Acoustic centers

[tvrgeek]

Continuing with my tweeter evaluations, I was measuring the acoustic center of the ZAPH driver with respect to my 5 test tweeters. I had always assumed tweeter centers were pretty close to the flange, and one need only really measure the midrange. Well, not so.

Seas 27TBFC/G measured 38mm offset
Vifa T25G30 at 23
Peerless HDS at 27
SB29RDC at 30
Vifa D26NC at 30 mm

Measured with a half wave pulse at 2 Khz (my target crossover), pulser box with mic at 1M centered on the axis between the two drivers, which were about 7 inches on center.

So, when building a library of drivers, the question becomes : Offset with respect to what? This is of course with no crossover, just getting basics on the raw drivers. My conclusion is that one needs to pick a "shop standard". I can then build my library with respect to the house standard. I think I will use the SB as it's very wide bandwidth means it can be a reference against any other tweeter.

Also interesting how vastly different these tweeters sound with this gated pulse as the driver. If anyone is wondering, I am liking the T25 and 27TB. Different in character. Surprised not to like the HDS. I thought that was my ringer n the batch.

[john k...]

Originally Posted by tvrgeek

Continuing with my tweeter evaluations, I was measuring the acoustic center of the ZAPH driver with respect to my 5 test tweeters. I had always assumed tweeter centers were pretty close to the flange, and one need only really measure the midrange. Well, not so.

Seas 27TBFC/G measured 38mm offset
Vifa T25G30 at 23
Peerless HDS at 27
SB29RDC at 30
Vifa D26NC at 30 mm

Measured with a half wave pulse at 2 Khz (my target crossover), pulser box with mic at 1M centered on the axis between the two drivers, which were about 7 inches on center.

So, when building a library of drivers, the question becomes : Offset with respect to what? This is of course with no crossover, just getting basics on the raw drivers. My conclusion is that one needs to pick a "shop standard". I can then build my library with respect to the house standard. I think I will use the SB as it's very wide bandwidth means it can be a reference against any other tweeter.

Also interesting how vastly different these tweeters sound with this gated pulse as the driver. If anyone is wondering, I am liking the T25 and 27TB. Different in character. Surprised not to like the HDS. I thought that was my ringer n the batch.

Since the acoustic center is really nothing more Than a references point (phase is always relative to some reference) I have recommended for years that phase be measured relative to the driver mounting flange, be it the face plate of a tweeter or the mounting rim of a mid or woofer. In that way any offset of the AC is already in the measured phase.

Thus, if you are building a flat baffle speaker, the phase of all drivers would be relative to the baffle surface. If the baffle is staggered then you only need account for the physical stagger of the baffle. Additionally, there is little point in trying to obtain the location of the AC because if you store your measurement with phase relative to the AC, you must also store the location of the AC, and then, when designing a system you must specify the AC location. That is, to get the phase relative to the AC you removed the excess delay associated with the distance the AC is behind the baffle (or mounting flange) and then you must add it back in when designing. No matter how you look at it, 0 = +1 -1 = 0. So when making a measurement I recommend removing the excess phase form mic position to driver flange. Then there is nothing to worry about. It removes the mystery of the AC from the equation.

[Loren42]

Wait. What are you saying? Different drivers have different dispersion angles, ZDP, and efficiencies. Also, you are only measuring one sample for each type, so how can you make a generalization with a sample size of one? Does the acoustic center change if you rotate the tweeter by any chance?

I am having a hard time visualizing your test setup. If you had a picture it might help.

[Loren42]

Originally Posted by john k...

Since the acoustic center is really nothing more Than a references point (phase is always relative to some reference) I have recommended for years that phase be measured relative to the driver mounting flange, be it the face plate of a tweeter or the mounting rim of a mid or woofer. In that way any offset of the AC is already in the measured phase.

Thus, if you are building a flat baffle speaker, the phase of all drivers would be relative to the baffle surface. If the baffle is staggered then you only need account for the physical stagger of the baffle. Additionally, there is little point in trying to obtain the location of the AC because if you store your measurement with phase relative to the AC, you must also store the location of the AC, and then, when designing a system you must specify the AC location. That is, to get the phase relative to the AC you removed the excess delay associated with the distance the AC is behind the baffle (or mounting flange) and then you must add it back in when designing. No matter how you look at it, 0 = +1 -1 = 0. So when making a measurement I recommend removing the excess phase form mic position to driver flange. Then there is nothing to worry about. It removes the mystery of the AC from the equation.

Ah! Is the acoustic center really the ZDP as Dickason calls it? In other words, are we talking about the voice coil's position relative to the driver flange?

[john k...]

Originally Posted by Loren42

Ah! Is the acoustic center really the ZDP as Dickason calls it? In other words, are we talking about the voice coil's position relative to the driver flange?

It depends on how AC is defined. The most common definition is the position along the driver's axis where the phase associated with the on axis response reduces to minimum phase. A true ZPD is not a single position but is frequency dependent. All drivers have frequency dependent phase. If that phase is a nonlinear function of frequency (which is always is) then this represents a time delay which varies with frequency. Thus, the "ZDP" can not be a single position but must also vary with frequency. The ZDP reduces to a single position only if the phase is linear with frequency. Now, if the ZPD is defined as the position where the delay at any frequency is the delay associated with the minimum phase response, then the ZDP and AC become the same point.

The VC position relative to the flange is just a guess for AC position. But as I said, knowledge of the AC position is completely unnecessary, and at best, locating the AC is always an approximation requiring some degree of assumption about how the driver behaves.

[tvrgeek]

Rotating the drivers does not seem to effect it any. This method accounts for both phase and offset at a single frequency. If you were to measure at several frequencies with respect to the sync pulse, plot it all out, you could get the overall phase and extract it from AC, but I don't think that matters unless one were building an all-pass variable phase filter of some sorts. This measurement is only to get alignment through the crossover region. Not true ZDP, but simple offset at crossover center. Besides, I don't have phase calibration on my mic.

My point was how much, and we are talking 11 mm between tweeters, they vary. Of course, one must prototype and them measure for phase with the target crossover etc, but this was just a comment on how wide the variance is between tweeters. If you figure out a reference, then programs like Sound Easy can use the values. ( if you can get SE to behave, I can't). After prototyping, one can add or subtract from the measured to make the model work. All this does is get you closer for the first prototype.

The shortest offset lines up about 2 mm in front of the ZAPH voice coil. I have two of each tweeter, and see no differences between samples. These are all upper mid-quality drivers so I don't expect a lot of variance.

My set up is pretty simple. I have the Zaph in a sealed box with it mounted as close as I can to the top edge. I have a fixture with 6 inch square cutouts for all of my tweeters to hold them in place on the cabinet top. This allows me to just slide the tweeter back and forth. Mic is placed a meter away. I don't see any difference at 2 or 3 meters, providing I am consistent with being in line with the center of the two drivers. Measurement is on-axis. I use the pulse box as described in Audio Amature about 15 years ago. It generates a single half wave with a sync pulse. The frequency of the half wave is adjustable, as is the interval. This allows using a delayed sweep scope to get a nice clean view. I set the frequency to the crossover point. Connect the Zaph and adjust the scope so the peak is centered on a graticule line. Then connect the tweeter, disconnect the Zaph. Move the tweeter until the pulse is exactly where the Zaph pulse was. Reconnect the woofer and play with it a little to be sure you are not fooling yourself. Alignment is obvious as the pulse becomes well shaped and maximum in amplitude. Then just measure flange to flange. No computers were involved.

[Loren42]

Originally Posted by tvrgeek

Rotating the drivers does not seem to effect it any. This method accounts for both phase and offset at a single frequency. If you were to measure at several frequencies with respect to the sync pulse, plot it all out, you could get the overall phase and extract it from AC, but I don't think that matters unless one were building an all-pass variable phase filter of some sorts. This measurement is only to get alignment through the crossover region. Not true ZDP, but simple offset at crossover center. Besides, I don't have phase calibration on my mic.

My point was how much, and we are talking 11 mm between tweeters, they vary. Of course, one must prototype and them measure for phase with the target crossover etc, but this was just a comment on how wide the variance is between tweeters. If you figure out a reference, then programs like Sound Easy can use the values. ( if you can get SE to behave, I can't). After prototyping, one can add or subtract from the measured to make the model work. All this does is get you closer for the first prototype.

The shortest offset lines up about 2 mm in front of the ZAPH voice coil. I have two of each tweeter, and see no differences between samples. These are all upper mid-quality drivers so I don't expect a lot of variance.

My set up is pretty simple. I have the Zaph in a sealed box with it mounted as close as I can to the top edge. I have a fixture with 6 inch square cutouts for all of my tweeters to hold them in place on the cabinet top. This allows me to just slide the tweeter back and forth. Mic is placed a meter away. I don't see any difference at 2 or 3 meters, providing I am consistent with being in line with the center of the two drivers. Measurement is on-axis. I use the pulse box as described in Audio Amature about 15 years ago. It generates a single half wave with a sync pulse. The frequency of the half wave is adjustable, as is the interval. This allows using a delayed sweep scope to get a nice clean view. I set the frequency to the crossover point. Connect the Zaph and adjust the scope so the peak is centered on a graticule line. Then connect the tweeter, disconnect the Zaph. Move the tweeter until the pulse is exactly where the Zaph pulse was. Reconnect the woofer and play with it a little to be sure you are not fooling yourself. Alignment is obvious as the pulse becomes well shaped and maximum in amplitude. Then just measure flange to flange. No computers were involved.

Thanks. I just did not understand what acoustic center really meant. I do now and what you are doing make sense.

[s3tup]

You can measure the AC offset between various drivers at specific frequency: Put two drivers on single baffle, connect them in paralel with reverse polarity one against other (+ of mid to - of tweeter and vice versa) and play a sine wave thru both of them, at the frequency you plan to cross. Place a microphone at listening level (say, in front of tweeter, with the midrange placed under tweeter on the baffle). Then skew the baffle until you see a cancellation of sine waves on the microphone output. At the cancellation point you'll get the mid and tweet in-phase (they will cancel out due to reverse polarity).
Then, all you need is to measure the sides of triangle (microphone, tweeter and woofer centers) and you could be able to calculate the tweeter-midrange offset (or midrange-woofer).

All you need is a microphone with a kind of volume level measurement setup.
I did it in adobe audition by recording the signal and running spectrogram view at the same time and it works.

Running this test with crossovers on both of drivers should work too, then you'll find the relative phase offset between drivers including the crossover phase shifts.