable for box builders.  They're available from
Albert Constantine and Sons in Bronx, New York, and Leichtung Workshops
in Cleveland, Ohio.  After the box was assembled I caulked the joints
inside the cabinet with Liquid Nails, a thick construction adhesive
applied with a caulking gun.

I mounted the crossover elements on a piece of 1/4" aircraft-grade
plywood.  The components were mounted to the plywood using nylon ties
and silicone caulk.  Once the crossover was wired I installed it in the
bottom of the cabinet using Liquid Nails.  The completed crossover is
shown in figure 7.

After letting the adhesives set for a few days, the drivers were mounted
using bolts and t-nuts with a bead of Mortite between the drivers and
the baffle.  I trimmed the 1646's plastic bezel slightly on two sides to
fit the baffle.

FINISHING TOUCHES

These days most bass guitar bottoms are covered with synthetic carpet;
one common brand is Ozite.  I dislike the way that stuff feels and
decided not to use it, though admittedly it does wear well and also
dampens panel vibrations.  In comparison, the old fabric-backed vinyl
material long used by amplifier manufacturers (Tolex was one brand name
that Fender used), doesn't wear well at all.  I prefer a hard finish and
chose to cover the MDF with Kydex, a long wearing, textured sheet
plastic.  Kydex is harder and more durable than the ABS laminate
sheeting commonly used on flight and road cases.  I selected black but
it's available in many other colors from many major plastics
distributors.  I glued it to the MDF with Weldbond contact cement,
available at most hardware stores.

The box was filled with a pound and a half of polyester pillow stuffing
(one pound per cubic foot) spread loosely within it.  The stuffing adds
to the apparent volume of the box and makes up for the volume lost to
the driver's basket.

The port is a flanged 2-3/4" inside diameter flanged plastic duct
available from many sources.  I got mine from Penn Fabrication; Parts
Express and MCM Electronics have a similar part.  My port ended up being
the stock 5-1/4" inches long.

The crossover's L-pad was mounted to a small black steel cup that I
mounted over a hole in the rear panel.  A knob from Radio Shack fits
over its shaft.  The 100 Watt L-pad came from Madisound and the cup is
available from Umbra (part number 7005).  I mounted the cup to the
cabinet with cap screws, washers, t-nuts, and a bead of Mortite.

Most speaker bottoms use standard Switchcraft quarter-inch phone jacks
on their connector plates.  Having had my cord pulled out of the bottom
more than once, I decided to fix that problem.  I used two of Neutrik's
quarter-inch locking phone jacks and mounted them in a steel cup I
bought from Penn Fabrication (part number D0601K).  I mounted the cup to
the cabinet with cap screws, washers, t-nuts, and a bead of Mortite.
The cup had to be filed slightly to fit the Neutrik connectors.  The
Neutriks provide a solid, secure electrical and mechanical connection.
I also used their phone plugs to make a cable with which to connect the
speaker bottom to my amplifier. All of Neutrik's connectors are
gorgeous, high quality products.  I bought the Neutrik locking jacks and
phone plugs from my local Sam Ash Music in Paramus, New Jersey.  The
jacks can be a source of air leaks; be sure to fill any unused jacks
with an unwired plug.

Handles are another of my bugs.  I've had to lug speaker bottoms up
unbelievably long flights of stairs and if I never see a cheap plastic
well handle (the kind that is sunk into a hole in the cabinet), it won't
be too soon.  The plastic simply doesn't hold up under heavy use.  The
metal ones, similar to the spring-down handles you see on flight cases,
are nice (J. H. Sessions makes a standard one) but they require a large
hole in the cabinet.  This cabinet is small enough that I wanted to
avoid compromising its structural integrity with too many large holes.
For medium sized cabinets I've found that strap handles, as used on the
older Fender amplifiers, look good and work well.  They mount with two
bolts and t-nuts and last forever.  The model I chose is made from
flexible vinyl that is reinforced internally with a steel strap.  I
chose an all black model available from Umbra (part number 0312).

Like the musicians they work for, speaker bottoms travel a lot and are
subject to a world of abuse.  If you want them to live a long productive
life, you must protect them.  That means they must have edge and corner
moldings.  There are a bewildering variety of molding choices out there.
I chose a simple three-quarter inch plastic edge molding and a
coordinated corner molding.  The edge molding slides into the corner
molding making a neat and well protected junction.  The corner moldings
come in two styles: one with three complete surfaces for the speaker
bottom's rear corners, and one with one surface cut away for the front
baffle.  I used four of each.  The moldings are available from Umbra
(part numbers 3144, 1402/1403).  The edge moldings were glued on with
contact cement while the corner molding were held in place with 3/4"
black drywall screws.

In some performance venues (a rowdy bar for example), if the musicians
could choose to work behind a wall of steel protection, they would.
Speakers drivers need protection too.  I used a perforated black steel
grill.  The grill has three-eighth inch holes and is sized to fit into
the cabinet's front baffle area.  The edges are folded over for
three-quarters of an inch on all four sides to provide stiffness.  The
grill is held down with five 1/4-28 black bolts, rubber stand-offs, and
t-nuts.  I specified the grill's dimensions to Umbra (part number
7300/7301); they cut and folded the grills to order.

When I put my speaker bottoms down I expect them to stay put.  To make
that happen I mounted four large (one-and-three-quarter inch) rubber
feet on each the bottom and the side opposite the handles, for a total
of eight feet.  The rubber feet are reinforced internally with steel
washers and mounted with bolts and t-nuts.  They too are available from
Umbra (part number 1606).

SUMMARY

This bottom cooks!  When it comes to low frequency extension and
overall sound quality, it easily outperforms every commercial bottom
I've ever played.  It solidly reproduces the low B on a five string bass
and also gives crystal clear highs.  It won't blast you deaf but it is
seriously loud; if you need even higher sound pressure levels you can
simply use more than one bottom.  It's easy to build, attractive, and
promises to be very durable.  And lastly, it's very small.

In using MDF for the cabinet I made a conscious decision to not focus
too heavily on light weight as a goal.  The completed cabinet weighs 48
pounds.  By comparison, an Eden D110T weighs 42 pounds; a D210T weighs
60 pounds as does the SWR Goliath Junior.

The cost of the cabinet is very reasonable.  The drivers and crossover
cost roughly $125.  The accessories (moldings, jacks, handles, grill,
hardware, and cups add another $50 or so.  The MDF, Kydex, adhesives,
and Neutrik jacks add another $50 or so for a total material cost of
roughly $225.  For comparison, a good price for an Eden D110T is over
$325.

My goal for this bottom was to design and build a very compact,
relatively lightweight, reasonably efficient speaker bottom that could
reproduce the full range of a five string bass without electronic
equalization.  Most important, it had to sound good.  I believe that I
succeeded on all accounts.  I hope that if you build this cabinet that
you'll enjoy it as much as I do.



                             List of Figures
                                    
Figure 1: Peerless 1759 Frequency Response Plot

Figure 2: Peerless 1759 in Ported 1.6 Cubic Foot Box (Perfect Box)

Figure 3: Peerless 1646 Frequency Response Plot

Figure 4: Crossover Schematic

Figure 5: Predicted System Frequency Response at Crossover (LEAP)

Figure 6: Cabinet Part Dimensions

Figure 7: The Crossover

Figure 8: The Completed System

February 24, 2005: Here's a letter from Olafur Gislason with some updated information about parts availability.

Len,

Just some stuff I ran into on sourcing my materials, things change over the years.

Umbra is now a Housewares company, they reorganized their case hardware division.

TCH www.tchweb.com
800-465-6281
416-299-3255 fax

Umbra part #'s are now TCH part numbers
7301 grill 503-7301900 (900 is black)
1606 feet 503-1606-900
3144 molding -507-3144-900
1402 - 1403 corners discontinued
7005 mounting cup - 514-7005-5900
0312 Strap - 500-312900

Madisound doesn't have the port anymore
www.simplyspeakers.com Pt 306 2 3/4 ID x 5.25"

Corners for molding were from www.steelsound.com
They also have a locking 1/4" jack.

Penn Fabrication got weird about selling small qty's retail, I used www.usspeaker for a D0-944 prepunched mounting cup.

Larry at Madisound did tell me they got in a few more of the 1646 drivers, they have 29 now, but that is the last they expect. He noted they had sold about 15 kits a year pretty steadily. He also said Peerless had been sold and that plant was closing with a few products being dropped.

I figured if someone should ask you, there are the updated sources, no need to re-do the search...

regards Olafur