Phase correct drivers

Status
Not open for further replies.
Jump to Latest
#3
EternaLightWith said:
What companies make phase correct drivers? I'd like to start a DIY speaker and want a drive that is inherently phase correct.

Eternalightwithin
Click to expand...

AFAIK no one makes a phase correct driver. Phase correct when used to describe loudspeakers has more to do with crossover construction, driver placement and enclosure issues.
 
#4
Maybe I took the guy wrong...

Richard Hardesty says there are only a few manufactors that make phase coherant speakers. Looking back I think he was referring to the speaker as a whole, not just the driver.

I feel kinda stupid now. :ashamed:

Well, I'm looking at a three-way, I think. With a 1st order crossover.

The one thing I know for sure, it's going to be either a TL or a TQWT.

Eternalightwithin
 
#6
Be careful in your design. Most designs optimized to keep the phase linear (something of very questionable audibility) end up sacrificing polar pattern smoothness, freedom from excess diffraction, power handling, and distortion (all very audible). If you're doing this from scratch and you have a lot of experience in speaker construction, design, and measurement, you've got a chance of pulling it off. If not, you may want to rethink priorities.

If you want to get a handle on the audibility of phase shift, you can build a unity gain all-pass filter with a turnover frequency at 2 kHz (a common xover point for 2-ways), and switch it in and out of a circuit while listening on headphones. You may be surprised to find that, although the circuit scrambles up square waves horribly, you can't hear its presence.
 
#7
You may want to try and find some Tannoy Dual Concentric drivers, and put them in a transmission line. They are wonderful drivers, and the tweeters are located right in the center of the woofer, giving them an inherently phase correct position. Then, use an active crossover and bi-amp the arrangement. This should give you pretty good results.
 
#8
SY said:
Be careful in your design. Most designs optimized to keep the phase linear (something of very questionable audibility) end up sacrificing polar pattern smoothness, freedom from excess diffraction, power handling, and distortion (all very audible). If you're doing this from scratch and you have a lot of experience in speaker construction, design, and measurement, you've got a chance of pulling it off. If not, you may want to rethink priorities.

If you want to get a handle on the audibility of phase shift, you can build a unity gain all-pass filter with a turnover frequency at 2 kHz (a common xover point for 2-ways), and switch it in and out of a circuit while listening on headphones. You may be surprised to find that, although the circuit scrambles up square waves horribly, you can't hear its presence.
Click to expand...

SPL and phase can be predicted with some simulation softwares, how do you predict those?
 
#9
You can get an idea of polar pattern with good software like LEAP, and a somewhat cruder idea from CALSOD (which is what impecunious sorts like me use). Likewise, power handling is easily seen in software simulations. Distortion should be measured directly. There may be ways of modeling diffraction from coaxial and similar drivers, but I don't know what that is. I've done enough measurements on those drivers to get an idea of how difficult it is to prevent diffraction; the polar patterns are usually pretty awful, and getting a crossover that will, in an acoustic sense, preserve phase is a huge challenge.

I'll be back in Alsace in a week or two- how's the weather?
 
#10
SY said:
You can get an idea of polar pattern with good software like LEAP, and a somewhat cruder idea from CALSOD (which is what impecunious sorts like me use). Likewise, power handling is easily seen in software simulations. Distortion should be measured directly. There may be ways of modeling diffraction from coaxial and similar drivers, but I don't know what that is. I've done enough measurements on those drivers to get an idea of how difficult it is to prevent diffraction; the polar patterns are usually pretty awful, and getting a crossover that will, in an acoustic sense, preserve phase is a huge challenge.

I'll be back in Alsace in a week or two- how's the weather?
Click to expand...
Are you alsacian?
The weather is beautiful here, since 2 or 3 weeks! (we only had some rain friday evening)

to be short: it's like every year, the summer starts in april here 🙂
we had 27°C today, too hot for the roller skating race we had
 
#11
I agree with SY,
the phase properties of the crossover is up for debate, Here's a link to do some reading, he states the same kind of test SY explains, he put it out to anyone to try and give thier feedback:

http://www.linkwitzlab.com/

Plus, if you look at the crossovers in those "Transiant perfect"/phase correct speakers you'd be surprised, some of them have up to 30 components(a lot of work for something that is yet to be proven audible), they are most of the time more complex than other order crossovers. Plus to have a truly phase accurate speakers it would have to be sealed, because any kind of back-wave is going to be out of phase with the direct radiation from the drivers, and this is audible! Plus once you've got the speaker out in real world enviroment the phasing of a crossover is going to be the least of your problems, you'll have sound reflecting of everything which means different time arrivals, but than again we listen to this all day long...
 
#12
bricolo: Malheureusement, je ne suis pas un Alsacien. 🙁 Mais je la visite souvent. Alsace is the most beautiful place in Europe.

ds21: In order to have a first order acoustic transfer function and have everything come out nice and flat in the frequency domain (necessary for phase coherency), the crossovers indeed have to be complex, expensive, and quite critical in design. Same with the drivers; it helps if you can have a factory do a few dozen iterations of a purpose-built prototype, but most of us can't do that.

A lot of designs with phase coherency in mind have been sabotaged by an insufficient appreciation of how difficult they are to do properly. The vast majority I've heard have major flaws in other areas.
 
#13
ds21: can you specify the full url of the article you're talking about?


SY: Alsace is nice; sure 😉 But I like some european places more. Having ever lived in Alsace has certainly soimething to do with this opinion...

when you guys are talking about phase correct drivers, do you mean always the same phase, at every frequency? How is this possible?
 
#15
Bricolo said:

when you guys are talking about phase correct drivers, do you mean always the same phase, at every frequency? How is this possible?
Click to expand...

It isn't possible. The best you can hope for is a flat phase characteristic well away from the drivers' LF and HF rolloffs. What we're speaking of is a linear phase system, where the phase is constant or linear (i.e., constant time delay).
 
#16
SY said:


It isn't possible. The best you can hope for is a flat phase characteristic well away from the drivers' LF and HF rolloffs. What we're speaking of is a linear phase system, where the phase is constant or linear (i.e., constant time delay).
Click to expand...
a constant phase for the driver, between it's crossover's high and low pass?
 
#17
A driver's phase curve will be flat (or linear) well away from its own rolloffs, without the complicating factor of a crossover; normally, both high and low rolloffs of the raw driver are at least second order, and that's not conducive to easily keeping the phase constant or linear.
 
Status
Not open for further replies.
Top Bottom