Yes, its linear phase. I havn't actually measured the total latency, but I think its somewhere around 2 seconds. Makes it difficult to do measurements, but my 20 year old CLIO can do it.
That has to do with how the AMT driver works and nothing else. As I understand it, the sound pressure generated by the inner parts of the membrane folds will cancel out the sound from the outer part. There was a thread in this forum about that, I will try to find it.
EDIT: this one:
http://www.diyaudio.com/forums/planars-exotics/161178-understanding-air-motion-transformers.html
EDIT: this one:
http://www.diyaudio.com/forums/planars-exotics/161178-understanding-air-motion-transformers.html
Last edited:
Mid bass EQ
The mid bass takes a lot of EQ. Are these room effects? That should be well below any dipole peaking.
The mid bass takes a lot of EQ. Are these room effects? That should be well below any dipole peaking.
You like curves dont you?
We'll start with a tweeter polar response. The measurements are taken at the following angles: 0 - 15 - 30 - 45 - 60 - 75 - 90, distance 1 meter, anechoic MLS measurement smoothed 1/6 oct. Not bad, but there is "something" around 6 kHz. Its not the frame posts, I think its diffraction around the tweeter itself. Will investigate this.
An externally hosted image should be here but it was not working when we last tested it.
I also notice the small peak at 13.5k and the very small dip at 12k at 0º not only become more exaggerated further off axis but by 75º off, they have shifted down in frequency to 12k and 9.5k respectively. Does this not suggest that both peak and dip are being influenced by changing distances to some reflective or refractive edge? Or is that already obvious?
Nice looking graphs! One question and one suggestion though:
Why are you using a non-symmetrical crossover between mids and tweeter?
The mid has to deal with only a relatively small frequency range. Would one unit not give you sufficient dynamic range? One would probably give you a slightly better looking polar - closer to a theoretical perfect dipole.
Why are you using a non-symmetrical crossover between mids and tweeter?
The mid has to deal with only a relatively small frequency range. Would one unit not give you sufficient dynamic range? One would probably give you a slightly better looking polar - closer to a theoretical perfect dipole.
One mid would be better, I agree, but the efficiency of those Excel drivers is pretty low... Can give it a try, no problem.
The electrical XO is non-symmetrical, but that doesnt matter at all - its the acoustic summation that important.
The 15k dip is there at all angles. May not look like that because of the graph smoothing, but its there. That one is caused by the folded membrane alone like I mentioned earlier.
The other stuff must be diffraction, like Rudolf was talking about here, and nothing can be done with it.....
The acoustic crossover is non-symmetrical as well. The tweeter rolls off very steeply below 1.5 khz, while the acoustic crossover on the mids seems cascaded, starting slowly but steepening above 2.5 khz.
Yes, the acoustic result of each driver will also be non-symmetrical, but why is that a problem when the summed response is flat? I'm using phase-linear crossovers, so any driver overlap would sum in the positive direction only. There are no phase issues at the XO point. Reversing the phase of any driver results in the same frequency response also.
With a minimum-phase filter this is far more important of course.
With a minimum-phase filter this is far more important of course.
Since my last update I've rotated the H-baffle subs so they point towards the listening position. I found that I actually lost 2-3 dB level by having them point straight forward, they are directional just like the main dipoles.
The current room layout looks like this:
One question: is the distance from the H-baffle subs to the rear wall critical? Should I have more space behind the baffle? I see pics on this forum of OB woofer placed very close to the rear wall.... do anyone have recommendations?
The current room layout looks like this:
An externally hosted image should be here but it was not working when we last tested it.
One question: is the distance from the H-baffle subs to the rear wall critical? Should I have more space behind the baffle? I see pics on this forum of OB woofer placed very close to the rear wall.... do anyone have recommendations?
It's a matter of preference and room interactions. I'd suggest moving the subs back and forth and seeing what you like.
The closer to the wall the greater the level of early reflections. Makes the sound louder but less precise. Personally I like Linkwitz's recommendation of a minimum of ~1m from the rear wall (some folks think of it as the front wall since it's the wall they're facing as well as the wall behind the speakers). I prefer 1.5-2m but rarely get to listen in environments with that much room. These are guidelines; move 0.5m, hit a different room mode, and things change. In your case you can easily redo the EQ. My personal preference, though, is to try to get things set up as best as possible without EQ; our styles diverge on that point, though for dipoles it gets a bit academic.
The closer to the wall the greater the level of early reflections. Makes the sound louder but less precise. Personally I like Linkwitz's recommendation of a minimum of ~1m from the rear wall (some folks think of it as the front wall since it's the wall they're facing as well as the wall behind the speakers). I prefer 1.5-2m but rarely get to listen in environments with that much room. These are guidelines; move 0.5m, hit a different room mode, and things change. In your case you can easily redo the EQ. My personal preference, though, is to try to get things set up as best as possible without EQ; our styles diverge on that point, though for dipoles it gets a bit academic.
Since your 21s will be two to a cabinet, you will have the potential to position them in four different locations instead of just two.
Earl Geddes advocates this approach, with variation in vertical positioning, when feasible, as being even more helpful in not exciting room modes.
His fundamental aim is for each sub to excite a different set of modes and for those to be 1/4 the strength of the worst case - a single sub. For this, assymetry is preferable, too (so the room "seen" by one sub is not a mirror image of what is seen by another).
Those who have heard his set up usually comment very favorably indeed on the bass.
It would be easy for you to try, especially given the fact you will be using separate amps per 21" pair
Earl Geddes advocates this approach, with variation in vertical positioning, when feasible, as being even more helpful in not exciting room modes.
His fundamental aim is for each sub to excite a different set of modes and for those to be 1/4 the strength of the worst case - a single sub. For this, assymetry is preferable, too (so the room "seen" by one sub is not a mirror image of what is seen by another).
Those who have heard his set up usually comment very favorably indeed on the bass.
It would be easy for you to try, especially given the fact you will be using separate amps per 21" pair
i would like to try the total opposite approach
Attaching the drivers to a solid pole via their magnets only
Reminds me of the guy who mounted drivers directly to the backwall, via a steel rod to the magnets
Anyone remmeber those pictures ?
We laughed at it
But my present experiment tells me he may have been on to something
Maybe we just didnt understand it
That he may have used the backwall as a huge ressonator is a different matter
Attaching the drivers to a solid pole via their magnets only
Reminds me of the guy who mounted drivers directly to the backwall, via a steel rod to the magnets
Anyone remmeber those pictures ?
We laughed at it
But my present experiment tells me he may have been on to something
Maybe we just didnt understand it
That he may have used the backwall as a huge ressonator is a different matter
Regarding the placement of the sub's, I wanted to know if anyone actually tested different distance to the front wall. Well... I will give it a try myself then.
Makes sense that the front wall reflection will affect sound quality also in the low bass.
I have been concidering having the sub H-baffles on each side of the main dipoles. Will give that a try as well!
Makes sense that the front wall reflection will affect sound quality also in the low bass.
I have been concidering having the sub H-baffles on each side of the main dipoles. Will give that a try as well!
Last edited:
Maybe something like this: ?
An externally hosted image should be here but it was not working when we last tested it.
