Scan-Speak 10F JMLC horn combo - large dip troubleshooting

[Defo]

I would love some help here...

I'm working on my new speakers consisting of a pair of Scan-Speak 10F coupled to a pair of 48 cm JMLC horns pictured below.

The drivers alone have a fairly flat response. But coupled to the horns they get, in addition to the horn boost, a huge wide bandwidth dip around 8,5 khz. See measurements below. Drivers simply laying in bed.

The drivers are fairly well coupled to the horns at the outer rim of the speaker surround. See pictures for this and mic position below.

Any ideas what might be the problem here?

[badman]

Lack of a phase plug?

[Defo]

Originally Posted by badman

Lack of a phase plug?

Shouldnt need one when theres no compression ratio between horn throat and driver.

[David_Web]

Is it axis dependent?
What happens further away?

[ScottG]

Make sure it's measured a meter away from the "mouth" of the horn. Then be sure to take off-axis measurements (15,30,45,60) from that same meter distance.

[Defo]

A meter would be too far for me as the room (about 12 sqm...) would affect the measurements too much. See the RT60 near and further away below. But I made some close and further away measurements on and off axis (dont know how many degree, but see the picture below), and this is the result...

How should I interpret this?

[cuibono]

Sorry to repeat, but you just can't get reliable data the way you're measuring. Go outside and measure at at least 1m. More is better.

Your horns look beautiful BTW - who made them?

[ScottG]

Originally Posted by Defo

A meter would be too far for me as the room (about 12 sqm...) would affect the measurements too much. See the RT60 near and further away below. But I made some close and further away measurements on and off axis (dont know how many degree, but see the picture below), and this is the result...

How should I interpret this?

My guess :

Basically you are seeing the effects (or "non" effects) of freq. bounding to the waveguide.

As you move off-axis, you get closer to the waveguide and the signal "bounds" to the waveguide and gets a "lift" in pressure. Unfortunately the 0 degree axis is the the point of operation where there is the least "support" or bounding to the waveguide.

At 8 kHz the wavelength is just 1.7 inches long, at 10 kHz it's 1.35 inches long.

Now the "lift" in pressure on the 0 degree axis from 9 kHz could be in part due to the driver having some increase in pressure in this range on-axis, but probably has more to due with some additional resonance or "grouped reflection" of the driver in combination with the waveguide. Don't know. Obviously the loss in pressure off-axis in this freq. range has less bounding "support" AND is lower because of the driver itself (which has a greater natural loss in pressure off-axis at its highest operating freq.s).

BTW, what compelled you to use a standard extended range driver in a horn?

Yeah, the waveguide is beautiful, and we know the driver is very good!

[Defo]

Originally Posted by ScottG

At 8 kHz the wavelength is just 1.7 inches long, at 10 kHz it's 1.35 inches long.

The throat is 7,5 cm in diameter, which equals a wavelength of 4,6 khz, so I guess the horn/waveguide no longer provide any boost for the driver above this point?

The measurements dont quite show this tendency, but I guess thats due to the comprimised testing limitations. The dip is present at any distance in my room though, so this is most certainly defenitive anyways.

Originally Posted by ScottG

Now the "lift" in pressure on the 0 degree axis from 9 kHz could be in part due to the driver having some increase in pressure in this range on-axis, but probably has more to due with some additional resonance or "grouped reflection" of the driver in combination with the waveguide.

This could very well be. I will try to place some small diffusors inside the horn, and see what happens. Allthough one would think the horns expansion rate is too rapid for issues like these to occour. This doesnt seem to be a issue with other horns using wide range cone drivers, like Ferguson Hill and others.

Originally Posted by ScottG

BTW, what compelled you to use a standard extended range driver in a horn?

Yeah, the waveguide is beautiful, and we know the driver is very good!

I like the dynamics of horns

Compression driver dont work well as low as 500 hz without having an oversized and deep horn. Also smaller throats and longer horns beams more, and large diaphragm compression drivers requires a tweeter. AND I have yet to hear a metal diaphragm compression driver sound good. Thats why I went for a cone driver instead

The horn/waveguide is made by jzagaja.

[ScottG]

Originally Posted by Defo

The throat is 7,5 cm in diameter, which equals a wavelength of 4,6 khz, so I guess the horn/waveguide no longer provide any boost for the driver above this point?

The horn/waveguide is made by jzagaja.

Well clearly there isn't much loss in output up to almost 8 kHz. I'd look to the bounding surface and "trace-back" from there and reduce that freq. by a bit for various losses.

BTW, a "dip" is present in many designs on-axis, including the Gedlee products (..though of course the dip is marginal in those products by comparison.)

In your case I'd just put a low-pass on it around 6.5 kHz.. and a *very* steep one at that (elliptical/Cauer). Listen a bit off-axis and "blend" in a vertically directive super tweeter (above or below the horn) with a steep high-pass filter (..and I'd probably make that portion an active solution with an integrated "volume" control).