What's going on with these bass horns? The mystery of the missing frequencies.

[Tim Link]

I've been scratching my head lately about a whole room suck out I have at 85 Hz with my horn woofer stacks. It's a little louder here and there but overall 85 hz is way down everywhere. 75hz too. So I got to wondering how that can be. How can the woofer be moving without making sound? Could the speaker cone be breaking up? Is the cone not moving for some reason? Is it a cancelation in the horn from the mouth feeding back to the driver? If it were, wouldn't the sound still be somewhere? Where there's a null there has to be a peak somewhere else, right? Not right?

The horns are 6 feet long, bifurcated and folded a few times before joining in a common exit. They've got an 18" Dayton woofer. The throat is 5" x 15" and the mouth is 18" by 20". There are four of them stacked in the corner with their mouths connected vertically.

I had some fun and did some experimenting this afternoon. I took the woofer out of one of the bass horn cabinets and put it in the prototype cabinet, which I happened to still have because I've been too lazy to dispose of it. Before installing the woofer I drilled a hole just big enough to let me drop the measuring microphone right into the throat area. I measured the sound pressure at the throat and also in the back chamber. The results can be seen in the attached graph.

The orange line is the throat, blue line is back chamber. What it shows is that while there is a deep notch in the throat, there's still a lot of sound pressure in the back chamber, which proves the driver is moving just fine and the cone is not breaking up or anything like that.

The next graph shows an average of the room response taken at several locations vs the measurement with the microphone a few feet into the horn, which is about 6 feet long. Notice the huge peak occurring inside the horn at about 79 Hz, right between the 85Hz and 75Hz nulls in the room measurements. There's the dip at 90Hz which corresponds to the dip measured at the throat but strangely that frequency plays in the room.

I've also included spectrograms for the measurement 3 feet inside the horn and another one 3 feet away from horn in room.

Also the REW file is attached if anyone is interested.


I'm trying to figure out if there is something tricky I could do to smooth things out. Could an FIR filter maybe do something better with this than my current IIR filters can?

You can see in my pictures that I tried putting some poly fill in the back chamber and in the throat, and stuffing a cushion into the mouth to see what would happen. The poly fill did practically nothing. The cushion caused the deep notch to become wider and not quite as deep.

 

[Duke]

I'm not a bass horn design guy by any means.

A "whole room suckout" seems unlikely; my impression is that a global (throughout-the-room) issue tends to be a speaker issue rather than a room acoustics issue. And the sharpness of that dip in your throat measurement makes me think it's a cancellation notch.

Is there a corresponding glitch in the impedance curve? If so, again imo that would point to a speaker issue rather than a room acoustics issue.

A strong reflection which has an additional path-length of 75 inches would reunite with the main wave delayed by about 5.5 milliseconds. A delay of 5.5 milliseconds would put the reflection out-of-phase with the main wave at 90 Hz, theoretically resulting in a deep cancellation notch centered there. So, do you see any features in your horn which would result in a strong reflection arriving at the throat after having travelled a round trip of about 75 inches?

Again, I am not a bass horn design guy; I'm just throwing darts blindfolded here.

 

[Tim Link]

I'm not a bass horn design guy by any means.

A "whole room suckout" seems unlikely; my impression is that a global (throughout-the-room) issue tends to be a speaker issue rather than a room acoustics issue. And the sharpness of that dip in your throat measurement makes me think it's a cancellation notch.

Is there a corresponding glitch in the impedance curve? If so, again imo that would point to a speaker issue rather than a room acoustics issue.

A strong reflection which has an additional path-length of 75 inches would reunite with the main wave delayed by about 5.5 milliseconds. A delay of 5.5 milliseconds would put the reflection out-of-phase with the main wave at 90 Hz, theoretically resulting in a deep cancellation notch centered there. So, do you see any features in your horn which would result in a strong reflection arriving at the throat after having travelled a round trip of about 75 inches?

Again, I am not a bass horn design guy; I'm just throwing darts blindfolded here.

Totally appreciate the blindfold thrown darts! I fancy myself a horn design guy but I really am not comprehending what's going on here. The 75 inches is pretty much the length of the horn from mouth to throat, so that makes total sense. The reverse wave from the mouth termination shock arrives just in time at the throat to cancel the output of the driver at 90Hz. So I'd expect a room-wide weakness in output at 90 Hz, as seen at the measurement at the throat. But oddly 90Hz plays just fine in the room. It's 85Hz and to a lesser degree 75Hz that have the strong room-wide cancellation. It's also interesting that at about 78Hz there's a huge peak all along inside the horn flare. But it doesn't manifest in the room as a peak at all. It plays just fine. Somehow the cancellation gets frequency shifted a few Hz in room.

I did another experiment today using a couple of 4" drivers on the ends of a 2' long PVC pipe section. I drilled a hole in the center and dropped the mic. in. Tried open ended on one side, hard terminated on one side, both drivers playing in phase, out of phase, and out of phase time aligned to make a miniature double bass array. I was able to get perfectly Smooth measured response up to about 1000Hz before things got too wacky to fix. That was interesting because I expected to see the usual comb filtering but that's not what happens. Instead you see gradual dropping to null but then sudden reversing up to a peak, and then gradually falling off. Sometimes it gradually rises to a peak and then rapidly falls to a null. The peaks and nulls are right next to each other and that looks familiar to things I'm seeing with my horn. I'm trying to get my head around that but I think it's a combination of the usual comb filtering and impedance matching dropping and rising between the air and the drivers.

This thing may look dangerous but it's mostly harmless. Did you know duct tape can hold drivers on to PVC pipe?

 

[Tim Link]

This bizarre result of spikes and dips right next to each other I suspect has to do with multiple bounces of the sound back and forth inside the tube. I think if the sound waves were just passing by each other once each way it would look like a normal comb filter. That leads me to believe that stuffing my horns with enough damping material of the right consistency along their length would calm them down significantly, and reduce their efficiency of course. I guess I need to put some stuffing in this tube and try it again.

 

[Tim Link]

This bizarre result of spikes and dips right next to each other I suspect has to do with multiple bounces of the sound back and forth inside the tube. I think if the sound waves were just passing by each other once each way it would look like a normal comb filter. That leads me to believe that stuffing my horns with enough damping material of the right consistency along their length would calm them down significantly, and reduce their efficiency of course. I guess I need to put some stuffing in this tube and try it again.

Replying to my own post. I did some simulating in Audacity of multiple bounce comb filtering and it looks nothing like the weird spike-dip combo I'm getting. I've seen these kinds of spikes before and it's not hard to find other people's postings of such spike-dips but I haven't yet found an explanation for them that I can understand. What I am seeing consistently is that there has to be some kind of strong feedback to a driver to get this sort of thing. It can't just be reflections interfering with each other in air.