The Open Source Open Baffle Speaker Project
I have for a few years now been fascinated by open baffle loudspeakers, and also fascinated more recently by full range drivers (especially field coils like the Voxativ AC-X).
I want to share my experience in building this project in hopes that someone else will use the design, improve upon it and share it here so others can benefit. This is the perfect speaker project for the curious DIY'er, and the tweaker. One can change drivers easily, and one can experiment with all kinds of speaker, wire, and vibration damping tweaks without too much difficulty.
Open Baffle Designs
Open baffle speaker designs have been gaining in popularity recently due to the successful designs and efforts of small audio manufacturers like The Pure Audio Project and Spatial Audio. They have crafted some excellent sounding and good looking products that, in audiophile terms, are quite affordable.
Above: The Pure Audio Project Trio 15.
Note the separate baffles for each driver.
Open baffle loudspeaker designs are very attractive to the DIY community as well. Mostly due to the fact that one does not have to build an elaborate box, or a box at all to enclose their speakers. Just a rectangular piece of wood large enough to house the drivers, and of course drivers well suited to open baffle designs. This makes building an open baffle speaker much easier and less expensive than conventional speaker designs.
I particularly like the open and boxless sound of the open baffle speaker. Since there is no enclosure containing the back wave of the drivers, the speaker radiates out the front and back. This also helps create a large soundstage. For the person looking to build an inexpensive, easy to assemble speaker, open baffle designs offer great performance and can be constructed without much knowledge of woodworking and carpentry.
Criterea for the design
For my open baffle project, I wanted the following criteria
- Modularity - I want to be able to test different drivers easily. This means putting the drivers on their own separate baffles, and being able to configure them in any way. Not only did the baffles have to be modular or swapable. The frame supporting the baffle had to be adjustable to accommodate the different baffle sizes and placements.
- High efficiency - I like high efficiency loudspeakers. It's strictly a personal preference, and it allows more flexibility in amplification.
- Full range main driver - I also like full range drivers because of their phase coherance and simplicity. The full range driver will of course be augmented by drivers covering the lower octaves.
- Active Design - Everything in speaker design involves a compromise. When designing a passive loudspeaker, much time and expense can be spent on the crossover and getting everything to sound just right. Not only that, inserting extra electronic components in between the amp and speaker creates all sorts of sonic issues and loss of signal. I chose the active route, using a MiniDSP crossover/processor box and my laptop because it would allow me to quickly adjust crossover points, slopes and correct for phase issues as well as room issues. Of course there are compromises when going this route as well. I am using the device as the DAC, which precludes me from using a more premium DAC, and there are the issues associated with active crossovers such as phase coherance. In using the full range driver, I hope to avoid most of these issues by only using an active crossover on the low frequency drivers, allowing the full range driver to run with no crossover. The tradeoff here will be output. I won't be able to play this configuration very loud due to the full range driver operating in free air, with no box to mitigate extreme cone excursion. But, if I do want to turn up the volume I can have a second crossover programmed in to the MiniDSP that would cut off the low frequencies to the full range driver for listening at higher levels.
I wanted the frame for my open baffle speaker to accommodate a large variety of drivers, and be wide enough to sound good. After much research, I chose 1.5" extruded aluminum tubing because of the stiffness of the metal, and the flexibility in design options I had with this type of tubing. for the main "mast" of the speakers where the baffles will be mounted, I chose a height of 48" and a width between masts of 19", with 17.5" being the distance between bolt holes for the baffles. The baffles will be 22" wide and will overlap the masts so they do not show when viewed from the front. The width will be able to accommodate 15" drivers easily. There are many sources for this tubing, and the parts to connect them together. I chose 80/20 on the recommendation of a good friend who knows how to build stuff. Below is my parts list.
- 4 pieces 1503 LS, 48" long - These are the "masts" where the speaker baffles will attach. I chose 48 inches so that I could accomodate 2 15" driver baffles and up to a 10" full range driver baffle. I needed three slots on these so I could fasten the baffles, cross braces, and angle bracket to the base.
- 4 pieces Part 2534 - These are 45 degree cut braces, 12" long that I used to brace the baffle portion of the frame to the base.
- 4 pieces 1503 LS, 18" long - I used these for the bases, running from front to back. I needed three slots on these so I could fasten the top brackets to the part of the frame holding the baffle, a slot on the inside to fasten the cross braces, and a slot on the bottom to fasten the rubber feet. There is no slot on the portion facing outward, creating a smoother look.
- 8 pieces 1501 LS 16" for the cross braces on the base and baffle frames. I only needed one channel for attaching the corner braces to the cross brace and main pillars of the bases and baffle frames.
- 40 pieces Part 13024 - T Nuts Used to fasten the tubes together. One still needs to buy screws. For that I chose McMaster Carr's Black-Oxide Alloy Steel Socket Head Screw, M6 x 1 mm Thread, 16 mm Long.
- 20 pieces 4302 corner bracket - 8 used to fasten cross braces, 2 used opposite the 45 degree angle to provide additional support for the baffle portion of the frame.
- 8 pieces 2849 rubber feet - I liked these because they attach to the bottom channel of the base easily and provide a grippy surface for wood floors. I am sure I will change these to something more audiophile certified if it makes a difference.
- 12 pieces 3267 to fasten end caps onto the end of the exposed bars. Purely cosmetic.
- 12 pieces 2030 plain - End caps for exposed ends. Again, purely cosmetic.
Below is one of the assembled frames. The 48" masts, joined together by the 16" cross braces, join to the 18" rails that run front to back. There are 2 additional 16" cross braces joining the rails of the base. The rails have the 12" 45 degree support on the back side of the mast, as well as a corner brace stabilizing motion on the front of the mast.
Things to consider on the frames
These are the items I will continue experiment with.
- Mass loading - the frames by themselves are heavy, but not terribly heavy. Once the speakers are assembled, I will be working on tweaking out the design and seeing if adding mass benefits the sound.
- Vibration damping - The frames as they are now ring a little. I am guessing that attaching the front baffles will mitigate that, but I will investigate other vibration damping options to see if they benefit the sound.
- Attaching the baffles - I will start by simply bolting the baffles to the frame. I would think that adding some sort of vibration damping gasket material between the baffle and frame would help to reduce the transmission of vibrations from the driver to the frame.
I like paper cones, as they have good self damping properties. I like large, efficient drivers as well. They can move larger amounts of air with less effort (and hopefully distortion) than smaller, less efficient drivers that need high excursion to move air.
Eminence Alpha 15, 15" Open Baffle Driver
I chose 4 Eminence Alpha 15 drivers for the low frequency portion of the speakers. Two per speaker. They are well known to be great drivers for open baffle designs, and at 97 dB efficiency, are plenty efficient for this design. For the first full range driver, I chose the Dayton Audio PS 220 - 8 8" full range driver. It is rated at 95 dB efficiency. My goal is to try several full range drivers in this design, and the modular nature of the design will make doing this easy. This will be the first of many.
Dayton audio 8" full range driver.
The frames I designed will allow me to experiment with driver positioning. The Pure Audio Project sandwiches their main drivers between the bass drivers (MTM design), which I can do also. But unlike Pure Audio I can also experiment with placing the main driver above both woofers (MMT design).
I will be using the MiniDSP DDRC-24 as the crossover, DAC and DSP for now. I have tweaked the MiniDSP out by using a third party linear power supply over the supplied wall wart. I am also using an ifi iUSB to help clean up the digital output from my laptop. If you wanted to get nuts, you could purchase a linear power supply for the ifi unit as well. These two add-ons to the MiniDSP make a big difference in the sound quality when listening to music. I highly recommend them.
I will also be using my laptop and Room EQ Wizard for measurement duty. I cannot use a streamer at this point because I need to take measurements, and that involves plugging in a microphone.
The microphone used for measurements is the UMIK-1, which can be purchased with the MiniDSP DDRC-24. These tools will allow me to optimize a crossover, correct for any frequency response issues in the drivers or room. DIRAC Live Room correction software is bundled with the MiniDSP.
Purists may scoff at the idea of using room correction on a full range driver. Like I said, everything is a compromise. Using the room correction will correct for any frequency response or phase issues, but may harm ultimate resolution. Running the driver wide open with no dsp may provide a higher degree of resolution, but may have tonal anomolies that I may find unacceptable. I don't know, but I am going to remain open minded to all options. That's what this is about.
Audiophiles may scoff at the fact I am not using a high resolution DAC for final processing to Analog. The DDRC-24 does not have digital outputs, only analog. MiniDSP does offer other processors that sport digital outputs. That would allow one to take the digital output from the processor and port it to an audiophile DAC or two (4 channels necessary for a 2 way speaker). One could use the inboard DAC on 2 channels for the bass drivers, and an audiophile DAC on the full range driver. But for now I will use the analog outputs from the DDRC-24 to keep costs down and because it's what I have right now.
So it begins. The Open Source Open Baffle speaker project. I hope you enjoy reading about this project, but more importantly, I hope you build it and can contribute to it's evolution. Please leave a comment if you have questions, and if you have an idea that might improve upon this particular design, please feel free to share.
Next: Attaching the drivers to the baffles, and first listening impressions.