Product Review -
Sunfire Subwoofer - January, 1997
By John E. Johnson, Jr.
Click to see larger photo |
Sunfire Subwoofer (now in the MK-II version - 1998); Passive
radiator; Two 9" drivers; 14 pound magnet; 3.2"
diameter voice coil; 2,700 watt built-in amplifier;
Current-sensing servo-feedback circuit; DC resistance of active
driver 3.3 Ohms; Frequency response 18 Hz - 120 Hz + or - 3 dB;
Volume control; Variable low-pass 40 Hz - 120 Hz; Variable phase
0 degrees - 180 degrees; Flat vs. Video Contour switch (rolls off
below 35 Hz); RCA inputs; RCA outputs (high-pass); Speaker level
inputs/outputs; Size 11"H x 11"W x 11"D; Weight 43
pounds; Black lacquer; $1,250; Sunfire Corporation; P.O. Box
1589; Snohomish, Washington 98291; Phone 206-335-4748; Fax
206-335-4746.
The Sunfire Subwoofer is one of the most highly awaited products
I have ever seen. There has been a great deal of confusion as to
what this sub is and isn't, and, unfortunately, in the rush to be
first with the review, some of the reviews that have been
published elsewhere have not resolved the problem. This subwoofer
breaks many of the rules in speaker design, and it would be easy
to just evaluate its performance from a "Black Box"
approach (the black box is an electronic device whose inner
workings are not totally understood, but is just taken for what
it is and left at that). However, this would be a disservice to
our readers, who are just as curious as we are about how this
subwoofer does what it does. Therefore, we will go into great
detail about its functioning. We have a lot of dealers who read
Secrets, so let me first say what the sub isn't. If you have a
favorite dealer who does not access the Internet on a regular
basis, make a copy of this review and give it to them.
The Sunfire Subwoofer is not bipolar, and it is not push-pull (at
least, not in the classic sense), but it DOES have
servo-feedback. With a bipolar sub, both drivers are active
(connected to the amplifier), and both drivers move into the
enclosure at the same time, and out of the enclosure at the same
time [click here
to see bipolar speaker animation].
In a push-pull sub, again both drivers are connected to the
amplifier, and both move into the enclosure at the same time, and
out of the enclosure at the same time. However, one is usually
mounted outside the enclosure, facing inward [click here to see push-pull
speaker animation]. In such a
design, even-ordered harmonics are diminished. With
servo-feedback, the motion of the driver is detected by one means
or another, the resulting signal fed back to the amplifier,
compared to the original music signal, the difference
electrically inverted, and passed through the amplifier to the
driver, reducing aberrant driver motion.
The Sunfire Subwoofer is a passive radiator design (or, more
accurately, a "mass driven radiator"), with
current-sensing servo-feedback (see Audio Engineering Handbook,
K. Benson, editor, McGraw-Hill). There are two drivers; in this
case, both are 9" diameter (measurement to the outer edge of
the rubber surround). One is connected to the amplifier
("active"), and the other is not ("passive").
They are mounted on opposite sides of the enclosure (see photo).
At the tuning frequency, which our measurements suggest is about
20 Hz, the active driver is moving very little, and the passive
driver is flailing away. From 24 Hz to 38 Hz, both drivers are
moving in similar, but not identical amounts of excursion, and in
phase. This is bipolar in a way, but with a classic bipolar
speaker, since both drivers are connected to the amplifier and in
phase, the drivers always move equal amounts and are always in
phase. Above 40 Hz, the passive radiator of the Sunfire moves
very little. Below 20 Hz, the passive radiator is the major
source of driver motion. There is a current sensing circuit that
acts as the servo-feedback. This does not work as well as a solid
state motion sensor mounted on the voice coil, but it is
certainly better than no servo-feedback at all.
The enclosure is very small (11" cube) with two 9"
drivers, but is designed to perform like a much larger subwoofer.
In order to do this, two basic things had to be present: large
excursion (Xmax) capability, and a very powerful amplifier. The
drivers in the Sunfire have an excursion of a little more than
one inch each, according to our measurements. That's an Xmax of
30mm peak to peak (that is, the driver cone can move a maximum of
15 mm forward and 15mm backward from its resting position). If
you look at the specifications of drivers in a catalog (for do it
yourself projects), you will find that 8" and 10"
drivers usually have an Xmax of perhaps 5mm peak to peak. Even
the 15" drivers will be listed as having an Xmax of about
10mm. So how do the Sunfire drivers obtain 30mm? Well that is
what's called "building from scratch." The active
driver has a heavy gauge 3.3" diameter voice coil, a 14
pound magnet, and a rubber surround that is big enough to use as
a spare tire in my Mercury Mountaineer Utility Vehicle.
The second condition mentioned above is a powerful amplifier. The
Sunfire Subwoofer amp is rated at 2,700 watts into 3.3 Ohms,
meaning that it will deliver 2,700 watts into a 3.3 Ohm resistor.
However, because of the back EMF that is generated by the Sunfire
driver as a result of a very large magnet, the system becomes
extremely efficient, and not nearly 2,700 watts are required by
the speaker itself (see below). The output section and part of
the power supply (the top of the "T" in the photo) are
the size of a very big candy bar, but the amplifier as a whole is
bigger than that [click
here to see photo]. The
preamplifier section (the vertical section of the "T"
in the photo), plus the low voltage transformer (plus or minus 15
volts for the ICs) and capacitors have to be considered when
mentioning the physical size of the amplifier. I think Bob Carver
(the designer of this and other innovative products) was simply
trying to point out that the Sunfire Subwoofer amplifier output
section has massive power capability for its size. There are four
output transistors: two Toshiba 2SA1302 and two Toshiba 2SC3281.
The power supply uses six HEXFETs and two Aero M LP85C capacitors
rated at 2,200 uF/200 volts each. Capacitance has to take into
acount not only the microfarads, but the voltage as well, which
translate into total energy storage, called Joules. The formula
is 1/2 CV2, so, for the Sunfire, this is 1/2 x .0044
x1692, or 62.8 Joules (this formula uses the DC
voltage rather than the rms value). To put this in perspective,
the capacitors are replenished at the rate of 60 times per second
x 62.8 = 3,768 watts of power supply availability. The Joule
approach to evaluating power supplies is much more precise than
simply stating the amount of microfarads.
The power supply in the Sunfire Subwoofer uses a Tracking Down
Converter [click here], and this
is a variation on the digital switching concept. That is, the
power supply switches on and off for varying lengths of time in
order to supply the output section with the required voltage. In
this case, the down converter delivers the signal (music) voltage
plus 6 volts over that amount. The amp as a whole could be
designated as Class D because of the power supply, although that
is somewhat of an oversimplification (Bob calls it a Class D
power supply with an A/B amplifier). Some of Bob's earlier
amplifiers were operated in Class H. This is because the rail
voltage was variable (more than one voltage rail). In a typical
Class A or A/B amplifier, the rail voltage is fixed. Class A is
usually very "fast", meaning that the signal voltage
and current can be delivered to the speakers very quickly in
response to changing musical waveforms. This is due to high
current constantly flowing in the output circuit, even at idle
(no music), and it is instantaneously diverted to the load (the
speakers) when music signals are introduced. In a Class A/B amp
(the majority of high quality amplifiers), a portion of the
amplification power is in Class A, but most is in Class B, where
current flow has to be turned on in response to the music, rather
than simply diverted. A 100 watt amplifier would be classified as
A/B even if 1 watt is in Class A and 99 in Class B, so you have
to be careful when comparing different A/B amps. In the Class H
amplifier, little to no current is flowing at idle, and the
current demands are turned on in response to the music signal.
Class A is very inefficient because about four times the power
output capability is being wasted as heat during idle (thus, if
it is a 100 watt Class A amp, 400 watts are being dissipated as
heat during idle). Class A/B is more efficient, and Class H is
even more so. Class D is the most efficient. Just for the sake of
throwing out some approximate numbers, Class A is about 20%
efficient, Class B 50%, Class H 80%, and Class D 90%. Efficiency
is defined as the ratio of the power supplied to the speakers
divided by the power taken from the wall AC outlet. So, the Class
D amplifier in the Sunfire Subwoofer allows it to run cool, and
maximize the input power for delivery to the speaker. It barely
got warm during use with normal music, but became somewhat hot
when we tested it with sine waves at full output. Its Tracking
Down Converter means that only 6 volts times the current flow is
being dissipated as heat (the output transistors "see"
the entire voltage, but they are limited only by the power that
has to be dissipated in heat, and the voltage drop across the
speaker load is not part of that limitation).
We could not measure the actual output of the amplifier into the
speaker load because the enclosure is sealed. However, we put a 3
amp fuse in the AC line just to the sub and turned up the volume
using sine waves (a severe test). At 16 Hz, the fuse held on even
when there was system clipping (our definition based on when the
speaker motion reaches maximum and begins to bump, as opposed to
amplifier clipping where one can hear the electronic crackling
before the speaker reaches maximum excursion). Leaving the volume
set at this point, we increased the frequency, and at 31.5 Hz,
the fuse blew. This represents 120 Volts (measured on a
voltmeter) x slightly more than 3 amps . . . let's call it 3.5.
But, in order to get the power, the rms voltage must be
calculated. To do this, you take the 120 volts, multiply it times
1.414, square the value, divide it by 5 since the power supply is
using 1/5th of the duty cycle, and then take the square root.
This represents 75.88 volts rms. So, 3.5 amps x 75.88 volts is
266 watts. Assuming 90% efficiency, this is 239 watts continuous
into the 3.3 Ohm nominal speaker. Although 2,700 watts is
theoretically possible from this Class D amplifier, we think it
is most likely to deliver between 200 and 500 watts continuous on
average (using heavy duty program material, such as nuclear
explosions), because 200 - 500 watts are all that are required.
The system puts out high SPL even with only drawing a small
portion of the potential power because of the unusually high
voltage (95 volts rms) that is supplied across the speaker
terminals with strong input signals, combined with the high back
EMF produced by the large-magnet driver, making for very good
efficiency (see below). It is this fact that makes the Sunfire
Subwoofer unique.
The AC is connected directly to the power supply capacitors (click to see
photo) rather than going through a
transformer. This means that the power can be replenished faster
than it is being used. Secondly, a power transformer sufficient
to drive this speaker would be too big for the enclosure. So why
are power transformers used in other amps? One reason is to
reduce the voltage and raise the current delivery capability.
Another is to filter out unwanted RF that is present in the AC
line but which won't pass through the transformer. A third is to
isolate the wall AC supply from direct contact with the amp
circuitry. The Sunfire has a nominal 3.3 Ohm load rather than an
unknown load that could be connected to an amplifier designed for
use with any type of speaker. Also, the amp is limited to
processing signals between 10 Hz and 120 Hz, so RF does not cause
a problem here. It would be nice to have the amp isolated from
the wall AC, but the transformer just would not fit, and
performance was the criterion for design. There is no on/off
switch, but rather, there is an automatic signal sensing circuit
that turns it on and off (a red LED on the control panel glows
brightly when the amp is on). We would prefer the
on/off/automatic sensing toggle that is present on the Sunfire
Cinema Grand, and to have the automatic sensing set at a lower
threshold. When we used the subwoofer with soft music, it
sometimes shut off with an annoying pop because there was not
enough low frequency signal to keep it on. Apparently, this
problem has been fixed in the latest production runs.
We tested the Sunfire Subwoofer in our home theater facility,
with a Yamaha CDV-W901 LD Player, Yamaha DDP-1 AC-3 Processor,
Yamaha RX-V990 Receiver, Rotel RSP-980 Processor, Sunfire Cinema
Grand Amplifier, Krix speakers, Eminent Technology VIII Planar
Magnetic Speakers, Nordost Flatline, and AudioQuest cables. Other
subwoofers used for comparison were the Velodyne F-1500R, M&K
MX-5000THX, and Mirage BPSS-210.
First, we connected the Sunfire sub to the subwoofer output on
the receiver, and removed all < 90 Hz signal from the
remainder of the sound track channels (movie laserdiscs) by use
of an AudioControl Rialto Equalizer. We set the low pass filter
on the sub to its lowest frequency (40 Hz) and started with the
introductory music and opening scene from "Jurassic
Park" (one of our favorites). The thunder produced by this
11" cube was incredible. In fact, I was so surprised by the
intensity, I spilled my sandwich on the floor. This was a really
good sandwich too. It had flaked tuna, diced pickles, celery, on
home made sourdough bread. I was ticked! After I made another
sandwich, and let the cat clean up the tuna (she doesn't like
pickles though), I continued the listening test with "True
Lies" . . . you know, the, "Here's my invitation"
scene . . . , "Return of the Jedi", and others.
The Sunfire Subwoofer did not outgun the F-1500R or MX-5000 (in
our tests), but the fact is, it came close! Sunfire recommends
putting the sub in a corner for maximum performance, but we
tested it in our standard configuration (several feet away from
all walls) for frequency (output) response comparison purposes.
90 - 100 dB is more than enough for me (see frequency response
tests below). When placed in a corner, as per Sunfire's
recommendations, the unit put out extraordinary SPL. Also, our
testing procedure is very conservative since we set subwoofers to
produce maximum volume without clipping (so that we can measure
accurate outputs throughout the usable range) at a frequency
which is quite stressful (in this case, 16 Hz). So, at the more
typical frequencies that are likely to be encountered with music
or movie sound tracks, the Sunfire will deliver even higher SPL
than we obtained in our frequency (output) response tests. This
is truly an amazing accomplishment in design, since the Sunfire
takes up only about 1/6 the volume of the larger subs. And, of
course, there is no other sub of this size that even approaches
it. The Sunfire is definitely a huge performer in a small
container.
We could hear only a faint amount of harmonics, even at 20 Hz and
full output (sine wave tests), and there was no boominess with
the low pass set to 90 Hz. This is probably due to the small
enclosure as much as the servo-feedback, and is serendipity to
say the least, since boominess can ruin the sound of human
voices, particularly female. At high output from sine wave tests,
it did jump around somewhat when placed on a hardwood floor (the
passive driver has a 1.5 pound weight, compared to the 90 gm cone
of the active driver, so even though they may be moving in equal
opposite directions, the difference in weight causes the movement
of the enclosure). Putting it on a rug prevented this. For
average listening levels, it can be put on felt pads if it has to
sit on wood floors. The AC plug is not grounded (it has two
prongs instead of three) and is non-polarized, so it can be
plugged in two ways. One way results in hum and the other does
not, so you need to try the plug both ways to determine which one
is best. There is a flat/video contour switch on the back if you
want to roll off the lows below 35 Hz.
Frequency (Output) Response Test Results, 1 meter, volume set
just below system clipping at 16 Hz, low pass set to 120 Hz
(These data represent tests in a real room with furniture, not
anechoic tests or simulations, and thus, may be somewhat
different than you might experience in your own listening room of
other dimensions and contents):
10 Hz - 60.6 dB
12.5 Hz - 71.5 dB
16 Hz - 82.2 dB
20 Hz - 97.1 dB
25 Hz - 105.8 dB
31.5 Hz - 107.9 dB
40 Hz - 108.1 dB
50 Hz - 98.2 dB
63 Hz - 101.3 dB
80 Hz - 107.2 dB
100 Hz - 95.3 dB
125 Hz - 92.6 dB
160 Hz - 81.6 dB
Frequency (Output) Response Test Results, 13 feet, volume set
just below system clipping at 16 Hz, low pass set to 120 Hz
(These data represent tests in a real room with furniture, not
anechoic tests or simulations, and thus, may be somewhat
different than you might experience in your own listening room of
other dimensions and contents):
10 Hz - 61.7 dB
12.5 Hz - 70.7 dB
16 Hz - 79.8 dB
20 Hz - 99.7 dB
25 Hz - 97.9 dB
31.5 Hz - 88.2 dB
40 Hz - 103.3 dB
50 Hz - 91.7 dB
63 Hz - 97.9 dB
80 Hz - 102.7 dB
100 Hz - 91.5 dB
125 Hz - 79.1 dB
160 Hz - 69.7 dB
As you can see, usable response goes below 20 Hz. For a 9"
driver, this represents a new level of performance that can be
expected from small subwoofers. The reason it works is that
Sunfire has gone beyond the regular Thiel parameters, by using a
very heavy magnet in the active driver. Normally, this would
result in such a large back EMF, generated by the movement of the
voice coil in the magnetic field, low bass would be diminished.
However, in the Sunfire, the back EMF is overcome by high voltage
from the power supply, and large, very intense cone excursions
result, delivering high SPL at low frequency. To explain this
further, when the maximum voltage of 95 volts rms is applied
across the 3.3 Ohm voice coil, the back EMF is 68.5 volts. This
counteracts the 95 volts, leaving 26.5 volts to do the job. So,
instead of 952 divided by 3.3, which would equal 2,735
watts, it is 26.52 divided by 3.3 which equals 213
watts, very close to our measurements using the 3 amp fuse in the
AC line (see above). A typical amplifier in most subwoofers
produces about 40 volts rms across the speaker terminals (and
that is with a big amplifier), and thus, the high back EMF values
generated by a 9" driver with such a large magnet, placed
into a small enclosure, would not allow such high SPL low
frequency output as obtained with the Sunfire Subwoofer.
We don't see this sub replacing the other, larger subs. We think
this product will find a new niche. For those audiophiles who
have a small apartment, small dorm room, small listening room, or
who just don't want a sub that is so visible, yet who demand
powerful deep bass, this is going to be just the ticket. We put
it behind a couch, in between a potted tree and the left channel
Monitor Audio Studio 20SE Speaker (the couch was moved there for
a listening test on the other side of the room), and as you can
see from the photo [click
here], the Sunfire is very
non-invasive to the furniture environment (it also added the
lowest octave to our CD reference system, bringing out those low
notes in Enya's music). The Sunfire Subwoofer is a very different
kind of product which gives us all a new type of choice (small
and powerful) in excellent subwoofers. It is a product that could
be called "NEAT". And we all know that the NEAT factor
is important when we puff out our chest to show off our hi-fi kit
or home theater. People will say, "What's that?", if
they don't already own one, or, "Where is all that bass
coming from?" One of the great pleasures of buying good
equipment is having visitors ask about it. If you want to be the
first kid on the block with one of these, you had better hurry.
Sunfire is selling them at a fast clip. But in any case, we
highly recommend the Sunfire Subwoofer for an audition, either in
your neighbor's home theater, or your dealer's showroom.
John E. Johnson,
Jr.
Editor
© Copyright 1995, 1996, 1997 Secrets of Home
Theater & High Fidelity
Return to Table of Contents for this Issue.