After about 300-400 hours of modeling since September, 2023, on Fusion 360, here is the first technical blog on the project. The soundbar will be housing 12 drivers in total, separated into the following numbers:
2 mini-subwoofers
6 woofers
4 tweeters, including two up-firing tweeters which allow the soundbar to have spacial audio capabilities.
The 2 mini-subwoofers have a surprisingly flat response, given its size, up until 80 Hz. They are placed facing against each other, as can be seen in the hollowed out middle part of the soundbar. This placement has been a somewhat tried and trued method of reducing vibration caused by the low frequencies since the air pressure caused by the opposing drivers cancel each other out, or "push-pull". In an ideal situation, this should work, though the limitation of the soundbar's size makes me heavily doubt its effectiveness in real life application. This will need to be tested and measured once the prototype is complete.
Two tweeters housed in the center floating module will be targeting the main listener. Since higher frequencies contain higher energy and are more directional, the listener should be constantly on-axis to the drivers in order to achieve an accurate and stable frequency response. I am less worried about woofers since they're far less directional to begin with. I have a few methods in mind on how to achieve that, but will include my final solution down the line once I finish.
In terms of control and display, I am still debating among several designs. I prefer physical buttons to touch control, so perhaps at the end it will come to physical or haptic buttons. Programming haptic or building one from scratch will certainly make things a lot more complicated, but it will be a lot cooler as well.
The master processor will be an Arduino Zero, which is the newer and lesser known brother of Uno. Zero operates at 48MHz, 32-bit while Uno operates at 16MHz, 8bit. It also has a much higher flash and RAM which will be needed for the amount of peripherals I will be adding.
The DSP units include a main Electrosmith Daisy DSP chip (Daisy Seed), which uses an STM32 chip, and 3 Class-D DSP amps from Dayton Audio that will be powering the 12 drivers. The DSP amps will be crucial for flattening the drivers' responses in a sub-optimal enclosure. The Daisy Seed will be used to process whichever way the audio comes in, ie. Bluetooth, HDMI, USB-C.
I am slightly worried about how many times the incoming signal will be processed. Each ADC and DAC process will degrade the audio ever so slighly that it might be audible when it gets to the output. Biggest concern in that is dithering. I do not know if the DAC on the Daisy Seed includes dithering, but chances are it does, as will the DSP amps. Assuming there is already dithering applied to the master track of a song when it goes from 24-bit to 16-bit, then goes through 2 more rounds of dither, the added noise can be high enough when on a high volume. For a prototype, as long as the system can accept a variety of audio inputs and play back at a decent enough quality, I will be happy with it. The long term plan, if I ever were to make it into a product, is to obviously build a PCB that integrates everything so that dithering can be applied at the end.
The tolerance test of some of the 3D printed parts is already underway. For the first prototype I will be printing at 0.2mm-0.3mm/layer height. 0.3mm is a big number, and will sacrifice a great deal of quality and accuracy, though it will be a lot faster which is more important for a short prototype time. layer height will be reduced down to 0.1mm once everything works out, and a new tolerance test will then need to be performed.
Written 1/7/2024 to 1/14/2024
Comments