D I O... P A P E R
FULL RANGE DRIVER
started offering our modified Fostex FE206EM
drivers (shown below) to the D.I.Y. market several years ago with
the most popular cabinet design being the HDT.
this span I don't think a week went by where we didn't listen to
these drivers - and on an ever changing variety of amplifiers.
At the same time we also swap in and out a rather large variety
of other fine speaker designs that don't employ the single driver
approach. This has given us a well earned perspective on the
strengths and weakness of our FE206EM driver. The FE206EM
was a significant step up from the stock Fostex FE206E and offered
a slightly smoother frequency response and better high frequency
response (something that the stock driver sorely lacked).
new to these types of drivers, they all share a few common denominators.
A) they are efficient, B) They are extremely fast, C) Bass
is not their passion, or put another way is something that has to
be earned with a good cabinet design. D) Midrange presence
is always above average. E) Perhaps most importantly is the
cones are paper thin.
on E for a second. Paper thin cones are super low mass, and
that is a good thing making A, B, C and D what they are. The
down side is that any sound on the back side of the cone is easily
transmitted to the front side of the cone. Quick example:
Remove driver from enclosure. Toss an AM battery operated radio
inside the cabinet. Reinstall the driver. The AM battery radio
will be crystal clear and actually be louder than when it was outside
of the cabinet in your hands. This is the main reason why
so many single driver speaker designs have certain issues that often
get labeled as midrange shout.
the original FE206E was a decent driver. Our modified version
of it, the FE206EM was a significant improvement made possible by
a number of changes. We have found other ways to improve it, and
after a year of testing are now ready to release what will likely
be the final version, the DFR-8 (Decware Full Range - 8 inch).
will define the improvements between the FE206EM and the new DFR-8
the frame and some minor cone treatments, the biggest modification
to the original Fostex driver was the phase
guide that was installed in place of the dust cap. The actual
shape of this active phase guide is something that I've been playing
with over the past year by having various designed machined and
then installing them into the prototypes that we listen
to every week. Of course we also applied these guides to other
full range drivers such as the audio nirvana 8 inch and the less
expensive Lowthers. I bring that up because the lesson that
came out of it is that every phase plug or in our case the
active phase guide must be unique for the driver it is used
above is where the phase guide design landed. This final
version has a shorter port which raises the resonant frequency.
That makes the guide more active in the top two octaves.
This new phase
guide was used in some of the final FE206EM drivers that were made,
and is standard in the new DFR-8.
takes all of the mod's of the FE206EM to a more refined level (including
cone treatment) but with a new suspension design that is also
active. If we can animate the function of a phase plug
why not some of the rest of the speaker? The new suspension
design creates a diaphragmatic absorber that sucks nearly all of
the modal resonance's out of the cone and coverts that unwanted
energy into heat.
active suspension system
FE206EM used an inverted roll cloth surround that is attached to
the under side of the speaker cone. The cloth is treated,
but not air tight making it impossible to create a truly sealed
diaphragm between the inside and outside of the speaker cabinet.
The cloth has the advantage of being low mass, strong, and quick.
It offers no damping like butyl rubber but then butyl rubber
is high mass, less strong, and slow.
If you look
at Lowther speakers they typically use polyether foam, not unlike
that found in many hi-fi speakers. I suspect the reason why
is because the polyether foam is the most responsive, or fastest
of all the available surround materials.
We have added
a specially matched positive roll polyether foam surround to the
top of the cone, while leaving the existing inverted cloth surround
in-tact. This creates a more linear resistance on the cone
in both directions. It also by nature adds some damping to
the cone, another reason Lowther probably uses it. But in
the case of this driver, having both creates a dead air space that
is sealed (but with a high Q loss from the original cloth to the
inside of the enclosure) that effectively becomes a perfect absorber
for the midrange frequencies that are most effected by modal resonance's
and reflected energy inside the cabinet. I call it a diaphragmatic
mass and FS are not notably effected by this new suspension system
but the linearity has improved significantly. Having such
a high tech push-pull suspension keeps the voice coil in the gap
a greater percentage of the time which actually increases the efficiency
in the areas where it suffered from 5K on up.
system also offers tighter bass control and enhances low frequency
power handling by a notable amount. In an open baffle situation
it makes it possible to run the driver full range below the cut
off of the baffle step without it unloading on low frequencies nearly
the suspension eliminates low frequency rumble at 20 Hz, and gives
the bass that IS there some real weight and percussion. The
positive roll also increases the actual surface area of the cone
making the speaker diameter larger than before and also seems to
give a better off axis launch from the outer diameter of the speaker
cone. (Better imaging)
of the FE206EM and DFR-8.
To try and
find a way to accurately show the differences between FE206EM
and the new DFR-8 I had to really think about it for awhile...
frequency response supplied by Fostex for the stock FE206E would
be the first place to start, and then compare it with the new DFR-8.
Problem is, these frequency response plots really tell you
nothing. Your probably better off when selecting a driver
if you don't even look at them. The weighting and scales used
can modify the response graph to look more attractive on paper
than it really is. When you combine that with the reality
that it will measure nothing like this in your cabinet or room you
have to take manufactures frequency response charts with a
grain of salt.
graphs published for a given driver do not tell you how the
speaker will sound, or what the actual frequency response will be.
There could be a narrow peak at 3kHz that simply doesn't show
up on the graph. It could be one that you would find
distracting or even annoying to the point of not liking the speaker.
In the listening room the drivers response and other characteristics
are greatly effected by what happens on either side of the speaker
cone. Interactions from the cabinet will effect the back
side while interactions with the listening space will effect the
front side of the speaker cone and thereby modify it's response.
The speaker itself reacts with the amplifier driving it which
further modifies the speakers response. Add to that the fact
that from spot to spot throughout your entire listening space, no
two spots will measure the same and it's easy to see why a
factory response chart is almost useless.
you look at the graph below you will see a logarithmic frequency
response chart done in two different scales. The 1/3 octave
is the blue line and is what you usually see from the speaker
manufactures. The red line is a more sensitive 1/12 octave
plot that begins to reveal the narrow peeks and valleys that make
up a speakers frequency response. You can see that the blue
line is easier to interpret than the red, but the red has far more
information in it.
How we measured
to answer the anticipated question; "What's the difference
between the FE206EM and the newer DFR-8" we choose to install
the drivers into a pair of ideal cabinets placed into an average
sized listening space. With computers we can now measure
the individual wave fronts coming off the cone before room reflections
modify those wave fronts. By gating the response to just a
few milliseconds it is possible to see what would be similar to
the anechoic response of the speaker while at the same time study
the decay from room reflections. This approach gives us more
usable data in regards to how something might sound when it gets
put into use in the real world.
To find the
best mic placement for testing a real time measurement was
taken showing a 3D waterfall plot that we could watch change as
we moved the mic around. The idea is to gradually move the mike
until you find the smoothest response. This ended up being around
15~20 degrees off axis on both speakers. Since mic placement
can wildly change the frequency response you'll measure, it is imperative
that both speakers are measured from the same point in space.
Also the distance
for the mic was 11 inches from the cone. The mic used is calibrated
as is the sound card in the computer doing the measurements. For
these tests the amplifier used was the SE34I.2
which is a class A single ended triode at 6 watts per channel.
measurement is a burst test that shows the impulse response of the
two drivers. It gives us a general idea of the speed at which
the driver will start and stop, however since the burst test is
not using music, it doesn't tell us how well the driver will maintain
this impulse response as the volume is increased while listening
to music. In listening tests it was easy to hear the new suspension
system deal with low frequency information in a far better way than
the FE206EM, yet it is only hinted at in the graphs below.
above is the impulse response of the FE206EM. The cone has
stopped moving within 2 milliseconds. Below is the DFR-8.
The cone has stopped moving within 1.5 ms and the initial burst
is cleaner. Both of these are killer good compared to a conventional
hi-fi woofer but between the two, the DFR-8 comes out on top.
below is the FE206EM. On a response chart like this we are
especially interested in the area between 20Hz and 4016Hz. This
is where the bulk of the music lives and the critical midrange where
our ears are most sensitive. Key points in this response are
the narrow 6dB peek at 3kHz and the large trough at 8Khz and the
roll off starting around 14kHz.
the same measurement taken for the new DFR-8. Notice that the peek
at 3kHz is non-existent. In fact all of the issues above have
been corrected. The narrow dip around 5kHz keeps the speaker from
getting in your face and while on paper may be considered an artifact
in the listening room it clearly seems desirable.
graph below is the FE206EM. Of interest is the apparent flutter
centered at 28 Hz. This also showed up on the stock FE206E
drivers. Remember the driver is suppose to roll at 40Hz.
next graph below is the same measurement taken on the DFR-8. Notice
the artifact at 28Hz is completely gone, and the roll off at resonance
is as it should be. We also see a much flatter response from resonance
(40Hz) on up.
are the same two graphs in a water fall plot showing the decays.
The first one is for the FE206EM and the second one under it
is for the DFR-8.
impressive improvement can be seen below with the new DFR-8.
range cumulative decays
here are the wide range waterfall plots of each driver. These
will match the linear frequency responses shown earlier on this
page. The first is the FE206EM followed by the DFR-8.
final point of interest is the phase response of each driver. In
this graph you are looking for the blue line to follow the red reference
line as closely as possible. The first chart is for the FE206EM,
the second is for the DFR-8.
again, the graph below shows the DFR-8 with a significant improvement
in the phase response
black dye originally used on the FE206EM is no longer used on the
DFR-8. The process leaves the natural color of the banana fiber.
We use a polyacrylic diluted with water and apply it to the
cone under a high rotation. As the cone gets wet the fibers
expand and the top layer of fibers are re-animated and brushed in
the direction of the rotation. When the moisture evaporates
from the cone, the polyacrylic encapsulated fibers are locked into
position. The reason we do this is to remove the stress locked
into the paper cone when it is pressed. The untreated cone
is glass smooth to the touch while the treated cone has lost the
gloss and developed a subtle texture. The improvement you
hear in the sound of the paper itself when rubbed also shows up
in the timbre of instruments and voices. A final benefit this
cone mod brings is a stronger cone and one that is less effected
by changing humidity.
basket has felt lined spokes behind the cone, and each spoke is
also dampened on the outside with viceoelastic foam pads. There
is also a small 1/2 inch square pad added to the back side of the
cone opposite the lead wires to offset the mass of the lead wires
and damp ringing from the neck joint.
you listen to a 1kHZ reference sine wave on the FE206EM it sounds
like any other decent driver would, but when you listen to a 1kHz
sine on the DFR-8, it actually has tone, sounding more like an instrument
than an oscillator. This was a rare observation made during
testing. In the listening room, the driver is far smoother
with noticeably better top end. The focus is better, and the
sound projects better as opposed to sounding like some of the sound
is coming out of the speaker cabinet itself. The ability of the
cabinet to disappear has greatly improved. If was to rate
each driver between 1 and 15, I would give the stock
FE206E a 7.0, our modified FE206EM a 8.3 and the DFR-8 a 11.5 in
over all sound quality just to give you an idea of the range of
anyone with the original FE206EM drivers wishing to have the new
suspension system added, you can contact me or just send the drivers
in with a letter stating what you want done. Cost of the upgrade
would run around $25 for each driver. Upgrading the phase guide
is usually not possible, however an original FE206EM with the newer
suspension will sound very close to the DFR-8, with a sound quality
of 10.5 based on the example above.
Decware is a trademark of High Fidelity Engineering
Copyright © 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
2007 2008 by Steve Deckert