http://speakerdesignworks.com
Loudspeaker Designs and Articles for the DIY Enthusiast
The
AviaTrix
Measurements
and
Modeled Data
Links to:
Aviatrix Main / Design Goals
Driver Selection

Sealed MTM Cabinet Construction
MLTL MTM Cabinet Construction
MTM Crossover Design

MT Cabinet Construction
MT Crossover Design

Measurements
Frequency response plot: (Typical of all variations)

This was modeled based on the MTM on-axis measured response. The
MT is nearly identical, but of course the sensitivity is less by about 6 dB.
MTM Mass Loaded Transmission Line Impedance plot:

With the exception of the frequency band between 100 and 200 Hz, the
impedance stays above 5 ohms. I deem this a 6 ohm design, but due to
the relatively low phase angles at the impedance minima, I suspect any 8
ohm capable amplifier will be capable of carrying the load. The little
bump just above 20 Hz is the second impedance peak, showing the
design is over damped.
MTM Sealed Impedance plot:

Not much difference compared to the MLTL plot. The main resonant
peak is about 10 Hz higher, and where the lower peak was, you can just
start to see the effect of the capacitive loading caused by the 1000 uF
cap in series with the woofer network. The impedance increases
significantly below 20 Hz. I'm still calling it a 6 ohm design.  
MT Sealed Impedance plot:

This looks identical to the sealed MTM plot until you notice the
impedance scale on the left has doubled. I don't think anyone will
contradict me calling this one an 8 ohm design, although it could be
easily considered higher.  

MT MLTL Impedance plot:

Not shown for the simple reason that it will look just like the MTM MLTL
plot above if you double the impedance scale.
Driver Phase plots:

The crossover network aligns the drivers in phase quadrature. Some of
you sharp eyed individuals might have already surmised this by noting
on the response plot at the top of this page the drivers sum -3 dB
instead of the -6 dB more commonly seen.
Horizontal Off-Axis response: (TM shown, MTM similar)

These plots suggest that the best listener position will be slightly
off-axis, and the off axis voicing will be quite consistent, other than the
unavoidable tweeter roll off. -The RS28F actually does quite well in this
regard.
MTM Center Channel Off-Axis response:

A lot of folks disregard MTM center channels out of hand due to many
designs poor off-axis performance in the past, or of any MTM center
channel design crossing at 2K or higher. The advantage of the AviaTrix
is the 1K crossover point, as well as the asymmetrical crossover and
phase alignment in phase quadrature.

At 30 degrees off axis, the response is acceptably flat at +/- 1.5 dB from
500 Hz to 17 kHz. The 45 degree and 60 degree plots indicate things do
not deteriorate too badly even at these extreme angles.

For the MTM mains, this would be representative of their vertical polar
response.  
Harmonic Distortion Plots:

1/24 octave stepped sine plots of the sealed MTM.
Shot at 2.83v/1meter in room in the Aviatrix
enclosure, and the SPL was normalized to 90 dB.
The actual SPL was somewhere around 86 dB. I
caution against comparing this to anyone else's
HD plots due to the differences in measurement
techniques. There may be some room artifacts in
the plots due to the measurement distance, but I
suspect they are relatively innocuous.     
Mass Loaded Transmission Line Models:

I used Martin J. Kings's Mathcad worksheets to
model the TL variants. The driver and port
positions are very instrumental in obtaining the
smooth curve with almost no ripple shown here.  
Note the vertical scale on the modeled response
plot is only 3 dB per division, and suggest the f3 is
40 Hz, while the f6 is near 30 Hz. The modeled port
noise is down 30 dB from nominal. Also of interest
is the correlation of the modeled impedance here
and the measured impedance shown above.   
MTM MLTL Driver excursion:
-And an SPL warning

I've indicated the AviaTrix MTM is capable of a
modeled SPL of 107 dB. This is not altogether
true.
If we strictly consider the published Xmax of 4
mm The SPL will be between 103 and 104 dB.
This
assumes, in the case of the TL, there is no energy
in the frequency band below 30 Hz, or a suitable
high pass filter is applied. -But don't forget the  
published Xlim of 20 mm. -It would be a shame not
to use some of that excursion capability wouldn't it?

Simulations with an excursion of 8 mm will provide
the SPL's I suggest, and in my experience, work
well without undue distortion or driver issues using
many different types of music as source material.
However this may not be the case with certain HT
sources with significant low frequency content
, and
certain common sense limitations should be
considered when attempting to reproduce this type
of content at 'realistic' levels. After all, these are 5"
drivers, not 8's. If you require a more visceral
experience, may I suggest Jim Holtz's Statement
design as a better, albeit more costly alternative.

The MT designs SPL limitations will be 6 dB lower
than those noted for the MTM, and the same
limitations and concerns apply to them as well.
Enclosure models:

As part of the research for the AviaTrix design I
modeled several different enclosure types and
volumes: Before the MLTL design was chosen, a
classic TL enclosure was also considered, but rejected
due to poor low end extension and size considerations.

In the plot shown here, I considered options for the
MTM center channel. Even in a relatively large 36 liter
cab, a vented design exhibited nearly 2 dB of ripple in
the response. The other two plots show a 15 ltr sealed
enclosure with and without the 1000 uF in series.
While the cap ultimately changes the low end roll off to
18 dB/octave, the lack of ripple and the f3 extension of
nearly 15 Hz over the same design without the cap
suggested it would be the better choice for the
enclosure size.