What's the point of having a driver in an open baffle? Some would say that it eliminates any "coloration" caused by cabinet resonances. Others claim that it just sounds more "natural" or "live" to have a driver operating freely in the room. Well, which is it? Well, it depends.
When I did my first dipole design, the Aethers, it was an exercise in simplicity. I used very well-behaved drivers and the least complex crossover I possibly could get away with. This made it easy for me to adjust the sound to find something that I found pleasing. No, it wasn't perfect, but it was fun to listen to and people seemed to like it. It had a very "live" sound that seemed almost to bound around the room. Following that success, I attempted the original Sunflowers, which used much higher-quality drivers and a more complex crossover than the Aethers. It had a similar "live" sound to the Aethers. Again, it wasn't perfect, but it was the best I could do at the time, and again people seemed to like the way they sounded.
So for the next few years, the Sunflowers took their home in my family room while I designed several more conventional speakers: bookshelf 2-ways, floorstanders, etc. While doing those designs, my crossover design skills sharpened. I got better and better at fine-tuning the sound of a speaker through its crossover, until I could consistently focus in on a particular sound that I wanted to hear from my speakers. Eventually, I looked back at my Sunflowers and wondered what I would do differently, with all this new skill and knowledge. What kind of improvements could I make? What would I do differently now than I did 3 years ago when I was still grappling with a lot of new ideas?
I decided to re-apply myself to the Sunflowers; a Sunflower Redux. I kept the drivers and cabinet the same as the original, but utilized my more-developed crossover design skills, and learned some new concepts as well! The result is an extremely high-quality 3-way (in my not-so-humble opinion), with very precise imaging, a large, enveloping soundstage, and a lack of coloration through the midrange.
I originally chose to use the Dayton Reference series woofers in the Sunflowers because I knew they were very high quality--competitive with many Scandinavian offerings at a fraction of the price. The RS225 is a very good 8" woofer, capable of very clean bass, and satisfying bass extension in a modest-sized cabinet. It also has a hefty amount of power handling--allowing peaks up to 103 dB!
The RS125-4 is also a nice little woofer. Even though Parts Express lists it as a 5" woofer, it's really a 4" woofer with a large frame--but that's a good thing. Because the cone is so small, the breakup is very high in frequency and well out of the way of the driver's usable range, and therefore easy to notch out and not worry about.
The tweeter is the Peerless/Vifa/Tymphany XT25, which has fared very well in the high-end commercial market, and for good reason. Its Frequency Response measures very flat and it has very low harmonic distortion. It's a bit quirky to work with, but well worth the effort when you get it right.
In case it's not glaringly obvious, the Sunflowers are not a "full dipole." The woofer is a normal, boxed, monopolar speaker. I did this for practical reasons. Dipole bass requires three things that I did not want to mess with: 1. large amounts of surface area (large, multiple woofers), 2. high excursion (lots of xmax), 3. active electronics and/or large amounts of (wasted) power. So to keep things simple, I did as small of a vented enclosure as I could on the bottom of the cabinet. It is around 42 Liters, and can be tuned a variety of ways. For no-fuss bass, tune low (around 23 Hz), with a 2" dia x 8" vent. For a bit of a subtle boost, tune up a bit (around 28 Hz), with a 2" dia x 5.5" vent. These also sound exceptionally good sealed; the bass still reaches low enough to have a "tactile" feel, but doesn't dig as deep as the vented enclosures. Of course, one could build the cabinet vented, then seal at his own whim by stuffing the port with a sock or foam ball. The woofer cabinet should have some sort of bracing, even if it's just a few strips of wood tying the sides of the cabinet together. The inside of the woofer enclosure is lined with convoluted "eggcrate" foam. Additional stuffing is not needed, but could be experimented with.
The mids operate as dipoles in a U-frame enclosure. The U-frame is required to keep the dipole response aberrations (more on those in a bit) as low as possible. You do not need to line the back of the baffle; though it's advisable to chamfer the back side of the woofer hole to let the driver breathe as much as possible. The tweeter is offset because it provides the least amount of diffraction, and thus the flattest Frequency Response.
As far as the overall look goes, I like to think of the Sunflower cabinet design as "open source." As long as you keep the front baffle 11" wide, the U-frame 4" deep, and retain the driver spacing, you can pretty much go crazy! Nothing changes electrically or acoustically. This link provides my detailed cabinet sketches (beware: chicken-scratch handwriting).
Below are a few concept sketches to get you inspired...
A sketch by [Robert F]
A concept sketch by Dan Neubecker
I also think the look of the old Mirage OB-series dipoles would suit the Sunflowers very well
Although it may not look like it at first glance, the crossover for the Sunflower Redux is almost an entirely different creature than the original. You see, with the original Sunflowers, I had the same mindset I did with the Aethers... basically, let the drivers "play into the room" and sound "live," and my part was simply to coax them into a pleasing sound. However, with this new crossover, my goal was to really figure out what was going on with dipole response and deal with these issues appropriately with the crossover. From there, I would then use my more developed skills of being able to "hone in" on various frequency bands and get the particular sound I wanted out of the speaker. This turned out to be more work than I originally planned for.
When I went back and re-measured the raw drivers to create the new crossover, I noticed that there were some strange issues going on with the midrange dipoles. Most significant was a massive 6+ dB spike in the output of the mids around 375 Hz. What precisely was causing this? After doing serious studying at Siegfried Linkwitz' and John Kreskovsky's websites, I learned that this is called a "dipole peak." The dipole peak is a phenomenon where frequencies which start out as out-of-phase in the back of the baffle wrap around to the front, and by the time they get there are now in-phase, and thus combine to create an increase in output. Of course, this only happens at the frequencies whose 1/2 wavelength equals the distance it takes to wrap around the baffle--in this case, 375 Hz, or 18 inches.
Regardless of whether or not that previous paragraph made sense, something had to be done to address this dipole peak. I tried several different contour and notch filters, however my ears alone were not enough to analyze the effectiveness of these filters. The problem with just using one's ears for such a task is that they are good at "getting used to" response aberrations and masking them. So with each new filter, I'd not only listen, but measure the speaker's anechoic and in-room response, with MLS and RTA Pink Noise, respectively. Finally, I had a filter that really hit the dipole peak dead-on with just the right amount of attenuation (C5 + L3 + R5).
With the dipole peak out of the way, I was surprised to find out how precise and clear the speakers sounded. The imaging was extremely accurate--as good any monopole I'd ever done, and the level of detail that was allowed to play through met or exceeded any of my monopolar designs as well. The only real difference I could tell between these dipoles and my best "boxed" speakers was that the midrange on the Sunflower Redux was very clear. Human voices, especially, sounded like they were really coming from a person and not from a box. I know that sounds incredibly cliche--and I don't mean to put down boxed speakers--but if you want to know the real advantage of a properly-designed open-baffle speaker, then that's it. The rest of the crossover is pretty standard fair for me. At a certain point I had to switch over from my favorite crossover design software, Jeff Bagby's Passive Crossover Designer, to the more complex LspCAD. I still prefer PCD, however in this particular case, it did not allow me the flexibility I needed to re-arrange the order of crossover components to achieve the type of response I was looking for. Like Ron Bugundy says, this project got out of hand faster than I expected. It's the first time I've ever run out of alligator clips!
Modeled Frequency Response. Ignore the response below 250 Hz. The drivers were measured anechoically. The true -3 dB point of the Sunflower Redux is around 34 Hz in a vented cabinet, and around 45 Hz in a sealed cabinet..
The Impedance is safe for pretty much any amp.
Individual Driver Phase alignment
Listening Impressions and Conclusion
This is a real game-changer--for me, anyway. When I started my DIY journey a few years ago, I always wanted to try a dipole design. However, when I looked at possible projects, and I really only saw two options. On one hand, you had Siegfried Linkwitz' Orion and John Kreskovsky's NaO, both of which were very complex electronically, and also very expensive designs (though both are very beautiful to look at, I'll give them that). And on the other hand you had pockets of "experimental designs" on various forums across the internet, mostly using large pro audio woofers on a large baffle. With the exception of those by Martin King, these designs were not well-documented, and hard to pin down because the designers seemed to keep altering and upgrading their designs about once a month.
What I wanted to see was a dipole design that had the elegant, modest look of something like the NaO or Orion, but was also affordable and approachable for an average DIY enthusiast or music lover. With the Sunflower Redux, I think I have really hit that nail on the head. Finally I have a speaker with the benefits of dipole mids, but that really carries "my sound" that I like to hear out of my speaker designs.
I described the sound of the Overnight Sensations as "big." These are "huge." The soundstage is very wide and deep, but the imaging remains very precise. The overall "sound" of the Sunflower Redux is akin to the rest of my designs, which I find very satisfying. I was always just the tiniest bit irked with the original Sunflowers, as they didn't quite fit in with the sound of the rest of my designs.
The Sunflowers as built by Carlton
These are my Sunflowers, built with a combination of reclaimed plywood and Oak boards from a monastery.
These are the Sunflowers as built by Bob M. I love the simplicity of it, and the way it showcases the shallow cabinet.
Here's one for the "oh no he didn't" category. These were built by [Harold]. Look familiar!?
I really like the look of this build by Mike. I think this is what my wife wishes I'd built mine to look like.
About the name
I decided a while ago to name my dipole designs after the demoscene. Since I don't live in Europe, I can't really partake in the scene firsthand. However, the DIY community here in the USA feels much the same to me: very talented people who love art create and share their projects for free, and occasionally meet and compete. Sunflower is the name of a French demo group who only created a handful of demos. But they were so overwhelmingly beautiful and graceful that the demoscene still gives them props, even though the group hasn't produced a demo in a decade, and the members have all gone onto other things in life.
If you'd like to try something new, do yourself a favor and download one of their demos and play it. Because they are several years old (in computer terms, anyway), you can run them on practically any machine. "Tesla" and "Energia" are both very beautiful. Now if you don't think those demos are the greatest demos ever, I will fight you .*
by Paul Carmody | this page was last updated December 26, 2020