linear phase 2-way kits

[noviygera]

AllenB,
thanks for that link.
Davey,
my objective was to find a 2-way monitor that places emphasis on good phase response, and some of the designs mentioned here do have good phase tracking.
Furthermore, I have learned that only a coaxial design will have consistent phase off axis (vertical). And so this leads me to believe that one possibility is to start with a good coaxial driver. Not many around. KEF and Tannoy? Others are ok but not great.

The next step would be to explore the possibility of phase compensation with the "lattice phase eq" -- active line level ot passive. Not much I can find on this implementation for 2-way monitors...

So this leads to the following conclusion:
Great coaxial driver with 1st order filters or
Great coaxial driver with 2st order filters + phase compensation.

 

[Draki]

Sorry, my mistake:
SBA-761

this is the 1st order design (open design, free) and the top section can be used as the 2-way monitor, I think.

Actually it is not a full 1st order design. While the lower bass does follow a 1st order acoustical at Fc 850 Hz aprx, and the upper unit does the same at Fc 2K5 Hz aprx, the tweeter rolls on with LR12 at 2K5 Hz aprx. The phase is tilted from -180 to 180.

Hint: the tweeter is connected in reverse.

 

[ernperkins]

AllenB,
The next step would be to explore the possibility of phase compensation with the "lattice phase eq" -- active line level or passive. Not much I can find on this implementation for 2-way monitors...

Here's an example using "lattice phase eq" (aka a ladder delay network) by John Krutke: LINK. It's the best example I've found.

The ladder delay network used in the example is an alternate method to compensate for differences in the acoustic offset between drivers (ie to match phase at the crossover point). John could have opted for a sloped or stepped baffle instead. The example implements a LR2 design which certainly is not a linear phase design.

 

[mark100]

1. Can someone tell me what to look for in the measurements to identify if a speaker can be at least close to being able to produce a good square wave? My guess is that to play a square wave, the phase has to be a flat line, with NO tilt.
Yes? Or is this an oversimplification and phase does NOT have to be a flat line?

2. So if the phase response is a TILTED continuous line, like the measurements of the "Revelation Two - M4 WG", what does this say about its ability to play a square wave?
3. What does the phase tracking tell us? Let's say I want the woofer and tweeter to start and stop in sync. This is how I understand the proper square wave reproduction. The phase tracking doesn't really tell about that?

To understand, I must ask, does the "phase response" tell us BOTH: When both drivers START reproducing sound and STOP reproducing sound? And what's inbetween..

Hi, my understandings/experiences/2cents

1. Square wave reproduction needs two flat lines....
magnitude (frequency response) and phase.
Needs both, nothing more/nothing less; that simple.

2. Tilted phase (straight-line or curved) on a normal graph where frequency is plotted logarithmically will not reproduce square waves.

If frequency is plotted linearly, a straight-line tilt will reproduce square waves.
All it takes is a constant delay across all frequencies to take the tilting straight line to flat horizontal.

3. As others have said, phase is relative between frequencies.
What it means is that if every frequency were simultaneously played all from the same starting instant, that they all start together from zero amplitude and then each would peak 1/4th their respective periods later.
This is what a dirac pulse is, the summation of the first sample (and first sample only) of every frequency.

But obviously all frequencies can't end together, as they have different periods. (Unless you find the least common multiple of all frequencies' periods and make that the time duration of the test...methinks we would grow quite old waiting for the test period to end ....hey, maybe that's the end of time loL)

The greatest meaning i take from phase, is that when it's flat (and magnitude is too) all the various measurements we look at will look great too.
Whether square waves, impulse response, step response, etc.....they are all just a function of mag and phase...all saying the same thing from a different vantage. In the end all there is is frequency, and timing imho.

Ime, square waves are the hardest of those to reproduce....they have the finest resolution. here's a set from a 3-way M-T-M. They were admittedly hand picked frequencies, and also from a single measuring spot, but even with all that i believe these represent super results.


Those were done with dsp, linear phase.

I see you are probably planning to use analog and i think your leaning towards a good coax and sticking to very low order makes total sense.
But honestly, going the ladder delay route looks nuts to me for a number of reasons, especially given today's excellent dsp tools.

fwiw I feel my attention to phase has brought great sonic benefits, but of course many others say flat phase/square waves/etc is no biggie.
Best of luck, hope the answers above helped...

 

[bwaslo]

There is another requirement for a speaker or system to reproduce waveforms like square waves. The phase response needs to be not just flat and horizontal, but that must also occur at 0 degrees (or 180 degrees)*. Something with a flat phase response at, say, 45 degrees will make a mess of most waveforms.

A lot of speakers can reproduce a square wave at some selected frequency and microphone location. A more honest test is to use a sweep of square waves and watch a triggered oscilloscope trace of the response. Here's a sweep like that done for a sort-of linear phase speaker from some years back -

square wave sweep 100Hz to 4kHz - YouTube

This speaker still has places in its response where it goes awry either from phase/delay and/or magnitude getting too far off. And if the microphone is moved back to where room or cabinet reflections take over, then of course it all goes to heck!

*flat, with compensation for time delay applied, that is -- adjust the effective delay for flattest response at high frequencies.

 

[mark100]

Agreed on both points....

i feel flat (at 0 deg) is implicit, but it's definitely worth stating.

That's a great looking sweep !

 

[noviygera]

ernperkins,

I did look at the Zaph ZD5 design you point to. Looks to me like the ladder delay network on the tweeter can just be substituted for a waveguide tweeter that delays physically, in a nice way. Why aren't waveguide tweeters used more often for that reason?

My point was a little different. I was wondering if it's possible to use the "lattice phase eq" for a wider bandwidth -- to flatten the phase of the entire design, say between 100hz and 10khz? What Mark100 said: "a constant delay across all frequencies to take the tilting straight line to flat horizontal".

 

[phase_accurate]

This can be done and it is also being done. But it means quite some effort. There is no closed-form solution for designing the allpass-equaliser and some type of numerical approximation has to be used.

Regards

Charles