Dave's Speaker Pages
Lamb's Wool Tweak - My Dayton/Morel Hybrid Tweeter
26 Feb 2002
This page is about "The Hybrid" as I refer to it. It is the magnet assembly and face plate of an original version of the Dayton with the dome/voice coil assembly of a Morel MDT-32 replacing the original. I tried this because the Dayton looked like, and was reported to be, a copy of the Morels. As it turns out, the Dayton is a close copy of the Morel MDT-29, but with the dome/voice coil assembly being a copy of the MDT-32. The replacement works well. In fact, it's far better than I expected.
Dayton Dome (Left) and Morel MDT-32
There is one difference which can be seen. Both the Dayton and the Morel MDT-30 ( I have no MDT-29) have a bevel cut on the interior of the pole piece vent on the dome end. It is large on the Morel and small on the Dayton. This is likely there for a smoother transition of the pressure wave into the vent, reducing turbulence and hence possibly reducing associated resonances. The Morel is likely better at this. I see no other identifiable visual differences.
There are likely differences which can't be seen. I don't know if either one has an undercut pole piece. I don't think that either one has a Faraday (shorting) ring, although the impedance curve should indicate the presence. Since both have quickly rising impedance at high frequencies, I assume that neither one has this.
The magnet strength may be different. The gap width may also be different. The combined gap width and magnet strength determine the magnetic flux density. The sensitivity of the Hybrid makes me think that there is no significant difference, though.
The first graph shows the impedance curves, before and after the modification. Adding wool into the pole piece vent lowers the Fs a bit, although curiously the peak is increased. The Q remains similar. It also smooths an obvious resonance in the 4khz area.
The first SPL graphs show the same driver, with and without wool in the pole piece vent. It's easy to see the improvement made by adding the wool.
The next graph shows the SPL of the second hybrid with a stuffed chamber (picture below) which I tested compared to the wool in the pole piece vent without a chamber. The chamber helps, but the wool, with the vent opening covered tightly (just as the new units are supplied by the manufacturer), provides an excellent response. There is flat extension to 2khz and a smooth rolloff below that. The more surprising result is the linearity from 2-8khz, and the relatively small deviations above that. The driver is +/- 2db from 1800hz to 20khz!
I've now created a pair of hybrids, the SPL of each shown below, with the lamb's wool in the pole piece vent. They are very, very close above 2khz. The difference below 2khz is likely to be (though I haven't verified) a differing amount of wool in the vent. I wasn't as careful in ensuring identical amounts of wool. It could also just be differences in the dome/voice coil, but I suspect the wool to be the more likely reason.
The CSD (waterfall) plots for one of the hybrids, Unit 1, shows it to be very clean. The same benefit is seen in this one as is seen in the MDT-20 with the wool. The decay is faster and smoother in the low frequencies, especially below 2khz.
Now for the step responses. I show these because the step response is a good indicator of the speed and settling of the driver. The first one is a comparison of the hybrid with one of the original drivers I still have. They have very similar step increases, but the recovery and settling are quite different. The Morel has a smoother and quicker settling, although the original Dayton isn't bad.
The next two graphs show the step response and the impulse response for the hybrid without, then with, the lamb's wool. The truncated peak of the step may be ignored, I think, because I believe this to be simply a matter of the sampling time not catching the peak. Note how much smoother the response is for the wool case. This certainly shows improvement.
Finally we have the impulse responses, from which all else (SPL, Step, CSD, etc.) is calculated. First is the full response for one of the two hybrids, shown below. However, the more interesting graph is below this one.
A comparison of the hybrid without and with wool in the pole piece vent is especially interesting. The case with wool is better in every respect, from what I can tell. The initial transient is steeper with a less noticeable bend midway, which should indicate (I believe) better high frequency response. The return is quicker, the return overshoot is lower and the settling is sooner, all of which indicate less stored energy (meaning less ringing, as also evidenced in the CSD plots).
The "stepping" in the settling is, I think, related to the "ringing" seen in the CSD plots. The rate at which it settles to the zero line indicates the low frequency response, but at this time I'm uncertain of all of this. From what I've seen the better drivers have fewer "kinks" in the step.
In the end, this hybrid shows two things. First, that some tweeters seem to benefit more than others from either pole piece stuffing or a chamber. Second, the dome/voice coil assembly has the most influence on the small signal characteristics. The dome/doping material quality may be the most important aspect of a tweeter.
