Here is an audio test pulse for checking time delay.

60 seconds of 1.6kHz Positive Pulse Repeating 30 Times Per Second


The cutaway  drawing of an Altec 605 SUPER DUPLEX shows the location of the High Frequency diaphragm and the Low Frequency cone. Since the HF diaphragm is several inches behind the LF cone, there is an inherent time delay on the HF signal. We hear the low frequency sounds 1st and some 200 to 500 usec later, we hear the rest of the signal . 


 

This passive Time Delay Crossover is designed, in addition to being a crossover,  to compensate for the mechanical  offset of the Altec 605’s LF & HF acoustic centers . This crossover design uses a Bessel tenth order low pass filter (L1 to C5) that produces a calculated flat delay of 467 usec to the woofer signals in addition to providing the low pass crossover function. By measuring the required delay acoustically, the filter can be made to cover all electrical and mechanical delays.


SignalScope and SignalSuite software are used with a Mac for these traces.

The Test Pulse FFT power density frequency  envelope (green). The lower trace (red) is the noise level. For accurate measurements it should be 20dB below the test signal. Here the S/N Is > 40 dB.

The Test Pulse into the crossover (green) and the crossover High frequency output (red), with the speaker   connected. Using the Test Pulse as a reference ;  Note the HF out is not delayed.

Now the crossover HF and LF output signals showing the LF with a time delay of 0.4 msec. These are the electrical signals fed to the speaker.

The Test Pulse into the Time Delay  Crossover (green) and the crossover Low frequency output (red), with the speaker   connected. Using the Test Pulse as a reference ;  Note the LF out is  delayed about 0.4 msec.

HF & LF acoustical pulses  combine in time ( marker 2 ). The delay introduced by the low pass crossover filter: is matched to the delay caused by the mechanical offset of the drivers.


N1600 Time Delay Trace


This trace shows the time delay of the 605 with the  N1600 Altec factory crossover (red). The electrical  Test Pulse is shown (green) on the left ( at 0.0 ms ). The acoustical output from the woofer is at marker 1 ( 1.4 ms ) and the horn output is delayed to marker 2 ( 1.7 ms ) . The differential delay is ( 1.7-1.4 ) 0.3 msec . Also note the polarity reversal : woofer signal up, horn down.


Here, with a high pass crossover filter,  is the horn acoustic signal at marker 2. The time delay, shown  as the distance from marker 1 to 2, is 408 usec . This is the  amount of low frequency delay the low pass  crossover filter should retard the electrical signal fed to the woofer.

Here the woofer  is driven barefoot ,without a crossover, to establish a reference zero time line.

The woofer acoustic signal is at marker 1.

The measurement is 12” from the speaker on axis

and shows as 1.2 msec of air time between the electrical pulse ( green ) and the acoustical pulse ( red ).

Software  courtesy of Omicron


This is a design analysis of a low pass crossover section with exact  values per Bessel’s formulas. The curves show flat time delay at and below crossover   and a flat low pass filter frequency response ( no ripples ).

This filter design is very sensitive to changes in  load impedance. R1 and C6 are the LF Zobel that flatten the LF speaker to 13 ohms.

FREQUENCY RESPONSE CURVES (FFT)


These curves were measured simultaneously. Shown are the electrical  low pass frequency response ( in red ,measured across the woofer terminals ) and the over all acoustic frequency response of a 605A Altec speaker with the Time Delay Crossover (in green, with the controls at max).

Note how the measured electrical frequency response ( red )  matches the design analysis curve above, for the low pass filter .

Top: Reference FFT curve (in red).

Bottom: acoustic frequency response with the Mid control set to 5 (in green)

The crossover LF (green) and HF (red) electrical  frequency response. The marker is at crossover , since the HF driver is 10 dB hotter, the acoustical signals match. The 14kHz bump is controlled by L6 &  C8 to lift the high freqs. The 1kHz  suck out    protects the driver at resonance , is produced by L7 & C9. The HF control  is at max, the Mid is at 5 dB . These ( all ) curves are with 605A connected as a load.

Schematic Drawing

N1600-C Crossover

These curves are the Low ( green ) and High ( red ) electrical frequency response  for the original Altec crossover under same test conditions : the only change was to use the N1600-C crossover (0 dB Attn)

This is the acoustical  frequency response curve ( green ) of the 605A with a N1600-C with the same test conditions : the only change was to use the N1600-C crossover (0 dB Attn)

HF Diaphragm


LF Cone

My Test Pulse for checking the time delay,  between  the Woofer ( LF ) and the Horn

( HF ), is a half a Sine wave  at crossover . A Mac with Sound Studio software were used to generate a 1600 Hz ( XO freq) single sine wave , that was edited to half a cycle ; followed by 0.033 sec of silence ; and looped to repeat when played. The result is a 30 cycle per second repetition rate with good energy at crossover. Using Toast I cut a CD with this test signal. For accurate measurements the noise level should be 20dB below the test signal.

To use this technique, play the Test Pulse CD through a power amplifier, XO and speaker. With an oscilloscope and microphone the HF & LF acoustical pulses are observed as they combine in the room. Trigger the scope on the XO input signal, and place the microphone on axis to pickup the acoustical pulse. Watch the LF signal while turning the HF horn on/off. When the combined acoustic pulse vs. the LF pulse does not shift in time, but only gets taller and skinner, the time delay is right on. The microphone was 12” from the speaker  on axis.

This simple  test was discovered   quite buy accident  while I was working with Seadweller.We were able to  test down to below the 50 micro second level. Note: the pulses may be inverted.

LF Acoustic Signal in BLUE

CD Test Pulse in RED

Combined LF & HF Acoustic Signals

With Proper Delay