Dave's Speaker Pages
Dayton Reference Series Tweeter

February 2005

The new Dayton Reference Series tweeter, the first model in the line, RS28A-4 I'm told, is impressive. It's a 1-1/8" (28mm) metal dome with a mesh cover that has a plastic phase shield under the mesh cover for extending the high end response, typical of metal dome tweeters. I had the opportunity to test and inspect a pre-production model. My thanks to Darren.

A cursory visual inspection gives the impression of quality. The motor structure alone is exceptional (more on this later), holding promise for low odd order distortion, a key parameter of a good driver. Mark Krawiec will probably be posting results of distortion tests at some point.

Closer inspection of the motor structure reinforces this impression. There is a copper faraday cap on the pole-piece, the vent opening is chamfered and there is a real wool dome on the vent opening. All in all an excellent design.

The first measurement (below) shows a response typical of many drivers with good low frequency extension and commendably close to flat below 10K, the area that many would consider the most important. This is the area that determines the primary perception of a tweeters sound. Coupled with the excellent motor structure the result is a driver that looks to have a broad range of applications. Crossovers below 2K should be possible while maintaining relatively high power handling due to the motor. Low order crossovers at low frequencies should also be in order (pardon the pun). The breakup doesn't show because it's beyond the frequency cutoff of my measurement system.

Dayton Reference Series Tweeter on-axis at 0.5m

Experimentation with some modifications showed that there is little that the user can do to improve it in any significant way. It is possible to make a couple of minor improvements, but the audibility of such is questionable. I would make them just because that's what I like to do. The changes do linearize the response, essentially removing or reducing resonances a bit in the mid-band and above 10K. Check out the before/after curves in the next graph for the first mod. The results will be case dependent, since it relates to the supplied hard felt dome that covers the vent opening.

If the hard felt dome is positioned such that the opening into the vent is too restricted, then the bump above the highpass knee will be more pronounced. Repositioning might be recommended in this case.

The next change I attempted was my lamb's wool tweak. There was some improvement, but too minimal for me to recommend it. This is one really good driver out of the box!

Now to a modification that did show some reasonable improvement. It's also easy to do and fully reversible, though it does mean acceptance of small change in aesthetics. This one relates to the phase shield.

The phase shield is a small, circular plastic piece with self-sticking glue on one side. It's attached on the inside, so the top portion of the driver with the shield must be removed. The phase shield is first removed, then the cover replaced. Next, a small piece of felt (real wool or synthetic) with a small hole cut out of the center is placed onto the mesh on the outside. This piece should be a diameter similar to the original shield.

The effect of this is to maintain the function of the shield, but reduce the resonances that it invariable creates under the shield, between the shield and the dome. The felt is not nearly as reflective as the hard plastic, so it absorbs some of the signal rather than reflecting it. The small opening is too small to allow much through, but I did this in an attempt to relieve the pressure of the wave, further reducing the resonances. It seems to be partially successful (below). It may even be debatable as to whether or not this is an improvement, but I feel that a smoother 10K+ result is possible with a slight flattening below 10K.

The next graph shows the full set of measurements on- and off-axis for the driver with a repositioned hard felt dome (not necessarily an issue) and with the phase shield tweak. Note the very smooth response at 15 degrees off-axis (I didn't try 5 degrees). This is an excellent result. I'm planning on testing a pair as a replacement for my modified Scan-Speak 9300s.

You can see the importance of having the phase shield next. The yellow curve is added here to show what the response is without it. The response is nearly perfectly flat up to about 12K where it takes a dramatic curve downward. This is stricly due to the phase difference of the response across the radiating surface. Since the speed of sound is so high in the metal dome, the entire surface radiates at nearly the same time. But since the sound essentially arrives from the tip before it arrives from the rim, there is a partial cancellation due to the delay from the extra distance between rim and tip.

The CSD looks just as one might envision given the SPL response. I include that here just for reference. There's nothing special to note other than the fact that the decay is very fast and smooth throughout most of the passband.

I've included some digital shots of the driver at various stages of disassembly and with the phase shield tweak.

The next shot shows my attempt at tweaking the response. This is pretty simple. Just remove the cover (as above), remove the small, clear plastic phase shield, replace the cover, replace the mounting plate, then apply the felt.

The felt is just a circular piece of synthetic adhesive-backed felt of similar diameter to the plastic shield with a hole cut in the center approximately 1/8" diameter. I also tried adding a piece of self-stick aluminum tape on TOP of this hoping that there would be enough absorption of any reflected energy. There is a difference, but it is extremely small. The piece of felt alone should give you the change seen in the measurements shown earlier. Certainly the aesthetics of just the felt is better than with the aluminum on top. The nice thing is that this is fully reversible. Just take care of the plastic shield when removed because it can be replaced directly due to having a self-stick side.

I also swapped the felt dome of the RS with that of a Scan-Speak 9300. They are similarly shaped, but the original RS felt is a bit less dense. There is essentially no difference between the two in response. The dome on the left is from the RS, the right is the S-S.

My conclusion for this driver (prior to any listening tests) is that it should compete favorably with many of the excellent (and more expensive) drivers from the major manufacturers. But the proof won't be had until they are available for auditioning. I have one unit for testing only at this time. I will be buying another one when they become available to have a pair to audition.