Oh, great!
I'm surprised that the interface pattern is very soft (broad dips) compared to the throat reflection. Maybe being inside a wg smears the polar pattern compared to a flat baffle. Dips are also higher in frequency than what I initially thought. This is good news since it enables the mid ports to be placed further from the apex compared to the 1/4-wavelength (mid spacing) rule. Easier to time align with a tweeter and less diffraction from the mid ports in the tweeter sound.
Thanks!
I'm surprised that the interface pattern is very soft (broad dips) compared to the throat reflection. Maybe being inside a wg smears the polar pattern compared to a flat baffle. Dips are also higher in frequency than what I initially thought. This is good news since it enables the mid ports to be placed further from the apex compared to the 1/4-wavelength (mid spacing) rule. Easier to time align with a tweeter and less diffraction from the mid ports in the tweeter sound.
Thanks!
Just to stay within context, this is the regular ST260 (again) -
("SP" is the equvialent total radiated power, i.e. a FR of a point source radiating the same total power.
The on-axis ripple can be further eliminated, as was shown before, that's not addressed in this model.)
- I keep using it because it's small enough not to take a long time to calculate, yet it still covers all the design aspects.
("SP" is the equvialent total radiated power, i.e. a FR of a point source radiating the same total power.
The on-axis ripple can be further eliminated, as was shown before, that's not addressed in this model.)
- I keep using it because it's small enough not to take a long time to calculate, yet it still covers all the design aspects.
Last edited:
Yes, the 1/4 wl rearward reflection from the mids output into the apex, arrives back at the mids for 1/2 wl out of phase for maximum cancellation......
it's what folks building synergies commonly call the "notch" frequency.
An electrical low-pass for the mids needs to be below that notch frequency, and establishes their upper useful range.
There's not much getting around this 1/4 wl 'rule' with regards to ports placement distances from apex, ime.
In contrast, the principle that drivers ports stay within 1/4 wl of each other, as Patrick says, can be bent some.
My measurements find exceeding 1/4 wl comes at the expense of worsening off-axis responses. Slightly, to majorly, depending on separation distance...(just like a MTM or anything else).
Anyway, my take is the 1/4 wl principle is mainly about smoothing polars, more so than about max SPL.
How do you set boundary elements like this? I'm wondering what happens when you include the conical portion of a compression driver such as a B&C DE250. In that case, 35mm length total (~10mm of phase plug above the diaphragm, 25mm of conical horn to the exit of the compression driver).
I was once paid to evaluate Tom's patents. I found that the 1/4 wl was specific to his claims. But I did not see any reason why this specific length was chosen. It has always seemed to me that closer would be better, but for some reason his patent did not cover that situation. Patent claims are tricky things. He was likely more specific than previous implementations - for reasons stated in the patent. So his was an "improvement" on "prior art", so he was covered for the 1/4 wl situation. But 1/3 wl would probably not infringe.
This was my reading of the situation at the time. Others may disagree. Those patents may have even run out by now.
This was my reading of the situation at the time. Others may disagree. Those patents may have even run out by now.
To me this looks pretty good!
Just a thought experiment - could you rerun this, making the red areas completely absorptive and use the throat the driven element?
The 1/4 wl notch is typically fairly steep, so it depends on the order low-pass desired.
Low order needs a little breathing room; high-order can bang right up against it.
In either case, best answer ime, would be pretty close.
https://jblpro.com/en/site_elements...d700-series-co-axial-mid-high-speaker-systemsIsn't a "reversed MEH" a better proposition. The more favorable location for a driver should be at the end of the WG - no? High frequencies, say > 6-7 kHz, could instead be injected via holes in the side of the WG - those holes should be possible to make smaller than for mids and be less intrusive to the WG in general (long waves skip small holes, short waves fall into large holes). So, a (larger - means we need to leave the "260" territory) WG fed by an ordinary cone driver usable with desired directivity from say 500-600Hz should be much more desirable than optimising a WG for being perfect above 13kHz ->.... Small CDs for 6kHz->
JBL solve this by using a coaxial horn arrangment.
Thank you mabat, these are some really useful findings!
The following is beyond the scope of your investigation, but I believe folks need to keep in mind that we want the both drivers to sum as if they were one wideband driver and for that we need to account for both the driver(s) and crossovers phase shifts.
You are always adding the the magnitude response and phase shift of the driver and this will determine the physical location of the ports (taps).
The distance between the ports will determine group delay and must be accounted for in the filters, reversely you can slide the ports back and forth to accommodate the filter and/or you will have to design the filters in accordance with the actual circumstances to achieve the desired summing.
The following is beyond the scope of your investigation, but I believe folks need to keep in mind that we want the both drivers to sum as if they were one wideband driver and for that we need to account for both the driver(s) and crossovers phase shifts.
You are always adding the the magnitude response and phase shift of the driver and this will determine the physical location of the ports (taps).
The distance between the ports will determine group delay and must be accounted for in the filters, reversely you can slide the ports back and forth to accommodate the filter and/or you will have to design the filters in accordance with the actual circumstances to achieve the desired summing.
I am sorry if I appear dense, but are you saying that if the individual polars look ok, irregardless of if the mid port is located more than 1/4 wavelength from the treble, then in your sim they will always combine correctly and not radiate independently?
Afaik you can only correct one dimension in space with DSP, so in this case spatial issues will have to be corrected in the spatial domain.
Then again I might not fully grasp what you are saying.
Afaik you can only correct one dimension in space with DSP, so in this case spatial issues will have to be corrected in the spatial domain.
Then again I might not fully grasp what you are saying.
Last edited:
Basically, yes, I believe this is how it works (inside a horn).
That's true, and if the polars are OK, there are simply no more spatial issues to be corrected. Relative phase responses can be set freely in the crossover.
- Maybe I haven't thought it through enough, I'm not sure.
Last edited:
They can't be located further away because of the cancelation notch. I consider this given.
The closer, the better.