In the plot below, the same test conditions as above are used but
notch filter is out. As evidenced in the plot, output between
100 and 200Hz is pronounced.
The photos below show how the notch filter can be switched out. The mounting screw holding spade terminal labelled "000" (left) is pulled from the connection and tucked under the terminal block hold-down screw (right). Note wire markers (Brady Bluestreak).
The photos below show how the notch filter can be switched out. The mounting screw holding spade terminal labelled "000" (left) is pulled from the connection and tucked under the terminal block hold-down screw (right). Note wire markers (Brady Bluestreak).
SINUSIODAL SWEEP, 1M, ON-AXIS
The plots below are the frequency response of a corner loaded factory Klipschorn (SN 18W290) with the network. The response is with the microphone position about 1m away from the front face of the top section of the cabinet. The microphone is aligned directly on the axis of the mid-range horn mouth. The input signal is a sinusoidal sweep with data collection in a reverberant environment. The mid-range horn in this measurement has been attenuated by 6dB. The bass unit notch filter is active. Three plots show same data set, shifted to show effect of data smoothing (resolution). Purple plot is raw data from analyzer, blue and red, 1/12 and 1/6 octave smoothing, respectively.
The plots below are the frequency response of a corner loaded factory Klipschorn (SN 18W290) with the network. The response is with the microphone position about 1m away from the front face of the top section of the cabinet. The microphone is aligned directly on the axis of the mid-range horn mouth. The input signal is a sinusoidal sweep with data collection in a reverberant environment. The mid-range horn in this measurement has been attenuated by 6dB. The bass unit notch filter is active. Three plots show same data set, shifted to show effect of data smoothing (resolution). Purple plot is raw data from analyzer, blue and red, 1/12 and 1/6 octave smoothing, respectively.
NEAR FIELD RESPONSES
The bass horn near-field acoustic response, located in a corner position, is shown in the plot below. The horn near-field response, measured at the left side mouth opening (purple) and right side (red) are shown. The microphone is placed at floor level, located at the geometric center of each respective mouth opening (~4.5" from the wall). The green plot is the phasor sum (derived by post processing) of both left and right responses. The two near-field responses are derived using DFFT on MLS impluse using a rectangular window. Note large dips in response at 145 and 260Hz in summed response consistent with same measured in far-field (further down).
The near-field, on-axis, response of the mid-range horn assembly (K401+K55V) and the high frequency unit (Electro-Voice T35) are shown below. The mid-range response is shown below without the filter (red) and with (blue). The T35 response is also shown both without filtering (green) and with the filter connected (purple).
.
MID-RANGE AND TWEETER CROSSOVER QUASI-ANECHOIC RESPONSE, 1M
The measurement microphone was placed 1 meter from the center of the Klipschorn top-section and an MLS impulse applied and response recorded. The mid-range and HF units were both connected to the new network. The quasi-anechoic, frequency response for frequencies >2kHz is shown below. Note the tweeter is wired 180° out-of-phase with respect to the mid-range polarity. Purple curve is all frequency domain information derived from DFFT, dark and light blue curves are 1/12 and 1/6 octave smoothing, respectively.
Excess group delay at crossover between mid-range and HF unit both with and without the filter. Microphone is located 1 meter from top section on the mid-range axis. The 4th order band-pass on the mid and 4th order hi-pass on the tweeter provides well established acoustic center at the mid-range. Blue curve is the top section with both drivers operating without a filter. Note the magnitude of the excess group delay of the midrange is approximately 1.75ms behind the tweeter down to about 1kHz. The magnitude then increases significantly below this frequency, a typical observation. Higher order attenuation rates (6th, 8th, infinite) will not reduce this value. Active crossovers are required. Excess phase derived from DFFT of MLS impulse.
The bass horn near-field acoustic response, located in a corner position, is shown in the plot below. The horn near-field response, measured at the left side mouth opening (purple) and right side (red) are shown. The microphone is placed at floor level, located at the geometric center of each respective mouth opening (~4.5" from the wall). The green plot is the phasor sum (derived by post processing) of both left and right responses. The two near-field responses are derived using DFFT on MLS impluse using a rectangular window. Note large dips in response at 145 and 260Hz in summed response consistent with same measured in far-field (further down).
The near-field, on-axis, response of the mid-range horn assembly (K401+K55V) and the high frequency unit (Electro-Voice T35) are shown below. The mid-range response is shown below without the filter (red) and with (blue). The T35 response is also shown both without filtering (green) and with the filter connected (purple).
.
MID-RANGE AND TWEETER CROSSOVER QUASI-ANECHOIC RESPONSE, 1M
The measurement microphone was placed 1 meter from the center of the Klipschorn top-section and an MLS impulse applied and response recorded. The mid-range and HF units were both connected to the new network. The quasi-anechoic, frequency response for frequencies >2kHz is shown below. Note the tweeter is wired 180° out-of-phase with respect to the mid-range polarity. Purple curve is all frequency domain information derived from DFFT, dark and light blue curves are 1/12 and 1/6 octave smoothing, respectively.
Excess group delay at crossover between mid-range and HF unit both with and without the filter. Microphone is located 1 meter from top section on the mid-range axis. The 4th order band-pass on the mid and 4th order hi-pass on the tweeter provides well established acoustic center at the mid-range. Blue curve is the top section with both drivers operating without a filter. Note the magnitude of the excess group delay of the midrange is approximately 1.75ms behind the tweeter down to about 1kHz. The magnitude then increases significantly below this frequency, a typical observation. Higher order attenuation rates (6th, 8th, infinite) will not reduce this value. Active crossovers are required. Excess phase derived from DFFT of MLS impulse.
QUASI-ANECHOIC RESPONSE, 1M
The quasi-anechoic response of the Klipschorn with the network is shown below. The red curve is the near-field, phasor sum of the left and right bass horn mouth responses combined with the quasi-anechoic response of the top section at 1 meter (the green curve, which agrees with near-field response, approximates this response). The quasi-anechoic response of the top section approximates the response to about 226Hz. The blue curve represents the phasor summation of the two responses. The near-field bass horn response was corrected to account for both measurement distance and the ratio of mouth areas between the mid-range horn (102 sq. in.) and bass horn (693 sq. in). A -6dB L-pad is used on the midrange. The bass horn notch filter is active. No data smoothing, raw data from analyzer is shown.
The quasi-anechoic response of the Klipschorn with the network is shown below. The red curve is the near-field, phasor sum of the left and right bass horn mouth responses combined with the quasi-anechoic response of the top section at 1 meter (the green curve, which agrees with near-field response, approximates this response). The quasi-anechoic response of the top section approximates the response to about 226Hz. The blue curve represents the phasor summation of the two responses. The near-field bass horn response was corrected to account for both measurement distance and the ratio of mouth areas between the mid-range horn (102 sq. in.) and bass horn (693 sq. in). A -6dB L-pad is used on the midrange. The bass horn notch filter is active. No data smoothing, raw data from analyzer is shown.
PURCHASE
The network can be purchased with either the switchable LF notch filter (KS2) or without (KS1).
KS1 $1050/pr.
KS2 $1450/pr.
Shipping networks shipped from US postal zip code 01864. For estimation purposes, KV1 (pair) shipping wt. approx. 25lbs, KV2 (pair) 32lbs. For US customers we use FedEx Ground. Full refund within 30 days of receipt, customer pays shipping back to North Reading Engineering. Units must be returned unaltered. Please note that designs and components are subject to on-going product improvements, revisions. Actual devices shipped may differ somewhat from photos shown.
The network can be purchased with either the switchable LF notch filter (KS2) or without (KS1).
KS1 $1050/pr.
KS2 $1450/pr.
Shipping networks shipped from US postal zip code 01864. For estimation purposes, KV1 (pair) shipping wt. approx. 25lbs, KV2 (pair) 32lbs. For US customers we use FedEx Ground. Full refund within 30 days of receipt, customer pays shipping back to North Reading Engineering. Units must be returned unaltered. Please note that designs and components are subject to on-going product improvements, revisions. Actual devices shipped may differ somewhat from photos shown.
THE ALTERNATIVE KLIPSCHORN NETWORK (KS1 and KS2 FILTERS)
A three-way crossover network for the Klipschorn loudspeaker system is described here. The network has been developed by iteration between computer model simulations using both LTSPICE and MATLAB simulators and experimentally determined acoustic responses. The design approach considers the acoustic response limitations of each component and seeks a network configuration that establishes as near a flat, on-axis, near-field, frequency response as is possible within the design practice allowances for useable bandwidth, phase distortion, physical size, part count and cost. Network simulations are terminated with loads that accurately capture both impedance magnitude and phase information of the bass, mid and high frequency horns providing accurate predictions of purpose specific design revisions. Near and far-field impulse, frequency and phase data files and curves are generated using CLIO (v. 7.13) acoustic analyzer.
Note regarding acoustic response curves - Real, unsmoothed acoustic response curves are not pretty (although we like them). We show them all. The only thing that matters is the acoustic response, not the predicted response from simulation (we have those too).
OVERALL DESCRIPTION
The network consists of two sections, a bass horn filter (below upper left) and a top section filter. The bass horn filter section sends low frequency signals to the bass horn whilst the top section filter sends signals to the midrange and high frequency drivers. Capacitors and inductors are Janzen manufactured (Capacitors: polypropylene Cross-Caps, Inductors: #18 air-core and #15 iron-core). We have incorporated one design feature found on the current factory network, a large notch filter in the bass horn section. We have however, altered the shape of the response to suit our design requirements.
HIGHLIGHTS
Midrange horn output can be user modified by switching out optional L-pads.
Midrange filter is 4th order band-pass, establishes acoustic center of mid-range.
High frequency filter is 4th order high-pass.
Bass horn filter is low-pass with notch filter (user modified, notch in, notch out).
A three-way crossover network for the Klipschorn loudspeaker system is described here. The network has been developed by iteration between computer model simulations using both LTSPICE and MATLAB simulators and experimentally determined acoustic responses. The design approach considers the acoustic response limitations of each component and seeks a network configuration that establishes as near a flat, on-axis, near-field, frequency response as is possible within the design practice allowances for useable bandwidth, phase distortion, physical size, part count and cost. Network simulations are terminated with loads that accurately capture both impedance magnitude and phase information of the bass, mid and high frequency horns providing accurate predictions of purpose specific design revisions. Near and far-field impulse, frequency and phase data files and curves are generated using CLIO (v. 7.13) acoustic analyzer.
Note regarding acoustic response curves - Real, unsmoothed acoustic response curves are not pretty (although we like them). We show them all. The only thing that matters is the acoustic response, not the predicted response from simulation (we have those too).
OVERALL DESCRIPTION
The network consists of two sections, a bass horn filter (below upper left) and a top section filter. The bass horn filter section sends low frequency signals to the bass horn whilst the top section filter sends signals to the midrange and high frequency drivers. Capacitors and inductors are Janzen manufactured (Capacitors: polypropylene Cross-Caps, Inductors: #18 air-core and #15 iron-core). We have incorporated one design feature found on the current factory network, a large notch filter in the bass horn section. We have however, altered the shape of the response to suit our design requirements.
HIGHLIGHTS
Midrange horn output can be user modified by switching out optional L-pads.
Midrange filter is 4th order band-pass, establishes acoustic center of mid-range.
High frequency filter is 4th order high-pass.
Bass horn filter is low-pass with notch filter (user modified, notch in, notch out).