This Frequency Generator Knows How To Get Down

What kind of clever things could you do with a signal that had a period of 2 hours? Or 20? Any ideas? No seriously, tell us. Because [Joseph Eoff] has come up with a way to produce incredibly low frequency signals that stretch out for hours, and we’d love to figure out what we can do with it.

To be fair, it’s not like [Joseph] has any ideas either. He thought it would be an interesting project, and figures now that he has the technology, maybe some application will come to him. They say that if you’ve got a hammer everything looks like a nail, so maybe the next project he sends our way will be a sinusoidal fish feeder.

[Joseph] says doing the software side of things with Pure Data wasn’t a problem, but getting it out of the computer proved to be tricky. It turns out that your average computer sound card isn’t equipped to handle frequencies down into the millihertz range (big surprise), so they need to be coaxed out with some extra hardware. Using a simple circuit not unlike an AM demodulator, he’s able to extract the low-frequency signal from a 16 kHz carrier.

So if you ever find yourself in need of a handful of hertz, now you’ve got the tool to generate them. At least it’s more practical than how they used to generate low frequency signals back in the 1900s.

19 thoughts on “This Frequency Generator Knows How To Get Down

    1. If you’re going to that much trouble to clean up the signal, it would be easier to start with a good DAC and a DC amplifier and then generate the low frequency signal directly.

    2. It looks just like what it says. It generates sine waves. The demodulator works just like the detector in an AM radio.

      There’s an image of a 24 second sine wave in the project pictures. That’s 42 millihertz, and it is a sine wave.

      1. I mean that with the extreme wavelength vs. bit depth of the DAC it *should* look like the signal is aliasing. When the DAC changes one bit the output changes far faster than the next bit shows up. I think your scope is showing what I mean as jagged edges on top of the sine.

        Come to think of it, you could generate pulse width modulation on top of the DAC if your soundcard has enough bandwidth. No extra circuits needed.

        Of course this may be completely redundant depending on your application. With no particular goal in mind we’re free to geek out. :)

  1. Plot it on a clockwork paper drum (like ones used by meteorologists in early 20th century).

    Make a VLF PLL and sync it with e.g. ebb and flow of the tide, to make a reality-tracking diorama of some bay which changes its look dramatically with water level.

    Make pair of the signal generators, 90 degrees of phase apart, and rotate a very slow magnetic rotor (or compass needle) with two electromagnets. Or, shift apparent direction of a diffuse LED ambient light, by shifting angle of its maximum on two vertical LED panel “windows” , to simulate natural daylight in room without windows (e.g. for an indoors sunflower plant)

  2. Photography Intervalometer…. I.e. Time-lapse Photos. Most SLR & Digital SLR will have a simple port which will allow you to focus & close the shutter with an external switch (to prevent moving the camera while taking long exposure shots. Typical applications night be time-lapse photos, long exposures (star trails, night sky, city-scapes, etc).

  3. When I was working in a shop on 2 shifts (one week morning, the next one afternoon), I thought about making an even slower astable multivibrator with 2 weeks period. There could be 2 LEDs showing on which shift I would work that week (maybe not lighting constantly, but switched on by some button to make the batteries last longer). Setting the correct phase of oscillations could be done by controlling voltages on capacitors by a regulated resistor, a switch connecting it to the proper capacitor(s) and a voltage meter.
    Using equations from I calculated that i.e. if capacitors had 1000 uF capacitance and resistors R2 and R3 1 MOhm resistance, half period would be about 693 seconds – that’s about 11,55 minutes. To make it 1 week, capacitance*resistance must be over 840 times greater. I think for such long periods it would be necessary to remember about capacitors self-discharging.

  4. I was surprised, when working on a project where I needed a precise 2-channels symmetrical output (, that each computer handles sound differently : on my thinkpads ( a T500, then a T430), whatever level is sent to the audio board stays here until it’s updated. So the sound board / processor can be use to output a very accurate level.
    The program I wrote was also tested on a small chromebook, on which it was obvious that sound was an 8 bit pwm, relying on the speaker or headphone impedance to smooth out the waveform.

    I was much more surprised to see that my 2010 iMac has a high definition output, but everything staying the same level for more than around 500ms is considered silence, and the output goes back to 0…

  5. Exfiltration of data from a secure facility by modulating the power lines with ELF. Not my idea, in the 60s Russians used this attack against our embassy.

    1. Line out from a PC sound card is only about 1 volt peak to peak. After that passes the diodes in the demodulator, there wouldn’t be anything left of the signal. The audio amplifier has an output of about 8 volts peak to peak. That’s high enough to get a useful signal at the output of the demodulator.

  6. reminds me of a dumb hack i did…i had an r/c car and i wanted to replace its radio with a discard phone, so i controlled its h-bridge chip with a PIC12 and used the audio out on the phone to control the PIC12. a very simple low-rate modem. it just output a 1kHz tone gated by my serial data stream, and demodulated it with a diode, RC, and transistor to drive one of the input pins on the PIC12. the crazy thing is, it worked. :)

  7. LFO for a synthesizer. LFO used to modulated a CV. This would be the ‘first’ idea for anyone who loves Eurorack. Let’s see an article on the company Non-Linear Circuits and their sloth module. A 2 hour sine wave would have marketing value in the synth world.

    1. I really like loooooong modulation signals. I have a DIY module that consists of a disc cut out of an xray mounted on the hour hand shaft of a quartz clock motor. An LED on one side and an LDR on the other with voltage run through it. Voila- 12 hours of non-repeating modulation for ambient patches created by the varying opacity of the xray.

  8. Used almost the same method to create a 24hrs triangle-ish wave to simulate a sensor. I didn’t lowpass the square wave but fed it directly to the sensor input using a diode. It had enough on board filtering for that to work.

  9. VLF and ULF radio signals move through water and earth effectively at much higher frequencies that this could produce. maybe using the signal to modulate RF would yield interesting results. one could be curios to see what kind of signal bounce back could occur, or more interestingly what one would pick up listening at this frequency (listening on a graph that is since its below audible frequency). IDK just my 2 cents.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.