In this blog post I would like to feature the BMS 4591 Midrange compression driver with Horn No.1670 which is a revised version of the ES-290 Biradial.
The 4591 uses a ring radiator style diaphragm constructed from polyester. The diaphragm fires into radial slots as you can see in the image below.
Horn No.1670 is a revised version of the ES-290 Biradial horn. It features a new dome feature which I found to eliminate a subtle resonance at 400Hz. During prototyping I modelled this dome feature using clay. I then incorporated it into the design which is now CNC machined into the wood. I point to the dome feature in the image below.
I've also added a ten degree draft angle to the side walls of the horn. This allows the CNC cutter to reach deeper into the part allowing more of the horn to be machined at one time. Now the entire bottom and top halves are machined independently. I no longer have to stack layers of wood. This reduces the chance of quality issues.
As you can see, the horns turned out flawless. The new ten degree draft angle provides a more aggressive appearance which I think looks cool!
Horn No.1670 is part of a larger system called 1672. Details on the complete system is shown below.
Horn No.1670 features a rear plinth to mount the high frequency horn. The plinth extends far back enough to allow physical time alignment between the HF and MF drivers. This is a critical design element in terms of successful integration of the two drivers. If the high frequency driver is left too far forward, then it will be perceived as a distinctly different sound source. In other words, the sound will not be cohesive or coherent between the midrange and treble.
Below is the frequency response of the BMS 4591-8 on Horn No.1670. I measured in my living room with a 2m mic distance. It's still too cold outside to conduct measurements outdoors. My gate is set to 4.8ms before the first room reflection.
For comparison, the published data from BMS is shown below.
If I truncate the graph similar to mine we can more easily compare.
We can compare further if we overlay the images and apply transparency to my results. This allows us to see the impact of the horn on the response. As you can see below my horn is providing increased sensitivity above 2kHz. This is expected since all constant directivity horns have a slight falling response that sometimes requires EQ to correct. My horn differs from constant directivity in that it uses an exponential horn flare geometry. The frequency response peaks are more pronounced with my results (700Hz, 5.5kHz,6.5kHz). I'm not sure why my results are showing more pronounced peaks, however we can ascertain that the peaks are inherent to the driver and not related to the horn. The peaks are also subtle enough to not warrant concern. Relatively speaking, these are still very good results.
Off-Axis Frequency Response
All exponential horns have a gradual narrowing of the directivity as frequency rises, and this horn is no exception. What is important here is that the response is consistently even as you move off axis. Any disturbances will be an indication of edge diffraction which is a source of resonances, which give horns a coloured sound. There is a minor irregularity at 4kHz, otherwise this is a textbook polar map.
Below is the step response. The step response is as clean as any I’ve seen. Basically it’s a non-event.
Below is the burst decay. Using a -35dB vertical scale highlights any potential resonances either acoustical or mechanical. There are some very minor resonances in the -25dB to -35dB region, but there is nothing above 25dB with the exception of the something at 400Hz. This can be seen as a +2.5dB peak on the frequency response graph. This is something inherent in the driver, although it is very minor.
The intermodulation distortion results shown below indicate essentially a distortion free driver. Distortion is -60dB down at 90dB SPL@1m listening level. This is the limit of my measurement system. I could not measure at higher SPL since there were too many random noises in the house at the time of measurement.
Subjective Listening Impressions
This driver has the same sound characteristics as the B&C DCX-464 midrange section, which is exemplary, providing immense authority to vocals. There is zero hint of harshness as well. This driver could easily become the underpinning of a great system. This driver is also slightly more affordable than the B&C DCM420. The BMS 4591 is also a true 2" driver where other drivers use an adapter from 1.4" to 2". Does this equate to better sound? I’m not sure. But if a system is going to have a dedicated mid-horn, why not stay with with 2 inch?
My listening evaluation was done with the horn integrated into system 1672 (see diagram above).
For my testing I rigged up a simple high pass filter using a single 39uF capacitor. I allowed the driver to roll off naturally in the high frequencies which is around 7kHz. I did not find the need to apply any low pass filter. The BMS 4591 produced acceptable sound quality even in the upper registers around 4-7kHz.
I was able to blend the high frequency horn with ease since the BMS provided a nice roll-off. This is good news for those wanting to integrate super tweeters such as the Fostex T925A.
Overall I was very pleased with the BMS 4591-8. During my listening, my partner looked up from her reading and said “that sounds really really good!” That’s always a good sign!
Please note that I have no conflicts of interest in this review other than wanting to find great drivers for my horn and speaker system products. I purchased the 4591 for use in a customer's project.