------------------
Juha Backman, "Low-frequency polar pattern control for improved
in-room response", presented at the AES 115th Convention, New York,
2003 October 10-13, preprint 5867:
"As a summary of the results
presented above for the room-speaker interaction it can be stated
that cardioid source has more immunity against changes in source
placement or room absorption in sparsely modal range. Below the
lowest mode the cardioid speaker does not have any advantage over
the monopole source, but both exhibit higher output and less
source position dependence than the dipole speaker. These results
indicate that creating a loudspeaker that has unidirectional polar
pattern in the sparsely modal region and omnidirectional below the
lowest mode represents a good compromise between low-frequency
output capability and avoiding room coloration effects."
------------------
The following polar pattern collection should explain why different
cardioids are quite immune against changes in source position.
Typically dipole has the weakest side wall reflection, but all other
early horizontal reflections are weaker with cardioids.
Supercardioid is probably more immune and flexible than basic
cardioid.
Excessive immunity for placement may also be a disadvantage. Speaker placement closer/farther to front or side wall may not change bass level enough - like with dipoles and monopoles. Electrical tone control or different listening position may be needed if overall tonal balance is not satisfactory.
Cardioid cannot fight against reflections from back wall - of
course. Single back wall reflection may cause deep dip because other
reflections are too weak to fill it. See blue, magenta & orange
curves in the following response set.
On-axis room modes will be exited if front and back walls are solid,
and speakers are close to front wall. See the red curve. Rotated
dipole may be better choice in that case if EQ or other room
correction is not possible.
Dipole woofer is not so sensitive for a single back wall reflection,
but otherwise response variations are bigger:
Placement map for previous two measurements. Mic is at 220 cm
towards arrow. Time window 500 ms, smoothing 1/24 oct.
Total and excess (in-room) group delay of cardioid woofer (2x18") at
240 cm, on-axis:
Total and excess (in-room) group delay of dipole woofer (2x12") at
240 cm, on-axis. Dipole woofer is located 115 cm from front wall:
Response variations cause also variations in minimum group delay. GD
peaks of (bigger) cardioid are shorter compared to (smaller) dipole.
However, these examples don't have measurable excess group delay
(<250 Hz) caused by reflections from front wall and corner.
Excess group delay 50-200 Hz is based on low-pass filtering of the
woofer.
Lower attenuation of front wall/corner reflections or front-back
unsymmetry (rear transmission line) or excessive rotation of dipole
woofer could cause excess group delay when reflected sound pressure
overrides direct. For example dipole woofer rotated 60° toe-in:
Excess group delay jumps to the sky where direct sound loses the
game. This could happen with any type of source, but the risk is
reasonable small with unidirectional source which is rotated towards
the listener.
Rear-ported is the most sensitive in this group. Trade-off's are
loss of directivity at frequency band below membrane's directivity,
and distance dependence - both caused by the distance/delay between
front and rear radiators. It's applicable as a subwoofer, and polar
pattern may spread to omni already at upper bass.
Side-ported is full-range application but it's not so sensitive as
rear-ported.
This is close to traditional side-ported resistance enclosure, but may have better polar pattern when properly tuned.
Horn loaded resistance enclosure is more sensitive and kicking than other resistance boxes. Side-ported is more directive at high frequencies but not so sensitive as versions where resistance ports are located more back.
There are several ways to construct cardioid with mixed sources. Basic rule is that acoustic centers of dipole and monopole should be as close as possible. Any transmission lines on the back should be damped to avoid higher acoustical impedance -> higher sound pressure and curves in responses -> difficulties in adjustment. Vertical (and horizontal) symmetry is good but not must (see 4th example). Monopole should radiate back if it's not on both sides, to eliminate back wave as high frequencies as possible.
Main advantage of this design is adjustable polar pattern.
Everything from dipole thru hyper- and super-cardioid to cardioid is
possible by adjusting level of monopole. Even monopole is possible
if dipole is muted and monopole is also on the front (5th example).
Mixed cardioid is not full-range application because of gap between
dipole and monopole. However, bass range could be covered without
severe problems. Volume requirement of monopole enclosure is one
disadvantage. Also polar pattern is distance dependent - just like
others (e.g. dipole) having delay between front and back radiators.
Two monopoles require more volume than other cardioid constructions. Delay requirement adds complexity and may require DSP. This is mainly sub-woofer application if delayed monopole radiator is located on the back. Placement closer to front baffle allows wider band but decreases sensitivity.
Separate volumes are not must with sub-woofer application when phase difference between front and back radiator(s) is continuously close to reversed. This also decreases volume requirement radically. Filling would decrease inter-modulation at mid-range.
Advertised as cardioid at lower and mid-bass if properly damped.
This would be simple but it's not even full bass range application
if the woofer is at least 10".
Leaking (resistance ported) transmission line could improve polar
pattern.
The most effective cardioid, but also largest construction.
Dimensions are 360 x 680 x 420 mm and net volume is 2x 35 liters. Drivers are SLS-12" in parallel. Leaking gaps are 20 x 600 mm, 40 mm steps, 9 gaps per side. Leaking area is totally 200 % Sd. There's one layer of SPO Yhtymä Oy's Lido Ural synthetic felt carpet in the resistance ports. Thickness of felt carpet is about 2.5 mm. Felt is clued onto outer surface with PVAc and stapled with 11x6 mm. Inner surfaces of the box are covered with 70 mm E25 plastic foam. 2/3 of volume is filled with polyester.
Enclosure drawing: KS-702.pdf
0-180º. Measuring distance 250 cm.
Polar pattern is close to supercardioid which is flexible for
positioning.
This is out-doors measurement. Wind was not totally calm and MLS
length was only 64 kt. The result is not accurate and reliable at
the low end.
Dimensions are 330 x 580 x 430 mm and net volume is 2x 25 liters. Drivers are SLS-10" in parallel. Leaking gaps are 16 x 235 mm, 35 mm steps, 10 gaps per side. Leaking area is totally 112 % Sd. First gap is 110 mm from front edge. There's one layer of SPO Yhtymä Oy's Lido Ural synthetic felt carpet in the resistance ports. Thickness of felt carpet is about 2.5 mm. Felt is clued onto outer surface with PVAc and stapled with 11x6 mm. Box is filled with polyester.
Secondary chamber is just an additional volume for lower frequencies which are low-passed into back chamber thru 50 mm (diameter) port. Back chamber material is 19 mm MDF. It's 100% filled with polyster to damp Helmholtz-resonance. There's also one layer of air filter above the port, on drivers' side.
Enclosure drawings for 2x12" and 2x10" drivers: MT-Resistanssi-3.pdf.
Measuring distance 200 cm.
Polar pattern is excellent and close to cardioid below upper
midrange.
This is out-doors measurement. Wind wasn't totally calm, but MLS
length was 512 kt. Anyway, the result is not very accurate and
reliable at the low end.
Measuring distance 200 cm, 0-180º
On-axis, signal 2.83 V.
Signal 2.83 V.
Non-linearity of port signal is higher at bass range. Midrange is
more linear.
Drivers in parallel.
Note! This design is not recommended because of limited SPL capacity at low frequencies.
This is quick and dirty mix of dipole and monopole. Construction is composed from miscellaneous boxes and baffles lying on the garage floor. Dimensions of dipole baffle are 440 x 560 x 38 mm. Dipole drivers are SLS-10". Closed monopole box is 277 x 560 x 325 mm (33.5 liters) with XLS-10". Gap between dipole and monopole is 110 mm. Target is to locate acoustical centers of dipole and monopole as close as possible to stretch controlled directivity up to midrange. Narrow gap increases acoustic impedance and causes response curves, which are better to damp with polyester.
Near fied response of monopole is slightly rising. Required EQ is
2nd order shelving low-pass +10 dB @ 30 Hz.
Near field response of dipole is slightly falling. This dipole needs
the same shelving low-pass as monopole. In addition, there is band
stop filter -6 dB @ 300 Hz, Q=2.0, which compensates dipole peak and
transmission line effects between the baffle and the box. Monopole
and dipole (far field) levels are set equal at lower midrange. Near
field responses of one dipole and monopole driver intersect at
mid-bass:
At 5 cm on-axis and 180 degrees, without polyester damping. Tight
gap between baffle and box increases pressure from 140 to 500 Hz.
Damping is needed for sure.
0°, 90° and 180°. Measuring distance 70 cm:
The result at bass range is quite good. Also listening tells that
it's decent cardioid bass.
All measurements are done in-doors at short distance; reliability is
not very good.
This kind of mixed cardioid is easy to configure up to 150 Hz. Lower midrange is mission impossible; front baffle of monopole box is too large and dipole baffle is not fully open on the top. One problem was XLS-10" with 400 Hz maximum.
Monopole is 60 liters closed with Peerless CSX-10" or XLS-10".
Dimensions of enclosure are 520 x 320 x 520 mm (W x H x D). Box is
filled with polyester.
Dipole driver is Peerless SLS-12". Dimensions of this "open horn"
are 520 x 375 x 230 mm (W x H x D). Width of back baffle is 360
mm. Material is 24 mm plywood.
See Alien2004.
- Measuring distance 250 cm, time window 170 ms, smoothing 1/6 oct.
- drivers SLS-12" + CSX-10"
Polar pattern is slightly omnipolar at 250 cm. Level of monopole
is too high for dimensioning at 250 cm. Result would be better
closer to speaker where relative level of dipole part will
increase at lower and mid-bass. This is working system but proto
was not perfect with common power amplifier for both radiators.
Page updated 14.05.2012
Copyright © Kimmo Saunisto, All Rights Reserved.