Copyright 2014 © Troels Gravesen



16-01-2016: The C18EN001/M has been discontinued by SEAS, thus no constructions will be made from this driver. 


One driver that caught my attention among the new drivers from SEAS was the C18EN001/M, coax driver. I've never done a coax driver before - except for Tannoys - and this one looked particularly interesting, being a dedicated mid-tweeter coax unit. Another reason was that I had recently heard a KEF speaker, KEF Q700 actually, and I was pretty impressed by its midrange/treble performance - and not so much by the lack of weight in the 100-200 Hz range despite three bass drivers. The Q700 sounded a bit like a sub-sat system.
Coaxial units excel in dispersion of the mid-treble range and no matter where you sit, you pretty much get the same sonic picture. What we too often have seen from the few coax drivers available is a horrific treble performance due to the cone working as a non-optimised waveguide for the tweeter and the tweeter being surrounded by rings and stuff seriously compromising frequency response, something many diy'ers put way too much weight on (based on response to my website).
One company that has taken the concept to the extreme is TAD. Beryllium dome surrounded by a magnesium cone and a very smooth transition from tweeter to the midrange cone. The top of the range from TAD even features a Be cone for midrange!
This SEAS unit looks very much like the TAD mid-tweeter unit and I was anxious to learn how it would perform. Based on the response graphs presented at SEAS website I wasn't too optimistic about treble performance, but who knows? Maybe it wasn't as bad as it looks.

Click image to view large

Download SEAS data file for C18EN001/M.

Click image above and see the very smooth surround meant to cause as little turbulence as possible and also designed for limited excursion (dedicated midrange). The very compact magnet system allows very limited area of reflection and if you click image below you can see the ventilation holes (voice coil) in the magnet system.

Click image to view large.

Generally the finish of this coax driver is second to none. It's a pleasure to hold and handle, like a Leica lens,  and hopefully it's as sonically rewarding as measurements and simulation suggest it will be. What is likely to appear later this year is a system with either 2 x 6" or 1 x 8" bass driver and a fully passive solution to supplement the active/passive construction found at SEAS website.


Let's start out with the impedance of the midrange and tweeter. Both units are shown and display very low spread in performance.
Above to the left the midrange driver. My C18EN001 drivers are out of the box and the ~120 Hz point of resonance seems somewhat above SEAS specs but is of minor relevance here with an expected point of crossover around 250 Hz. The limited peak height suggest an aluminum voice coil former and may be part of the reason for high mechanical damping (low Qm). There's no trace of resonance around 800-1000 Hz suggesting a very well made rubber surround. Too often we have a serious resonance problem in this range, sometimes causing a serious dip in frequency response above the point of resonance.
To the right we have the tweeters and again very similar results from the two units and the low peak suggest use of magnetic oil.
The following measurements are from one driver only. All measurements normalised for 2.8V/1 meter.

Left: Middriver response from two baffle sizes, 22 x 34 cm (green) and 65 x 100 cm (red). As can be seen, a wide baffle eliminates the usual baffle step loss, and this driver looks ideal for fairly high efficiency systems on wide baffles. Very smooth response throughout the midrange! With a point of crossover around 250 Hz and appropriate baffle width, it seems possible to design a speaker with a system sensitivity around 89-90 dB. Not bad at all.
To the left the tweeter response, which is better than expected and the dispersion graph below suggest an even better overall performance. The peak at 26 kHz is well above the audible range. Listening to the MLS signal during measurements gives an immediate sense of drivers' performance (based on experience) and from both membranes nothing shredded my ears as I remember from former magnesium drivers, not to mention alu drivers.

Left: Horizontal dispersion at 0, 10, 20 and 30 deg. from middriver, flush mounted on a 22 x 34 cm baffle. This is very good indeed. Although I expect upper point of crossover to be around 2 kHz, this driver does well up above 3 kHz. To the right the tweeter response at 0, 10, 20 and 30 deg. and this was what got me excited! Very even response all the way up to 17 kHz. The tweeter may have done with a couple of dB more in sensitivity, but let's see what can be achieved. Considering the mid cone works as a waveguide for the tweeter, this performance is very good indeed!

Above distortion measurements of mid-driver and tweeter. Green = 2nd harm. Blue = 3rd harm. Input is 2.8V and measurements performed at 0.25 meter distance, hence the high SPL levels. I'm always suspicious on distortion measurements, but these were done several times without any trouble, so I believe this is what we get. Overall this is excellent performance and particular the tweeter seems second to none. Very low distortion all the way down to 2 kHz.

Left: Response from middriver from three baffle widths, 22, 26 and 30 cm (red, green and blue respectively). The response can be smoothed somewhat from a wide baffle, but I'm sure most builders will want as narrow a baffle as possible. Problem is we may have a dip at 300-400 Hz from e.g. 26 cm baffle width.
Right: Step response from middriver. I'm a bit puzzled by the twin-peak step response and I'll have to check my old W18E(X)001 files to make a comparison to other magnesium drivers.
Comment from Håvard Sollien, R&D Manager, SEAS:
We find the same profile from our step response measurements. The double-peak is a result of on-axis measurements and is due to the nature of the break-up modes of the magnesium membrane material. On this unit it is very early in the step response and coincident with the main peak since the break-up is so high in frequency. A lower break-up frequency would show up later in the step response. This is present in all hard membranes. When the midrange driver is used with a low-pass filter then this will go away.

A beauty it is! Click image to view large.


Time alignment

Now, this SEAS coax unit features the tweeter right in the heart of the midbass voice coil. Will this provide a time-aligned mid-tweeter system?
1. To find out the difference in acoustic distance of the two drivers (mid and tweeter) we can do this: We place the microphone at 1 meter distance and record amplitude and minimum phase for both drivers and finally the summed response (impedance data not needed here).It is important the microphone is not moved during these measurements.
2. Next we take the CLIO data files to LspCAD6 software and simulate a 3-way crossover with tweeter, mid and summed response as the three drivers. The measured summed response is disabled in the SPL/Xfer function, thus does not add to the total summed response. This will produce 4 graphs, tweeter, mid, mid+tweeter and summed response of mid+tweeter.
3. Next we change dZ for middriver until summed response coincides with mid+tweeter response at suggested point of crossover, because what we also find is that the dZ is not consistent with frequency. Usually a middriver will change its phase behavior because not all parts of the cone may produce the same frequencies. At higher frequencies it may only be the center of the speaker, e.g. dust-cap, that does what it should do.
Once done we can start simulating our crossover with the correct dZ, here 0 mm. Quite unusual.

Above we see the response of individual drivers and summed response. Green is midbass, red is tweeter. Orange is summed response of the drivers driven in parallel. Blue is the summed response (calculated, not measured) from bass and tweeter. By changing dZ (see graph below) for the midbass we can set the acoustic distance between the two drivers until their summed response coincides with the measured summed response (bue). Between the two black vertical lines the two summed response graphs are coincident.
As can be seen the two graphs coincide between 1,600 and 4,000 Hz - and fairly well way above and below. Actually the phase difference @ 1,000 Hz i only 17 deg.


Above the driver window inserted displaying dZ for the midbass driver (encircled). Usually from a flat panel we have to insert some 20-30 mm here to make sure we simulate crossover that will also work when we start constructing the crossover. 20-30 mm is the distance the midbass driver is behind the tweeter when we mount them on a flat baffle, hence many of my recent constructions feature stepped baffle to time align driver and allow preferred crossover. On a flat baffle we often use asymmetrical crossover to compensate for the time difference.

Crossover mock-up


Based on measurements this came up from simulation in LspCAD.

The C18EN001 calls for LR4 filters. The tweeter from performance, the midbass from necessity. Simulation was straight forward and measurements of result was spot on as predicted in LspCAD, even the notch-filter needed for taming the 6 kHz break-up.


Above to the left the basic on-axis performance from 2.8V input and 1 meter distance. To the right the response from mid and tweeter plus summed response. Point of crossover is 2.2 kHz.

Left: Reversing tweeter polarity yields this image (blue line) indicating proper phase alignment in the crossover region. To the right tweeter level from R1011 = 0 and 1R0 (green). My guess is that either 0.47 Ohm or 1.0 Ohm will be needed in final set-up.

If you can't wait, the crossover here can be used for the C18EN001 in a 12 liter closed box with the front panel dimensions of 22(W) x 34(H) cm with the driver center 14 cm from the top. Use LR4 electronic crossover for the high-pass section from around 200-250 Hz and add a bass driver suitable for some 88-89 dB system sensitivity.