Effect of flat surfaces facing IB drivers.


It has often been recommended that IB drivers should not face flat surfaces. Particularly close walls.

The subject was raised on AVForums with regards to an IB project. (DIY Subwoofer Build forum)

It was desired to have a pair of well spaced manifolds but this brought the 18" driver face to about 18" from a solid wall. There was some discussion (and disagreement) of the likely effect of their relatively close proximity to each other.

It seemed worthwhile to try and prove, using REW, whether an IB is really affected by flat surfaces. Reflected pressure wavefronts returning from the wall could theoretically affect the large driver cones. Or the wall might increase the pressure seen by the cone. Not a desirable effect in an IB if it affects the TS parameters.

I decided to place cardboard boxes on the floor to provide nominally flat surfaces facing my 15" drivers. The boxes reached almost to the top of the lower four drivers but were open at the top and sides.

The flat surfaces were placed about 6" from the faces of the lowest four drivers in my 8 x 15" driver manifold. Due to the access door to the IB enclosure I could not place the box on that side any further away. Only the bottom four drivers of the IB manifold were driven.

My Galaxy 140 SPL meter was placed on top of a pile of cushions. About 6" from the mouth of the manifold and about a foot above the bottom of the manifold.

The lower 4 x 15" IB was swept 3 times using the longest 1M REW frequency sweep both with and without the boxes in place. Each half of the manifold is physically separated by a shelf midway up the 6' high box.

A clear and repeatable difference of about 1.5dB (maximum) existed between each set of graphs. The individual traces were identical within each set above 8Hz. The set with the boxes in place consistently reproduced the sweep at the higher level.

Only at 12Hz did the sets match: Below 8Hz variations are commonplace even using the longest REW sweep. This could be due to random air movements or draughts.

It seems rather surprising that such a clear result was possible from such a simple test.

There was a difference between the waterfall graphs too. Mostly at infrasonic frequencies:

 With boards
No boards in place.

Now I ought to repeat the test using more massive boards of much greater size to see if this affects the outcome. Also to examine the effect of distance of the flat surfaces from the cones.

There is also the question of whether a diver would be similarly affected if the magnet faced the wall. Rather than the driver cone. 


BTW; I am still suffering from intermittent switch problems on the cheap CX2310 Behringer active crossover. Perhaps some specialist electronics switch cleaner would help? I have never felt the use for any until now.