This seemingly endless virus lock-down has freed up a lot of time to pursue long-dormant interests. I decided to look into VLF listening, which led to a search for an inexpensive receiver. I came across the NASA VLF-3, among others. After building one prototype from the schematics, I discovered a couple of annoyances. I cannot state with certainty that these issues are as objectionable in the NASA receiver kit, but I suspect that most are inherent in the design.
One of the biggest issues was the noisy performance from the headphone amplifier. The LM386 is notorious for ‘hiss’, especially at low signal levels. I also observed an increase in data channel background noise on a waterfall display when the headphone amplifier was switched on. The level of noise changed depending on the position of the volume control. A few experiments on the bench indicated that this stage could benefit from improved power supply isolation.
I also found a couple of revisions to this design while researching this project. These are incorporated in my ‘fork’ of the design. These include the addition of a connection for a loop antenna and the ability to conveniently power the receiver from an external power source. I also added a jack to access the ‘second’ data channel to make it easier to connect a microphone or time receiver to annotate recorded data.
In addition to the changes described previously, I added a 60 Hz notch filter. I also repurposed the data channel switch to be able to switch the notch filter out if the situation permits or requires it, rather than swapping the right and left channels.
The receiver repackaging necessitated a new layout, with careful attention to the power and ground connections. I added power isolation and improved bypassing throughout. I also added some feedback to reduce the hiss and reduced the gain of the LM386 headphone amplifier, as recommended by my research on the LM386.
Internal electronic noise is an unavoidable byproduct of any electronic device, but it can be minimized by the careful selection of the proper components. At least two, and possible three of the active devices are becoming difficult to purchase from established primary vendors. Though some may be available from surplus and closeout channels, once those are gone, replacement with more current parts is the only option.
I selected a low-noise FET and transistor for the first two stages. Though not likely to impact the overall noise figure significantly, the op-amp is also a low-noise type. In addition, all resistors in the signal path are metal film.
I have just completed the rev 2.3 layout and am sending it to production this week. I anticipate turn-around to be two weeks, and the estimated cost per board to be between $10 and $15 each. I plan on releasing the board design, with schematics and parts lists for individual production.
The board is designed to fit into a Serpac A-27 enclosure, readily available from Mouser or Digikey. All of the other components are available from these two suppliers as well as several other major vendors.