Using a PA compression driver for audio is not as simply as it seems. Compression drivers on horns (like the 1505 multicel, but also the Oris 103) do not have a straight frequency response, simply because they sound better like this in big theatres. So I needed correction for this to lift the system up to a high-end audio system.

Thanks to Bert Doppenberg I was able to measure the 291-16A driver in the Oris 103 horn. The frequency response had quite a lift around 1.5 kHz, and the system also sounded like this. It was quite harsh sounding, not really nice to listen to. These curves were measured with an enlarged backchamber (about 1 liter) behind the 291 driver for better low frequency response.

So I designed a RC-CR network to correct this. This worked quite nice and made the system a whole lot better (It played like this on Arhus 2001). But it had the disadvantage of dropping the signal with about 27 dB, leaving me to need a very high amplification in my preamp to get some music... Also it just did not do the correction like it should.

To solve these problems I designed a LCR correction network. This really worked out great. Together with Bert Doppenberg we optimized the theorectical (almost perfect ;)) circuit by measuring the speaker. The results are very great. The speaker is almost flat from 300 Hz to 10kHz! This curve is not bad for a real measured speaker, and absolutely not for a horn speaker! (I can lift the dip in the center to make it really flat (replace R4 by 6k8), but listening to the system I like it better with the little dip).

Altec 288-16K driver

I also got my hand on a very nice pair of Altec 288-16K compression drivers. These drivers have ferrite magnets, but a better (lighter) diaphragm than the 291 driver. I had to try it out. We also measured these speakers on the Oris 103 and I designed a correction network again, which we optimized again by measuring the speaker.

The frequency response of the 288 (black line) looked better right away. I only had to reduce the two bumps at 1kHz and at 6.5 kHz. This was done by using two LCR networks parallel. The resulting frequency response is straight from 200 Hz to 10 kHz. The resonance at 12 kHz probably is caused by the back chamber, I still have to experiment with that. It is not really audible, using correction on the peak did not result in real difference in sound of the system...

Sound of both drivers

Both system sound really great! Great dynamics, unbeatable by any system. Also with the correction they sound very natural, absolutely not like horns and they have ultimate placement and are great in playin low level and high level at the same time.

Both systems have their advantages and disadvantages. The 288 clearly is better at the low frequencies. I sometimes have the idea the 291 breaks up a little when crossed at 250 Hz... Crossing higher of course will solve this. The 291 sounds more natural, more analoge. The 288 is more perfect, analytical, more CD-like. I still have not decided which one I like better. For now I use the 288. Next I will try to use the 288 diaphragm in the 291 magnet, they say it is the ultimate combination..

Notes:

• C3 in both corrections are highpass at 250 Hz, to prevent the drivers of getting too much energy. I also use the coupling capacitor in my poweramps to cross one order at 250 Hz. If you do not filter them at least 12 dB at 250 Hz, you might have a risk of damaging the drivers. I can play very very loud like this, so it works ;))
• Both filters reduce the signal with about 16-20 dB, so your preamp need to give enough output to drive the corrections. Both corrections are quite easy to drive.
• If you amp has different input impedance, you have the change it, these filters are designed for 100k input.
• Parts are quite critical, please use the exact values, otherwise it might result in large differences and bad sound.
• If you are looking for the coils, please ask me, I can order some more.