In this blog post I feature a recent front horn built. The horn is by far the largest I've built yet. It measures 44.50" x 44.50" at the horn mouth and measures 22" deep. (113cm x 113cm x 56cm) The black oak frame adds another 8” (20cm) to the overall size for a total width of 52.50” (133cm). The horn is constructed with four petals. Each petal is comprised of 44 layers of beech hardwood.
A video time-lapse below shows how I CNC machined each petal.
The horn flare geometry was provided by the customer who got it from a company called Rullit. Rullit makes custom field coild drivers and has developed this front horn. The original design utilized a 10" driver. I adapted the design to accommodate an 8" driver. I also decided to improve the throat transition with a custom CNC machined piece.
This piece fills in the voids presented by the round driver to square throat segment. The idea here is that it better acoustically couples the driver’s diaphragm to the horn lens.
Below is another IG post of the horn in my workshop just prior to delivery. I had time to test two different 8” fullrange drivers before they left the premises. The last clip Is a video where I play some music. This provides a glimpse on the sonic character. If it sounds good to you in the video then you at least know it doesn’t sound bad!
The first driver I tested was the AER BD3-B, and the second one is the SEAS FA22RCZ. Both drivers are what I had on hand at the time. Scroll down to see the test data.
Further project details
The customer has commissioned Kanzen to construct a custom field coil driver. He currently uses a Sonido SWR200-8FC as shown in the picture below. 
The customer also has a custom Rullit field coil. He plans on experimenting between drivers. I hope to visit the customer in the coming months to hear the various results.
Initial customer feedback
The customer reports that he is extremely happy with the sound using the Sonido field coil drivers. He remarks that there is no primary and secondary reflections from the speakers in his room, leaving only the direct sound which is the customer's subjective preference.
Measurements
The test data is shown below with the two drivers that I had on hand. The first test was done with the AER BD3-B 8" fullrange driver.
Frequency Response
Below is the frequency response of the AER BD3-B in the No.1548 front horn. The mic was placed at 1m and the room was not physically large enough to allow adequate gating of the low frequency reflections. So the result shown below is shown without gating so that you can see the overall low frequency cutoff (Fc) of the horn.
Below is the same result as above except with 1/3 octave smoothing applied. This allows you to see the general trend. The horn depth is 22" (56cm) which means we should see extension down to 150Hz. This certainly seems to be the Fc of the horn based on the graph below. Note as well that the horn was placed in the middle of the my workshop with no boundary reinforcement from corner placement. Corner placement would help elevate the lower frequencies somewhat.
Unfortunately I do not have the raw response of the AER BD3-B for comparison.
Burst Decay
Below is the burst decay. There is limited value in this measurement since I was not able to gate out the room reflections. The only take-away here is the behavior in the upper treble. But generally we can see a resonant peak just above 2kHz as well as 9kHz.
Intermodulation Distortion
I tested the lower and higher frequencies using a multiband test signal.
Below is the intermodulation distortion at 85dB SPL for the lower frequency range. There are pronounced intermodulation artifacts (dark grass area) at both 500Hz and 2kHz. The overall noise floor is -35dB (1.8%) in those areas.
Raising the SPL level +5dB to 90dB SPL at 1m sees the following result below. The grass rises +10dB which means we lose dynamic range. Distortion at 500Hz is now -30dB(3%).
Raising the SPL again another +5dB to 95dB at 1m we see that the distortion continues to rise in a linear fashion. We have around -25dB (5.6%) at 500Hz.
We can see that by observing the distortion levels in relation to rising SPL we can get a glimpse into the nonlinear function of the driver. We could conclude from this that the driver prefers to be played at lower listening levels. This may be related to the limited xmax of the driver.
Below is the IMD for the high frequencies with the playback level set to 85dB at 1m. Generally the distortion follows the up and down frequency response of the driver, remaining low at -55dB below the fundamental.
Below is the SPL raised to 90dB SPL at 1m. The noise floor is now at -52dB.
Below is the SPL at 95dB SPL at 1m. The noise is now at around -50dB.
I then mounted the SEAS FA22RCZ and conducted the same set of measurements. Below is the unsmoothed frequency response. We can see severe frequency response irregularities in the upper treble. However I don't see the +7dB peak at 2kHz as with the AER.
Apply 1/3 octave smoothing is shown below.
Burst decay for the SEAS FA22RCZ is shown below.
I performed IMD tests on the SEAS FA22RCZ which is shown below. I started with the same 85dB test signal. The noise floor is -55dB. This seems to be a better result compared to the AER which was at -35dB.
Below is the IMD at 90dB.
Below is the IMD at 95dB. Distortion continues to be low at -50dB.
Looking at the high frequencies we see the following result. Noise is around -60dB down.
Increasing to 90dB we see the following result.
Increasing to 95dB we see the following result. Distortion remains low at -50dB (0.30%).
The SEAS FA22RCZ seems to have superior distortion performance against the AER BD3-B. AER is short on published data to help guide design directions. The website has only this to say... "Wonderful sounding driver suitable for basically any type of loudspeaker.
They excel on an open baffle supported with a bass system like the Quasar MkII but also within a FAST-system but is a suitable rear loaded horn or TQWT these are able to produce decent bass as well."
Listening Impressions
My listening session was limited to my workshop the night before pickup. The horns had to be picked up by the customer the following morning. I hope to visit the customer in the coming months to get a better sense of the sound. I did get a chance to run through some of my favourite songs. I can report the the AER sounds very good. The overall presentation is balanced across the frequency spectrum. There was no horn colouration to the sound. The large scale horn sound was certainly there. This is bound to impress anybody that hears it.
I replaced the AER for the SEAS fullrange and I immediately noticed a drop in sound quality. The overall timbral balance was lost. There was too much mid-bass which muddied the sound. I don't think this was the only reason for the loss in sound quality. The midrange had lost some of it's liveliness as well. This may be attributed to the large dip in the frequency response at around 3.5kHz.
Conclusion
I think I can safely conclude that the horn is only helping things here. What you are left with is selecting a driver that pairs well with the horn. The horn demands a driver with a low QTS (strong motor). The horn really needs a driver with a QTS below 0.3, and ideally something around 0.20. I thought the AER driver sounded really nice, which didn't correlate with the distortion tests I conducted, however this may be due to good match-up where the SEAS failed in that regard. I received emails after posting to IG asking me how the horn sounded compared to the other products I test. All I can say is the horn sound very good. I look forward to hearing these horns more once I visit my customer again. He is excited to receive his Kanzen fullrange drivers.