Recently I was commissioned to design and build a small horn for the LCY K100 ribbon tweeter. The customer provided the tweeter so that I could get accurate physical measurements of the throat transition area along with the mounting hole pattern.
The horn I designed is based on the latest biradial technology using the ES horn flare curvature.
I began testing the stock tweeter by making a test baffle from some scrap walnut I had around. (see below)
I decided to add some bevels to help with edge diffraction. The baffle width is 15cm. The baffle is 5cm thick. I also flush mounted the tweeter.
Below is the acoustical measurements for the stock tweeter. The on-axis frequency response is shown below. We see a reasonably flat response from 6kHz and up.
Burst decay is shown below. There does appear to be a small High-Q resonance in the 9-10kHz region.
The waterfall plot (CSD) is shown below. Again, something happening at 10kHz region.
Step response is shown below. Transient response for the this tweeter is very fast. However you can see that after the initial rise and fall there is a sharp interruption in the middle of the wave as it attempts to settle.
This indicates a reflection very close to the diaphragm. It’s pretty obvious where this could be coming from. The tweeter bezel creates an obstruction with its square to round shape change. The design makes little effort to remedy this drastic shape change.
Conducting off-axis produces the following colored polar map. We see the dip at 9kHz on-axis being filled in as we move off-axis producing a strong ridge. We see a 60 degree listening window at 15kHz narrowing to only 40 degrees at 20kHz. Normally we’d want to see this at least double that angle for decent soundstage width.
I then removed the stock bezel from the tweeter and mounted horn No.1702 using the same four screws that came with the tweeter.
I then re-tested with the frequency response shown below. There does appear to be decent flatness 10kHz and up. But our issue at 9kHz is no longer a dip in the response but now a sharp peak. However below 9kHz we see a drop in the response. With the stock tweeter we saw a bump at around 7kHz.
Overall sensitivity has increased by 2.5dB in the 10kHz and above region. This would be welcome for those looking to pair this with an 8” fullrange driver for example.
Below is a comparison between the two responses. The horn adding sensitivity across the spectrum. Oddly the dip at 9kHz has now become a peak. The horn adds about 7dB sensitivity in the 2kHz-4kHz region. But this is not enough to sum to a flat response with the upper treble. So we are left using this tweeter for 8kHz and up.
Burst decay with Horn No.1702 is shown below. There still seems to be the pesky issue at 10kHz, although it is not nearly as bad as the stock tweeter.
The waterfall plot for horn No.1702 is shown below. We definitely see things clean up however the 9kHz region still has a resonance-15dB down in the noise floor.
Step response is shown below. Notice how the sharp disturbance is almost completely gone.
This could be remedied even further by ensuring an absolutely perfect physical transition even where the CNC cutter forms an arc. This could be cleared out to a sharp corner to match the exit shape of the driver.
I then conducted off-axis measurements for the No.1702 horn which are shown below.
The off-axis colored polar map shows a more consistent and controlled behavior. The listening window at 15kHz has increased from 60 degrees to 100. Coverage at 20kHz has increased from 40 degrees to 100. This means that your listening room’s early side wall reflections in the treble region will be about +5dB louder. This has the effect of really widening the speaker’s perceived soundstage.
Measuring intermodulation distortion using a 12 band per octave multitone test signal with stock tweeter at 90dB is shown below. Distortion is high at -45dB at the 10kHz region and slightly lower at -50dB in other parts of the bandwidth.
Horn No.1702 (Below)
Comparing the same with the horn in place is shown below. We see an overall distortion improvement of about 5dB. However the issue at 10kHz still persists which we already knew from our previous test results.
Where to go from here?
Lowering the test SPL down to 85dB does see distortion improve in a linear fashion, so this tweeter can reward if a lower listening level can be accommodated. This is in keeping with what is generally understood with ribbon tweeters, they sound great if you don’t ask too much of them.
Although we are still not at the performance target of -65dB for sound quality. The 10kHz region is only achieving -55dB resolution. Something needed to be done. I decided to plug the driver throat with 30ppi open cell foam. (see below) to see if I could damp the resonance occurring. Not knowing the exact source, the addition of the foam is just a guess.
I then retested the horn with the foam in place. The resulting test data is shown below. Distortion is now -60dB down. This is with a test signal of 85dB along with an 8kHz high pass filter.
I then looked again at the frequency response with the foam in place, which is shown below. I changed from a 8kHz to a 1kHz high pass filter for this result. As we can see, things are much better in the 9kHz region.
It is clear that this small ribbon tweeter has limitation on overall output capability. The sound quality is strictly conditional on the following:
- 8kHz or higher crossover point
- 85dB maximum SPL using multitone test signal at 1 meter
If your specific application warrants, this tweeter can offer razor sharp transients along with very wide off-axis coverage even into the upper treble.
I setup the tweeter for critical listening and found the tweeter to sound pleasant overall. It definitely has that signature ribbon sound character which is ultra delicate and revealing of source material. I found the tweeter to sound a little too gentle, lacking in realism when called upon to reproduce the vibrancy of cymbals for example.
It’s not that I want the tweeter to sound vibrant, but I would expect cymbals to sound convincing which this tweeter certainly could not do. Perhaps this tweeter should be used in a near field or desk monitor application where close driver spacing can be achieved. The tweeter sounded great just by removing the factory bezel and placing it on top of a speaker cabinet and using it as a super tweeter. I’m not sure why LCY doesn’t just sell the tweeter body without the circular bezel. The tweeter body itself is quite small and unobtrusive. It could even reside inside a small wood cube enclosure with a protective mesh screen on the front. It could look quite nice as an add-on to many speakers.