With my recent gamut of driver testing, I was quite impressed with what AliExpress brand Melodavid had to offer at a competitive price. Apart from the great out of the box performance, the Melodavid tweeters have a nice flat mouting mounting surface, so removing the face plate and mounting them to a waveguide is simple. Or so it appeared...
I started with the budget Melodavid CB25. A copper beryllium allow dome with large half roll surround. Total diameter is 38mm, so it should fit well directly mated to a Viston WG148, however all of mine were in use so I don't have that data to share. I attempted a waveguide model for CNC milling, using 3/4" material. The waveguide was about 165mm wide, and made elliptical with the height 2/3 of the width. The throat was made to 38mm so it would end right at the edge of the tweeter surround. It looked like this.
(sorry, the photo didn't capture the waveguide curvature very well, so you may have to use your imagination here)
Unfortunately I didn't find great success here. With as simple on-axis measurement, the waveguide provided a nice boost, however with a large cavity in the 11kHz range. This image compares the original response of the CB25 to it's waveguide loaded response.
I didn't bother to continue with a full spatial response set, since I had already declared this a failure, but the dip does go away about 30 deg off axis.
I tested as well with Melodavid CM25Nd, which has a different dome material, but dome shape and the surround are the same. I didn't bother to save the result, it was just the same as the CB25 above, so this problem is definitely a geometry issue.
But what about Melodavid Be28. It is also 38mm total width, with larger 28mm dome, and the half roll surround a bit smaller as a result. It looks like this:
To my surprise, Be28 did not experience the high frequency cavity at all, so I continued with 90 degrees of spatial data. This tweeter performs very well in this configuration. This is not measured on infinite baffle, but the same bookshelf cabinet I used for all my other speaker testing.
So what can be done about the CB25 and CM25? Is there any hope for these? Quite possibly. One option could be to try a diffuser disc in front of the dome. I also studied a commercial option, the Wavecor TW030WA11. What I found interesting about this tweeter is that the waveguide throat did not end at the edge of the dome surround, but rather at the peak. It includes a quarter round on the back side, which I'm sure is not there by accident.
For an easy test, I designed a small shim that could be 3D printed, then the waveguide can still be easily CNC milled, with a bit smaller throat, in this case I believe it was 31.5mm instead of 38mm.
The final waveguide, a tight fit, but nothing is touching the dome or surround, so it's a success.
The change was a success. No longer was there a big cavity, however the top end did have a fair bit of a peak. This is much less of a problem, as it is pushed up further towards the inaudible range, and can be more easily tamed with a crossover filter.
I can imagine CM25Nd will be no different. A bit more complicated to design, but not much more for those who are 3D printing the entire waveguide anyway.
I hope you find the information in this story helpful.
I started with the budget Melodavid CB25. A copper beryllium allow dome with large half roll surround. Total diameter is 38mm, so it should fit well directly mated to a Viston WG148, however all of mine were in use so I don't have that data to share. I attempted a waveguide model for CNC milling, using 3/4" material. The waveguide was about 165mm wide, and made elliptical with the height 2/3 of the width. The throat was made to 38mm so it would end right at the edge of the tweeter surround. It looked like this.
(sorry, the photo didn't capture the waveguide curvature very well, so you may have to use your imagination here)
Unfortunately I didn't find great success here. With as simple on-axis measurement, the waveguide provided a nice boost, however with a large cavity in the 11kHz range. This image compares the original response of the CB25 to it's waveguide loaded response.
I didn't bother to continue with a full spatial response set, since I had already declared this a failure, but the dip does go away about 30 deg off axis.
I tested as well with Melodavid CM25Nd, which has a different dome material, but dome shape and the surround are the same. I didn't bother to save the result, it was just the same as the CB25 above, so this problem is definitely a geometry issue.
But what about Melodavid Be28. It is also 38mm total width, with larger 28mm dome, and the half roll surround a bit smaller as a result. It looks like this:
To my surprise, Be28 did not experience the high frequency cavity at all, so I continued with 90 degrees of spatial data. This tweeter performs very well in this configuration. This is not measured on infinite baffle, but the same bookshelf cabinet I used for all my other speaker testing.
So what can be done about the CB25 and CM25? Is there any hope for these? Quite possibly. One option could be to try a diffuser disc in front of the dome. I also studied a commercial option, the Wavecor TW030WA11. What I found interesting about this tweeter is that the waveguide throat did not end at the edge of the dome surround, but rather at the peak. It includes a quarter round on the back side, which I'm sure is not there by accident.
For an easy test, I designed a small shim that could be 3D printed, then the waveguide can still be easily CNC milled, with a bit smaller throat, in this case I believe it was 31.5mm instead of 38mm.
The final waveguide, a tight fit, but nothing is touching the dome or surround, so it's a success.
The change was a success. No longer was there a big cavity, however the top end did have a fair bit of a peak. This is much less of a problem, as it is pushed up further towards the inaudible range, and can be more easily tamed with a crossover filter.
I can imagine CM25Nd will be no different. A bit more complicated to design, but not much more for those who are 3D printing the entire waveguide anyway.
I hope you find the information in this story helpful.
