When last we left our intrepid adventurer, a bunch of orcs had attacked Middle Earth with a virus causing Bjorn to flee to his bathroom to make speakers. Much like this opening, that effort really didn't turn out well. Here's the original thread, including some of the discussion of the Helios design itself.
Also, I just saw that Meniscus has the kit (and the write-up/plans) on its website. It looks like the final published dimensions are closer to my approximations, below.
To re-cap that thread, I made the first cabinet in the bathroom of my apartment, and it turned out okay, but it was much larger (deeper) in real life than I wanted. I did try to make a second cabinet, but I could not repeat the success that I had with cabinet Home So, I ended up converting the second mk1 cabinet into a dual-opposed subwoofer box, because that's what was feasible by the time I cleared out The Yips.
The mark 2 version is a wide-but-shallow small tower, with approximately 38 liters of net volume, and exterior dimensions of 11.5w x 37.5h x 9d. As-designed, Helios specifies 35L of internal volume, but that is before the substantial amount of internal bracing which is at least another liter or two of volume (I'd guess two...). I'm sticking with the ScanSpeak 26W passive radiator because it has improved low bass output (in simulations) vs the as-designed SBA passive radiator, and it does so with a smaller cabinet foot print (27cm vs 29cm). Using a different PR is really what allows this tower to work without resorting to some other extreme option. Finally, the cabinet construction is all carbonized vertical bamboo, and it is a mix of both single ply and three ply versions. Both have a nominal thickness of 19mm, or 0.748". I primarily used Northwest Bamboo to source my raw material.

Before I dive into either the design modifications or the build discussion, here's the current progress as photographed from the workshop of the San Diego Fine Woodworker's Association. I'm incredibly pleased that I have a resource like this available to me. Highly recommended for those within striking distance of Miramar and the 805.

In making the shift from a monitor to a tower, I'm making changes to two design elements. The first is obvious because it's one of the biggest rules among the DIY speaker community: don't change the baffle! The second issue is one that will be less obvious to people, and it can matter a fair bit, as well: changing the depth. Changing the baffle dimensions and layout affects the diffraction signature of the design. In practical terms, the change in the baffle will alter both the frequencies and amplitudes of the baffle-induced constructive and destructive interference. More on this in a moment....
The change to the depth of the cabinet changes the frequency and amplitude of the effects of the rear wave of the drivers. Since (mid)woofers are not acoustically opaque front-to-rear, the rear wave can have an audible effect on the overall speaker performance. While I do not have any simulation tools at-hand to quantify this, based upon how the discussion went a long time ago when the Grimm LS-1 was launched into the market, I would expect to see higher amplitude rear wave effects (less absorption) with a shallower cabinet, but with those rear wave issues happening at a higher frequency. In this case, the half-wavelength should move from about 566Hz based upon the original design to closer to 1kHz in my mk2 tower. With a Fc around 1300 Hz, the move up to 1kHz may not create too much of a penalty.
One will also note that the original design did not include any stuffing or surface treatment of the rear wall of the cabinet. Thus, the original design took a "let the chips fall where they may" attitude towards this rear wave issue. For the mk2, I have Sonic Barrier in inventory, and based upon Augerpro's findings in the construction methods thread, I'm likely to invest in some denim insulation, as well. Lacking a reliable modeling tool, I'll skip further commentary on this second issue until I've made more progress on the build.
Now, let's talk diffraction...

Using Jeff's Diffraction & Boundary Simulator (because it's fitting that I use his tools, plus I'm an Excel junkie), I put together the chart, above, showing five different scenarios overlaid on one another.
Based upon the simulations, above, I would expect to see about 1dB of more extreme (positive and negative) swings in diffraction between about 500Hz and 2kHz. The major deviations are between 500 - 900 HZ (excess SPL), but this is more than a half octave below the 1.3kHz Fc - so it will be -12dB or so below the design level. Between this and the directional nature of the waveguide's output, I'm optimistic that it might not be audible. If I had kept the original spacing, then the deviation would have been less than half of the as-built simulation. Also, as mentioned above, a waveguide increases the directivity of the tweeter, which should reduce the interaction with the baffle even further than what should be predicted by the increased effective radiating area alone (I'm guessing here).
All of that is great in theory, but I don't know that I could hear it and having such a big tweeter so close to the top of the cabinet just looked a bit odd to me. Aesthetics won out on this decision, and if I need to "fix it in post", then that's what I'll have to do.
Net-net: changing the baffle width had a trivial effect. However, changing the baffle height and the placement of the tweeter does seem like it will have a measurable effect.
Also, I just saw that Meniscus has the kit (and the write-up/plans) on its website. It looks like the final published dimensions are closer to my approximations, below.
To re-cap that thread, I made the first cabinet in the bathroom of my apartment, and it turned out okay, but it was much larger (deeper) in real life than I wanted. I did try to make a second cabinet, but I could not repeat the success that I had with cabinet Home So, I ended up converting the second mk1 cabinet into a dual-opposed subwoofer box, because that's what was feasible by the time I cleared out The Yips.
The mark 2 version is a wide-but-shallow small tower, with approximately 38 liters of net volume, and exterior dimensions of 11.5w x 37.5h x 9d. As-designed, Helios specifies 35L of internal volume, but that is before the substantial amount of internal bracing which is at least another liter or two of volume (I'd guess two...). I'm sticking with the ScanSpeak 26W passive radiator because it has improved low bass output (in simulations) vs the as-designed SBA passive radiator, and it does so with a smaller cabinet foot print (27cm vs 29cm). Using a different PR is really what allows this tower to work without resorting to some other extreme option. Finally, the cabinet construction is all carbonized vertical bamboo, and it is a mix of both single ply and three ply versions. Both have a nominal thickness of 19mm, or 0.748". I primarily used Northwest Bamboo to source my raw material.
Before I dive into either the design modifications or the build discussion, here's the current progress as photographed from the workshop of the San Diego Fine Woodworker's Association. I'm incredibly pleased that I have a resource like this available to me. Highly recommended for those within striking distance of Miramar and the 805.
In making the shift from a monitor to a tower, I'm making changes to two design elements. The first is obvious because it's one of the biggest rules among the DIY speaker community: don't change the baffle! The second issue is one that will be less obvious to people, and it can matter a fair bit, as well: changing the depth. Changing the baffle dimensions and layout affects the diffraction signature of the design. In practical terms, the change in the baffle will alter both the frequencies and amplitudes of the baffle-induced constructive and destructive interference. More on this in a moment....
The change to the depth of the cabinet changes the frequency and amplitude of the effects of the rear wave of the drivers. Since (mid)woofers are not acoustically opaque front-to-rear, the rear wave can have an audible effect on the overall speaker performance. While I do not have any simulation tools at-hand to quantify this, based upon how the discussion went a long time ago when the Grimm LS-1 was launched into the market, I would expect to see higher amplitude rear wave effects (less absorption) with a shallower cabinet, but with those rear wave issues happening at a higher frequency. In this case, the half-wavelength should move from about 566Hz based upon the original design to closer to 1kHz in my mk2 tower. With a Fc around 1300 Hz, the move up to 1kHz may not create too much of a penalty.
One will also note that the original design did not include any stuffing or surface treatment of the rear wall of the cabinet. Thus, the original design took a "let the chips fall where they may" attitude towards this rear wave issue. For the mk2, I have Sonic Barrier in inventory, and based upon Augerpro's findings in the construction methods thread, I'm likely to invest in some denim insulation, as well. Lacking a reliable modeling tool, I'll skip further commentary on this second issue until I've made more progress on the build.
Now, let's talk diffraction...
Using Jeff's Diffraction & Boundary Simulator (because it's fitting that I use his tools, plus I'm an Excel junkie), I put together the chart, above, showing five different scenarios overlaid on one another.
- Scenario 1 is the predicted diffraction of a 29mm dome tweeter placed 4.375" below the centerline of the speaker (edit: apparently 4.44 is the official number, which is 4 7/16 rounded to two decimal places). This is my guess about where the tweeter centerpoint is located in the original design. It's pretty ugly.
- Scenario 2 is the diffraction effect changing the 29mm dome to a 15cm driver (close to the usable diameter of the wave guide). This is much better, obviously.
- Scenario 3 expands the baffle from 11.1" to 11.5". There is almost no difference compared to the original design.
- Scenario 4 models the as-built cabinet, with the tweeter being 5.75" below the top of the cabinet. Ouch. That may be an expensive aesthetic choice.
- Scenario 5 maintains the 11.5x37.5 tower baffle, but moves the tweeter back to being just 4.375" below the top of the cabinet.
Based upon the simulations, above, I would expect to see about 1dB of more extreme (positive and negative) swings in diffraction between about 500Hz and 2kHz. The major deviations are between 500 - 900 HZ (excess SPL), but this is more than a half octave below the 1.3kHz Fc - so it will be -12dB or so below the design level. Between this and the directional nature of the waveguide's output, I'm optimistic that it might not be audible. If I had kept the original spacing, then the deviation would have been less than half of the as-built simulation. Also, as mentioned above, a waveguide increases the directivity of the tweeter, which should reduce the interaction with the baffle even further than what should be predicted by the increased effective radiating area alone (I'm guessing here).
All of that is great in theory, but I don't know that I could hear it and having such a big tweeter so close to the top of the cabinet just looked a bit odd to me. Aesthetics won out on this decision, and if I need to "fix it in post", then that's what I'll have to do.
Net-net: changing the baffle width had a trivial effect. However, changing the baffle height and the placement of the tweeter does seem like it will have a measurable effect.
Comment