EQ and mid-bass horns

[smf]

This is probably a stupid question, but that's not stopped me before ... I'm trying to design a mid-bass horn running from c.100-120hz to 6-800hz to fit in below a 400hz spherical wave horn and tweeter horn. I'd like to use a pair of Altec 414a's that I have. Due to WAF, I'm constrained as to horn mouth size, overall depth and placement, and so I've been looking at a square 150hz truncated tractrix (as the shortest horn profile). The response is (obviously and according to Hornresp) bumpy as a result - +/- 3dB over 150hz-800hz. However, if I EQ the horn and lose 3-4dB, I can obviously flatten the response out and gain extension down to 100hz - or at least that what it looks like I can do from the simulations. So what am I missing? Is there a reason this isn't sensible? I'm likely going to have to attenuate the horn anyway down to 100dB as I'll be pushing my planned bass/subwoofer if I try to go higher than that.

The alternative is just to try to run the horn from 150hz and run the bass/subwoofer up to there (with quite a bit of care on placement). Any thoughts?

Thanks in advance.

[smf]

A picture might help to explain what I'm looking at:

[mdocod]

Hello smf,

+/- ~1.5dB is what I'm seeing, which would be remarkably flat through that range if simulation holds true. flattening things out beyond the level of flatness shown in your simulation, IMO, is going to require in-room measurements and very fine tuning. I don't see anything "wrong" with the simulated response at all. Furthermore, the driver's natural response interacting with this simulated response may actually result a flatter, or less flat response than simulated. Again, at that point it's fine tuning and most ears wouldn't be able to tell the difference.

Regards,
Eric

[whgeiger]

It is at the bottom end of the lower 2-octive (50 Hz to 200 Hz) overlap region where you are in trouble. Horn [Fc] and dimensions need to support a 50 Hz signal. As the driver becomes unloaded, the potential for generating audible distortion increases dramatically, particularly when equalization is employed to offset the response decline occuring there. Also this condition may impose an output limit on the entire system due to the displacement limit of the driver used. I believe the Altec 414a has an under-hung coil with an [xmax] = 0.15" and an [Fs]=30 Hz.

At the upper limit, response should extend an octave above your crossover frequency as well. To maintain a smooth frequency response in the upper frequency region, use of a phase plug in the front chamber is required.

Regards,

WHG

[weltersys]

I agree completely with mdocod, your room will cause far more frequency problems than your speaker will most likely have.

I'm currently working on a pair of speakers for my backyard, they will be mounted on the wall of the house, so the design has to allow for the interaction with the wall.

Below are outdoor measurements of the same speaker, a cheap Fisher 10" in a 16" deep one cubic foot box set on three milk crates 32 inch high against a wall, microphone on the ground at 2 meters from the cabinet front. The tan trace is the speaker pointing out, the green trace is the speaker pointing at the wall about 2" from the wall.
At 440 Hz, there is a 14 dB difference in level from this change in position.

The two traces hardly look like the same speaker.

By the way, Hornresp is a great program, but the predictions are not always right, I have found some designs within a couple dB of the simulation, others have varied by as much as six.

In a low mid horn design such as yours, the low end probably will come out fairly close, though the room will have great effect, while the upper end may not look much like the simulation at all due to cone breakup that Hornresp can't predict.

Art

[smf]

Thanks all, but I obviously didn't explain clearly enough what I'm thinking of. My target system SPL is 100dB. If you look at the Hornresp simulation, you'll see that the horn hits 100dB at about 100hz and stays above 100db until past 1600hz. Any EQ will therefore be restricted to reducing the 'humps' above this level. I do not want to try to increase any dips and I do not want to try to extend the low frequency response of the horn by equalising the low frequency response up. I certainly have no desire to try to take the horn down to 50hz - well OK I have the desire but know I can't do it! The simulation says that the horn is already loading the driver to 100hz at 100dB - that's bang on where I'd like it to be. I'm well within my Xmax limits over my intended response range. I'm not doing this to get a flatter response per se. This is really driven by a desire to take advantage of a response that says (according to the sim) that I can high pass to the horn around 100hz. That gives me just a bit more flexibility in terms of what I do with my subwoofers which will run c.40hz to 100hz.

NB: all EQ is via a MiniDSP digital crossover. I run pretty steep L-R filters on my current set up - the Altec 414as in a 2 cu.ft sealed box running up to 1khz, crossed over to JBL 2420s in 400hz Stereo lab horns at 1khz, which are then crossed (passive 2nd order) to an old Goto tweeter at 8k. I want to move from the sealed boxes as I built them for Altec 755cs and 2 cu.ft is just a bit small for the 414a's which were really meant to be used in 3 cu.ft vented boxes - but this was what I had at the time and they've turned out to work quite well as a temporary solution. I could just build a better box for the 414a's but I'd rather have a go at building a mid-bass horn for fun.

[smf]

whgeiger - I was just thinking again about what you said about the horn having to support a 50hz signal. I assume that you mean that if I really want to cross a horn at 100hz, I ought to build a 50hz horn and cross at twice the flare rate. Sadly, my wife isn't that tolerant of my hobbies.

[whgeiger]

Originally Posted by smf

whgeiger - I was just thinking again about what you said about the horn having to support a 50hz signal. I assume that you mean that if I really want to cross a horn at 100hz, I ought to build a 50hz horn and cross at twice the flare rate. Sadly, my wife isn't that tolerant of my hobbies.

Or alternatively raise the crossover point to 200 Hz. That will require at least two subs to avoid directionality issues. When horn dimensions are not comparable to 1/4 wavelength of the lowest frequency to be passed, the horn function rapidly declines in both frequency response and pattern control as wave length becomes longer. In an at-home setting, the only base horn loading alternative is to us the room walls and floor as a continuation of the horn structure as devised by Paul W. Klipsch [1] a long time ago. This design approach continues to be persistently ignored here and elswhere.

Regards,

WHG

[1] Paul Wilbur Klipsch - Wikipedia, the free encyclopedia

[smf]

Yes, thanks. That was the alternative I mentioned in my first post, albeit using a 150hz crossover. I am going to be using two subs. The issue with this approach is that to avoid localisation problems I'll end up having to put the subs in what is likely not the best positions to smooth out room modes, but I'd have to experiment and measure I think to see what happens. It's all a question of choosing a sensible compromise.

I've looked long and hard at Klipsch's solutions to some of these issues over the last couple of years, but again WAF and space issues are a problem. (I live in Hong Kong and I don't have even one appropriate corner in our sitting room.)

Thanks again.