Noise Toy Build Night Recap

Noise Toy (v2) - Complete

The Noise Toy build night was a success. There were six of us and we ended up with five working devices (one person had to leave early).

The plan turned out to be too involved for a single evening. We got far enough to get everyone making noise, but we did not get the programming headers wired up, and I did not get to the programming presentation.

There’s a write-up on the Crash Space blog along with some pictures, including a 3D shot.

Noise Toy Family

One of the reasons I like the Loud Objects Noise Toy is that it is simple. It’s a little too simple, and has a couple areas it could use some improvements. I modified the design of the original Loud Objects Noise Toy to address them.

The audio channels seemed backwards to me, so I swapped them.

The audio output signal has a couple issues. It is a direct output from two digital outputs of the AVR. Each channel swaps between 0 and 3 volts, which is ridiculously loud. Line level is generally between 1-2 volts peak-to-peak, not 3. And it is all positive voltage, which creates a significant DC offset. DC offset in an audio signal is an issue because it makes inefficient use of bandwidth, and makes half the signal overhead, which adversely affects sound quality when recording. I added a voltage divider to reduce the output level.

I’m working on the design of version 3 now. This will add a high pass filter on the output to correct the DC offset, and a voltage starve control that can be used to simulate a dying battery. The AVR begins to act weird when the supply voltage drops below its rated minimum, and the effect can be interesting.

Post a Comment

Your email is never shared. Required fields are marked *