Could you please clarify what you mean? If I understand correctly, the link is about not having a phase rotation at the crossover frequency, but it does not mention anything about directivity.
Last edited:
Hi TBTL
There is no mention of dispersion properties in my Orion. The reason I provided the link is because what lead to the Orion was exactly what you said earlier.
I started it out with an active 24dB/oct LR crossover at 2.5kHz. While it sounded fine on axis, I picked up some strange behaviour off H axis. I was hearing some strong phase cancellations about 30 deg off center onwards. I figured the narrow vertical pattern must be interfering with the woofer in the crossover region. I didn't do any measurements because the anomaly was obvious.
That led me to go for 1st order. The problem disappeared. And since I'm at 1st order, I decided to go for transient and phase coherence.
I have another tweeter, a HiVi RT-1C Planar which doesn't behave this way. But then, I crossed it at about 4kHz.
I intend to do more work with the Fountek CD3.5H ribbon. This time, I'll have a horizontal mtm with the ribbon vertical. Similar to line arrays. It should integrate better.
Hi Allen
The horn may look square but my guess is the radiation pattern is more like a narrow wedge.
There is no mention of dispersion properties in my Orion. The reason I provided the link is because what lead to the Orion was exactly what you said earlier.
I started it out with an active 24dB/oct LR crossover at 2.5kHz. While it sounded fine on axis, I picked up some strange behaviour off H axis. I was hearing some strong phase cancellations about 30 deg off center onwards. I figured the narrow vertical pattern must be interfering with the woofer in the crossover region. I didn't do any measurements because the anomaly was obvious.
That led me to go for 1st order. The problem disappeared. And since I'm at 1st order, I decided to go for transient and phase coherence.
I have another tweeter, a HiVi RT-1C Planar which doesn't behave this way. But then, I crossed it at about 4kHz.
I intend to do more work with the Fountek CD3.5H ribbon. This time, I'll have a horizontal mtm with the ribbon vertical. Similar to line arrays. It should integrate better.
Hi Allen
The horn may look square but my guess is the radiation pattern is more like a narrow wedge.
The available plots show different angles on each (H/V) plot, which is disappointing to say the least. Could be a trend coming through around 10k?
In my experience general sensitivity is less at 1kHz. Not just due to larger wavelengths, but also for acceptable errors in magnitude.
No need for perfect polars, I'd just hold the mic in my hand, maybe take a handful of fibreglass, RTA spectrum. The two sides of the waveguide may be playing against each other? or the sides of the box?
Ok, that's worth a try. Is the midrange now running through breakup?
and
With the wavelengths involved, both the large tweeter (dome) and line-tweeter are effectively radiating uniformly at any reasonable distance. In other words, lobing is at worst minimal in the top octave assuming the drivers are placed near each other and 6 db loss in the top octave for the dome (low-pass behaviour).
You can also "fudge" this with distance and vertical tilt (also relative to any occurring polar-tilt from the crossover).
I have same question about possibly running mid into "breakup"
Also I wonder about the "line" often drawn at the "breakup" region. After spending most of my life listening to full range ribbons ( large free swinging ribbons with little to no tension) , I hear color in even smaller cones that are taken much above 1.5 -2 khz wich is often a bit before what is considered breakup. As if the flexures in a cone even before they get wild can at times be audible. Some do it with more sophistication than others but.....?
Not sure I get what you're saying, but if I understand you correctly, you're saying that lobing in the 10-20 Khz range isn't very critical?
I do have wondered how one could actually integrate supertweeters properly due to the extremely short wavelengths requiring extremely short c2c spacing.
Where the cone is a wavelength there can be a resonance, but sometimes it takes a keen eye to see it. It is before the big nasty peak. Directivity has narrowed and the cone is in its early breakup range.
You can cross here, but it needs to be steep. Eg. preferably out by 1/2 an octave. If you want more space you should start lower.
I would say physical time alignment is more critical between 2K~3K. This is where I usually cross my woofers to tweeters.
When they are time aligned, the sibilance in the voices sound natural, not harsh.
Ok that would explain things a bit. I have experimented with physical time alignment BUT I have never heard a solid improvement. However nearly all of those experiments have been with lower crossover points typically between 700 hz and 1.2 Khz
Same here. I can't tell whether the drivers are time aligned from 700Hz to 1.2kHz.
What led me to time alignment is the sibilance. I reference it to my Grado and AKG headphones.
What led me to time alignment is the sibilance. I reference it to my Grado and AKG headphones.
No. I usually keep well away from cone breakup.
I've attached the crossover of the HM130C0 and the CD3.5H.
Both are 1st order electrical. No active or passive EQ is applied.
Attachments
Yea part of what shaped my present diaphragm design was sibilance. Nearly every single pleated ribbon Ive heard ( both foil only and plastic carryer types) have issues with this and a unique damping method was necessary to overcome.
The "Flat foil" designs can side step the pleated issues BUT that construction has serious reliability and power handling issues below about 2 Khz in smaller ribbon format.
Interesting to me that sibilances reared their head in two very different ways
- Status
- Not open for further replies.