Some Kind of WAW Part 6.

Build is done lets see how it measures, but first some general ideas. I have a rough idea of how these elements measure from my fun with foam core. I am aiming for the W5-2143 to naturally roll of with a -3 dB around 100 Hz. Using a second order low-pass crossing the woofer over at 100 Hz would be the schoolbook way of doing it with a regular two-way, right?

Most readers are probably familiar with the Linkwitz transform. It is an elegant way of extending the response of  sealed woofer. Rod Elliott has another interesting take on the subject with his EAS circuit. The response of the ARGSUB840 is quite similar to the response expected from something driven by the EAS circuit and the plate amp got a variable crossover and gain.Could it be used to emulate some of the Extended Low Frequency principles?

This is a quick and dirty simulation (in Basta!) based on my cabinet, the manufacturers specification for the W5-2143 and my measurements of the ARGSUB840. The W5-2143 is expected to be -3 dB at 110 Hz and around 7 dB more sensitive than the ARGSUB840 (and a 5 dB gap at the -3 dB point of the W5-2143). This is with both elements fed by a 2.83V signal.

Simulating a second order low-pass filter on the ARGSUB840 at 110 Hz and matching the SPL level requires 6V to the ARGSUB840 (the W5-2143 is still fed 2.83V). The summed response is now -3 dB at 55 Hz and we got a bit of a 2 dB hump around the crossover frequency.

Lowering the crossover of the ARGSUB840 down to 50 Hz requires 9V to match the SPL level with the W5-2143. The summed response is now -3 dB at 43 Hz the response is almost flat in the rossover region.

It seems like some of the Extended Low Frequency principles can be gained by lowering the crossover point and at the same time increase the gain but it requires a lot of power. This plate amp is probably not the best way of executing this principle. It would be interesting to build Rods EAS circuit and couple it with a serious amplifier and a high quality sub-woofer element.

This was all simulated, now lets see how it corresponds to the real speaker. I am using near field measurements because I am only interested in the response below 400 Hz. Below is the response of the W5-2143 in this cabinet. It seems like it is rolling of naturally with -3 dB at around 100 Hz and that is what I aimed for. Please do not pay attention to absolute SPL levels since I did not calibrate with a SPL meter (and I am only interested in relative measurements of SPL in this case). 

The next graph is the near field response from ARGSUB840 at 50, 100 and 150 Hz. The gain is kept the same between measurements. 

And this graph is the combined output of both W5-2143 and ARGSUB840 with the microphone some 30 cm from the baffle right between the two elements (with psychoacoustic smoothing applied). As can be seen, 150 Hz at 50% gain creates a hump around the crossover frequency and a steeper roll-off. At the opposite end 50 Hz requires 100% gain but still doesn't reach all the way up to the level of the W5-2143. 100 Hz at 100% is a bit over powering but 100 Hz at 75% seems to be the sweet spot. Flat response and a gentle roll-of from around 43 Hz. I am presently surprised. 

That was near field measurements, time to move the speakers into the living room for some in room measurements. The difference in size between the 3WC and this new kind of WAW illustrated by the following image. The WAW is significantly more compact and my wife really liked both the shape and size so higher WAF for sure.

I discovered another pleasant surprise when I started to measure. The plate amp got speaker level inputs for both left and right channel. You might think that it would sum it up to a mono signal before amplification no matter what input you use. I measured a 10 dB difference if I only fed one channel compared to if I bridged the channel over to the other input. Feeding both channels from the same source added the headroom needed to lower the crossover to 50 Hz and still keep gain below 50%, pretty sweet. And this is how the speaker measures from the middle of the room crossed at 50 Hz. Not as even as the 3WC but digging deeper and with no crossover filter from 50 Hz and up.

I have had my doubts about the whole concept of WAW but this actually works. The mid-range magic of the W5-2143 is back without the constraint of the limited X-max when running it full range in a ported enclosure. Where it excels is vocals and single instruments like in small settings jazz and soul. Is it a better speaker than the 3WC? In general no but it is better in the few areas I feel the 3WC falls a bit short. The quality of bass from the cheap ARGSUB840 is better than the one from the SB23NRXS45-8 in the 3WC. It came as a total surprise but it is probably because of the sealed vs. ported enclosure and extra X-max of the ARGSUB840 in combination with the extra power from the plate amps. It is hard to judge the difference in bass quality sub 50 Hz for music. But this WAW is delivering a firmer bass and more of a punch when used to watch movies on the projector. The mid-range of the W5-2143 is more pronounced but it could be attributed to the higher sensitivity in the 1 to 3 kHz area (and no -1 dB/oct slope). But it is also unaffected by any crossover above 50 Hz. It is not as flat as the 3WC could be made to be (without the -1 dB/oct slope) but it works better in our listening environment for what we are using them for (mainly low volume music and movies).

I think the mid-range of the 3WC can be improved in the current enclosure. I also think the bass of the SB23NRXS45-8 in the 3WC can be improved with another alignment (and it would probably require a totally different box). But this new kind of WAW is what got me really exited. It would be really nice to take this concept further. With a high quality subwoofer element, maybe 10 or 12 inch instead of 8, and a purpose built amplifier. I am thinking about something like a combo of Rod Elliott's P68 amp, P48 sub controller and P38 auto power on...