Lockdown Experiments

Yeah it looks a bit bad on that plot, even considering the scale of the plot. Here's another one of just the mid





Still some dips, I think this one was taken in the mouth too. The dividers in the mouth are not far off parallel in this box, I think this contributes to the dips when measured close up. Here's a nice paint illustration of the problem- lots of different length paths to the mic





I expect these problems to lessen at a distance as the path differences become less significant. Hopefully. This box isn't intended to be listened to close up, I doubt anyone will be within 15m of it when in use.


Your plot looks good, do you have a link to more information about the box?

I finally managed to get outside and do some testing on a dry day with very little wind.





Here's a full range plot after setting delays and eq





I then did 10 measurements in different places on and off axis. The traces with dips around 3kHz were the measurements made further away, not sure what causes this dip yet.





Here's the average of them





I also tried turning the box on it's side to get measurements on and off axis vertically, but this caused huge cancellations wherever I put the mic. I really can't work out why though.

I dont really understand why you put the horizontal dividers infront of the mid and hf sections. These will for sure cause resonances between the parallel faces (maybe you can spot the resonaces in your frequency response by calculating the corresponding wavelenghts with the distance of your dividers).

JulianDA wrote: I dont really understand why you put the horizontal dividers infront of the mid and hf sections. These will for sure cause resonances between the parallel faces (maybe you can spot the resonaces in your frequency response by calculating the corresponding wavelenghts with the distance of your dividers).

That is a good point, I was unsure whether to use them or not. They are not glued in, I might remove them to see if there is much difference.

The partial drawing below shows the box I intend to use below the current one, it is noticeable that the dividers make a series of horns which get progressively more vertical flare and less throat area from top to bottom. None of these pieces are actually parallel, although the top pair aren't far off. The idea is to create a system where the amplitude is asymmetric allowing less SPL difference between front and rear of the audience, with this achieved mechanically rather than using many amp and DSP channels.





Then I remembered this array from a few years back





No top or bottom horn walls at all on there. Presumably this works due to the physical displacement and the aim of the waveguides as they form a progressive curve. My mid and high components have a similar vertical curve, albeit on a much smaller scale.

One advantage of using the dividers is in construction of the side walls. The very slight angles between each section make it very difficult to build without the dividers to act as guides. I could always remove them after making the horn walls, gut some reinforcement would be needed on the back of the walls otherwise it would just be glued butt joints holding it in shape.

The walls of the mid/high flare are particularly difficult to make without a former of some sort. Dividers in the lowmid definitely need to stay, they add stiffness to the box.

doober wrote: I have tried measuring at the exit of each cell- there is certainly havoc there, I can't get a decent measurement anywhere. There isn't any consistency with the cancellations. Thanks for the input, it has given me more to look into and consider.