Today I rewired the 8 drivers with single runs of 2.5mm^2 stranded speaker cable to replace the 0.75mm^2 flex. I could have used the larger size of 4mm^2 but had found the terminals on the EP2500 (and both sets of drivers) unable to take anything larger than 2.5mm^2 when doubled.
After checking driver polarity with a 9 Volt battery on the left and right channels I played Track 4 of Millennium Bell by Mike Oldfield. Switching polarity on one bass channel of the CX2310 crossover quickly confirmed that I had all eight drivers firing in unison. What was rather worrying was how little difference there was between correct polarity and reversed. Muting the stereo speakers and winding up the volume on the IB to a reading well over 100dB(C) on the Galaxy 140 SPL meter found more rattles around the room than was at all desirable. It seemed the lack of bracing was causing each set of drivers to act as an isolated vertical array. The output was stunningly loud and brutal but the manifold was obviously flexing badly. I had underestimated the reaction forces of the eight drivers in comparison with just four. The much greater height of the manifold was also compounding the problem.
It would be tempting to fix the three plywood bracing shelves which I have already cut to size. However, unless the shelves are very firmly bonded to the manifold sides, there will be very little bracing against reaction forces in the reverse direction. So I am rather tempted to duplicate the screwed rod brace of the last manifold but this time in triplicate. The use of sturdy 10mm studding with tightly fastened nuts and large, load-spreading, roofing washers inside and out should ensure the sides of the manifold remain quite still. Both inward reaction forces, and those which would tend to spread the sides of the manifold apart, should be safely restrained. I could even use up to to six or more such braces arranged symmetrically between the drivers if it proved necessary or beneficial.
Here's another image of the 10mm studding brace I used on the former 4 x 15" manifold. There are washers and nuts both inside and outside the manifold sides. The nuts are carefully adjusted so there is no inwards or outward pressure on the plywood before tightening well with large spanners. Such a rod is quite capable of accepting a load of several tons in tension. When in compression, using tightened nuts and large washers, I imagine it is just as capable of resisting the cyclic reaction loads in the plywood manifold. The reaction loads cannot be very high or the box would be crushed or ripped apart at the joints.
Luckily I had enough 12mm (1/2") studding handy to make three bracing rods. I can report a considerable reduction in structural vibration above 100db(C) [sub only] on Track 4 of Millennium Bell. It is extremely unlikely anyone would ever want to listen to full range music at this level so I can relax a little on that particular problem.
Photographing the manifold is proving extraordinarily difficult even with flash. This is the best of 40 odd images! I can't get a straight shot because of the LCD TV in the foreground. The slightly odd angle makes the rods appear non-parallel despite their hole positions being measured to the millimetre.Here's a video of the new IB playing 10hz sinewaves at a reading of 90dB(C) on the Galaxy SPL meter. With a 10Hz conversion factor of 13.45 dB this equates to 103dB. The drivers are well within their capability. I had to open the door slightly to stop it rattling like mad! The draught through the narrow gap was incredibly fierce! Interestingly, the cone excursions dropped with the door open and increased when I closed it.
http://www.youtube.com/watch?v=C_2C9M3_HLo
The video is also embedded just below:
If you decide to watch the You-Tube video in HQ full screen to get some idea of scale then click on Exit to return to standard blog size. If you have your computer speakers on you will hear the whistling buzz of the replacement (quieter) EP1500 cooling fan and me trying to be very quiet rather unsuccessfully. The 10Hz sinewave tone was absolutely silent to my ears. Nothing tickled my senses at all at this output level. Only the door shaking violently gave away that anything was happening.
Later I reached 106dB(C) @ 1 meter on the Galaxy 140 SPL meter with the door open. Add the 13.45 dB correction factor for the Galaxy @ 10 Hz. A totally silent 120db? Why can't I hear anything of the fundamental or any of the harmonics at this level?
Having played for some hours I have come to the conclusion that the different sets of drive are not well balanced with regards to reaction forces at 10Hz. The greater extension of the new drivers means that they take the brunt of the load in reproducing 10Hz. The older (more emasculated) drivers are hardly moving so contribute little or no reaction forces of their own. Further up the frequency scale they are much better balanced and both sets contribute more equally to output and thus help to cancel each other's reaction forces.
The basic advice to use all the same models of drivers still holds good. Imbalances of drivers will suffer much heavier vibration than identical sets as I have just discovered. The drivers on each side of the manifold are acting more like independent vertical arrays in my case. With all the disadvantages this entails. I owe it to myself to order four new AE IB15s to kill the vibration and probably extend the VLF performance of my IB. Away from sinewaves though the new eight driver manifold seems to be a considerable improvement on the last. Particularly with regards to sheer realism and shock value on both films and music.
*
Just in case you thought 8 x 15" drivers was slightly over the top: Here's a 12 x 15" driver IB depicted on a really great website:
http://www.subwoofer-builder.com/ib.htm
Have a good look around at the endless other projects and inspirational construction features. This website offers a great deal for the speaker and sub builder! I wish I had studied it more closely in order to do a better job on both of my IB manifolds. This website makes my blog look like childish scribble!
*
