Between a typical bass and tweeter, the bass need to go forward rather than backward to align the coils in the vertical plane. But you seem to suggest the opposite?
//
F is 3K or 3K5 , by memory
when both flush mounted , let's say that difference between acoustical centres is 15-20mm
pushing bass driver on back of front baffle will (almost) double that
where we are now ?
however , not good as proper (and precise) backing of HT driver
when both flush mounted , let's say that difference between acoustical centres is 15-20mm
pushing bass driver on back of front baffle will (almost) double that
where we are now ?
however , not good as proper (and precise) backing of HT driver
It is actually embarrassing to get praise for a question I did not ask but no one can doubt that was a simple misunderstanding. Mr. Bright must have started mid thread.
I was going from memory on the baffle thickness. It was decades ago when I had a pair of these and remembered it being thinner than was normal, even then. Being even thinner than 13 mm does take me by surprise and I appreciate the correction.
Having owned these after the Dahlquist "revolution" I was surprised by their exaggerating even further the natural discontinuity of the driver centers when placed on a flat baffle. I figured all of this was corrected with the crossover. It never occurred to me that this was part of the correction.
When it comes to this stuff I defer to ZM. No one I am aware of has investigated what is really happening with audio devices as carefully as he. Get to learn something every day when I read his posts.
I was going from memory on the baffle thickness. It was decades ago when I had a pair of these and remembered it being thinner than was normal, even then. Being even thinner than 13 mm does take me by surprise and I appreciate the correction.
Having owned these after the Dahlquist "revolution" I was surprised by their exaggerating even further the natural discontinuity of the driver centers when placed on a flat baffle. I figured all of this was corrected with the crossover. It never occurred to me that this was part of the correction.
When it comes to this stuff I defer to ZM. No one I am aware of has investigated what is really happening with audio devices as carefully as he. Get to learn something every day when I read his posts.
The LS3/5A was one of a series of BBC and KEF designs that used 18dB/octave slope Butterworth filters. Butterworth is a constant power response, not flat frequency response. Like all odd-order slopes, it would ideally produce a 3dB peak in frequency response when time-aligned on either polarity.
However, if you adjust the time alignment by another quarter wavelength, you can lose the 3dB peak in FR. But retain the flat power response.
It's just how the maths works. It's a clever solution with flat baffle. The BBC often used recessed woofers and negative polarity to effect this. KEF mixed it up a bit with the CS1 and CS1A kits. Thing is that BW3 gives you a varity of solutions.
The fact is that BW3 sounds very good in real life, and Harbeth still employ the BBC principles down to the light ply cabinets and screw-on baffles which like a cracked bell, don't ring.
Joachim Gerhard designed a fine small BW3 speaker called the Anima. Gramaphone remarked that it fell off above axis, so flipped the polarity to keep the reviewer happy. But in general, BW3 has much better dispersion off-axis than LR2 and LR4. I like it.
I always listen to a speaker off-axis to assess tonal changes. A good BW3 will sound good 30 degrees off axis. What you are really doing with BW3 is boosting the power at the crossover point and getting 90 degree phase.
However, if you adjust the time alignment by another quarter wavelength, you can lose the 3dB peak in FR. But retain the flat power response.
It's just how the maths works. It's a clever solution with flat baffle. The BBC often used recessed woofers and negative polarity to effect this. KEF mixed it up a bit with the CS1 and CS1A kits. Thing is that BW3 gives you a varity of solutions.
The fact is that BW3 sounds very good in real life, and Harbeth still employ the BBC principles down to the light ply cabinets and screw-on baffles which like a cracked bell, don't ring.
Joachim Gerhard designed a fine small BW3 speaker called the Anima. Gramaphone remarked that it fell off above axis, so flipped the polarity to keep the reviewer happy. But in general, BW3 has much better dispersion off-axis than LR2 and LR4. I like it.
I always listen to a speaker off-axis to assess tonal changes. A good BW3 will sound good 30 degrees off axis. What you are really doing with BW3 is boosting the power at the crossover point and getting 90 degree phase.
Regarding the recess of the B110 my understanding is that, because the LS3/5a was designed as a field monitor, the B110 was mounted behind the baffle to provide an additional margin of protection from damage as it was knocked around in mobile units. Same reason the tweeter had a shield.
As for power response, all odd order Butterworth crossovers will yield flat power response with flat axial response for point source drivers. This requires 90 degree phase difference at the crossover point. However, the vertical polar response, assuming a vertical driver alignment, will exhibit a lobed pattern, dependent on driver separation, with the potential for nulls and +3dB peaks at various angles. As driver separation becomes small relative the wave length at the crossover frequency, the lobes disappear.
I presented a series of simulations for various crossover orders, odd order for Butterworth and even order LRs starting with point sources and then including directionality, and then finally considering the effects of baffle step. The study can be examined here: Power
As for power response, all odd order Butterworth crossovers will yield flat power response with flat axial response for point source drivers. This requires 90 degree phase difference at the crossover point. However, the vertical polar response, assuming a vertical driver alignment, will exhibit a lobed pattern, dependent on driver separation, with the potential for nulls and +3dB peaks at various angles. As driver separation becomes small relative the wave length at the crossover frequency, the lobes disappear.
I presented a series of simulations for various crossover orders, odd order for Butterworth and even order LRs starting with point sources and then including directionality, and then finally considering the effects of baffle step. The study can be examined here: Power
Ha. Well obviously it is not only you aging. Must master this newfangled iPad thingy.
Harbeth cabinets are thin, but they're MDF, not ply.
As for the LS3/5A and phase, yes phase is an issue to get the drivers to sum, but beyond that the system is not time-aligned at all:
- Status
- Not open for further replies.