System Efficiency when using multiple drivers

[GRollins]

Wizard,
Use a chimney for a transmission line...it's easier. Sounds good, too.

Grey

[P.Lacombe]

Extensive measurements done in large fields have gived results such +1.8 dB for two drivers, and 3.2 dB for four drivers, with constant power of 1 Watt *for the entire array* of 1, 2 or 4 drivers.

Drivers were mounted on a piece of 20 mm thick plywood, fastened in the embrasure of an open window. Measuring microphone (Bruel & Kjaer 4133) was disposed outdoor, at 1m in the axis. Measuring frequency : 50 Hz. Precision of the B & K microphone : +/- 0.1 dB. Drivers : Peerless L 825 WG 8 ohms, resonnant frequency 20 Hz or so.

Similar results have been found 50 years ago by G.A. Briggs, R.E. Cooke and many others. This loss over theoretical values is caused by incomplete acoustical coupling of the drivers, I presume.

Regards, P.Lacombe.

[AudioFreak]

Just for clarification I was originally referring to the 2 mid-woofers in an MTM configuration enclosure

[walker]

I tend to believe a measured effect over a calculated or perceived effect but what we are trying to do in the end is enjoy a perception, not get off on measurements. In the end the ears (and mind) must win.

P.Lacombe’s post is very interesting (thanks) in that the acoustic coupling was measured with reputable equipment. If it were in error it would not be the first time that extremely accurate equipment was used inappropriately. I tend to believe that it’s accurate but it is much lower than I have experienced. I’ve used quality microphones (not calibrated) and VU meters with signal generators when setting up PAs.
Note; you will NOT get +6dB with 4 drivers but something less.
He also lists more than one source, (the evidence mounts) it’s pretty compelling. I will measure the gains on my new system as soon as I can.

Kelticwizard, the effect works for ported systems, (the bass bins that I have used were often ported) as long as you increase the box and port size appropriately. The rear wave is in a small closed system (therefore impedance matching would be better) so I don’t believe the gains will be as great out the port. Hopefully some one can confirm this for us.

With regard to measuring the gain at different frequencies, at low frequencies the drivers can all coexist within the same part of the waveform so only one wavefront is produced. At higher frequencies this can not be achieved and more than one wavefront is produced. These mix and produce nodes and anti-nodes, nodes are where two high pressures mix to produce twice the pressure, anti-nodes are where a high and low meet and null out. The points in space where these nodes and anti-nodes exist are dependent on a number of parameters one of them frequency.
Example:
Two tweeters are fixed to an infinite baffle, driven with a signal generator, say 1kHz signals, (in phase) and measured at a distance of 1m on the centre line. The frequency is now varied at a constant voltage. As nodes and antinodes pass the microphone the SPL recorded will vary from +3dB (compared to a single driver) to zilch, (in practice you won’t get the extremes due to reflections and diffractions). See I told you it gets messy.

Why would anyone in their right mind manufacture a speaker with more than one tweeter? Because these nodes and antinodes always exist caused by, you guessed it, reflections and diffractions. Still I prefer one driver when high SPLs aren’t needed. I’m building a speaker with two tweeters, for most listening one tweeter will be shorted out, (only used at high volumes).

I know that this isn’t the whole story, hopefully it covers the salient points.

Regards WALKER

[P.Lacombe]

Walker,

It's possible that poor power increase was caused by the drivers used for experiments. Peerless L825WG are cheap woofers with relatively small magnet, magnetic field is 9000 Gauss or so. This experiments have be done 20 years ago, perhaps it's time to start again...

Regards, P.Lacombe.