In this blog post I test the 18Sound 10NMBA520 10” Midrange on Horn No.1723 which is a 200Hz Fc exponential horn. I chose this driver for its unique Active Impedance Control (AIC) which is summarized by the manufacturer below.
AIC (Active Impedance Control) secondary voice coil for superior intelligibility, very low distortion and inductance linearization. The 10NMBA520 incorporates Active Impedance Control technology (AIC), consisting of an additional coil fixed on the pole piece and connected in parallel to the moving coil. The magnetic field generated by AIC coil has the following effects: 1) Impedance linearization 2) Acoustic and electric phase linearization 3) Significant increase of sensitivity and total SPL 4) Total harmonic distortion reduction 5) Constant power transfer By not absorbing the moving coil's electromagnetic energy, the AIC offers substantial advantages to the sound quality.
The driver has a very low QTS resulting in the typical rise in output as you move upwards in the frequency band. 
This makes it suitable for horn loading. The published impedance sweep clearly shows the AIC circuit flattening things out from 200Hz-10kHz. An impedance curve this flat is normally only reserved for planar transducers.
With the throat adapter removed on the 1723 horn, we are left with a 10cm wide x 7.5cm tall rectangular throat which is equivalent to a 4.00” diameter throat. This creates nearly a 5:1 compression ratio between the driver surface area (SD) and the horn throat area.
The driver is housed in a 15L sealed rear chamber which raises the FS from 60Hz to 100Hz.
Measurements
I started by measuring the impedance. We can see that the 15L sealed rear chamber raises the FS from 60Hz to 110Hz. 
I then measured the frequency response at 2.83V (1w) input power. I've used a 1/3 octave smoothing here since gating was not possible in my listening space.
I then flattened the response using a PEQ notch of -6dB at 440Hz with a Q of 1. I also added a low frequency shelf at 2kHz with a Q of 1 and -4dB magnitude. This results in the following overlay.
Next I looked at the burst decay. Since this is ungated it's hard to know what to attribute to my room versus the horn itself. But generally the burst decay looks acceptably clean.
The CSD plot is shown below. Again, it's hard to know what is the room versus the horn itself. The main goal for this testing is to see the distortion performance. If the distortion sets new benchmarks then perhaps taking this horn outside is worth the time and effort to see how it does in the time domain.
Harmonic Distortion
This is the main point of interest for me. The 10" driver is going to have much more radiating area than any large format compression driver, so there is potential for low distortion.
I measured harmonic distortion at 85dB and 95dB. For the 200Hz region we see H2 at only 0.04%.
For the 95dB test signal level the 200Hz region rises to 0.07%.
Intermodulation Distortion
I then measured intermodulation distortion at 85dB, 95dB, and 105dB.
For the 85dB test signal we see IMD at -70dB in the 200Hz to 3kHz.
Increasing to 95dB we see IMD unchanged at -70dB.
Increasing the test SPL to 105dB we see IMD reduce to -60dB.
Gedlee Distortion
I then measured Gedlee distortion as a function of frequency at a test SPL of 95dB. Gm is 0.05 at 200Hz dipping to only 0.002 at 1kHz.
Subjective Listening
Below is my ranking for subjective sound quality. My evaluation was limited to just the single midrange horn covering 200Hz-8kHz. I focused in on the frequency region of 200Hz to 5kHz paying attention vocal sound quality with a variety of singers both male and female.
Clarity —9/10
Soundstage depth — 9/10
Smoothness —- 10/10
Sense of dynamic range —- 10/10
I couldn't fault any aspect of the sound in the operating range from 200Hz to 2kHz. What I noticed was an extreme level of ease to the sound, effortless, and palpably dynamic. I did detect some mild coloration in the 2kHz - 5kHz region, but this is outside the operating range of the horn, as you would typically pass this onto another horn.
Conclusion
From a test data standpoint the frequency response smoothness and time domain results are inconclusive until I can get the horn outside to measure in a reflection free environment. However the distortion measurements clearly show record breaking figures for low distortion. My in-house target for low distortion is -70dB for the 85dB test signal level. But here we see -70dB even at the 95dB test signal level. This means that distortion is 10dB better than our in-house target! I don't recall seeing figures this low in any of my prior testing.
The 10NMBA520 seems to do very well when mounted to my large format midrange horn.