I have always been fascinated with how seismic detection can see what is under the ground. So much so, that I decided to try it for myself. It works by generating a wave from the surface and measuring the reflections bouncing back from the layers below use something similar to a microphone, called a geophone. I created some amplifiers to amplify the signals from the geophones, and read them with a raspberry PI. I will show how I did all this, and what the results were. Below is a simple functional diagram of how this setup I made will work:

Geophones
It seemed like these are semi hard to find. I found some here, and ordered 4 from BG micro. In order to calculate the distance to the reflection, multiple geophones will be required. Multiple geophones could also be used to cancel out noise, depending on placement and how they are wired up. It looks like there are quite a few sellers of these on alibaba, but I did not try them. There are even fully built seismic detection rigs pre-built from alibaba. If something on one of them broke, I imagine it would be more difficult to troubleshoot in a rig from alibaba.

Amplifier
The cool part! I just used the design from the application notes in this datasheet. The whole design is based around the LT1677 op-amp, which is a very low noise (and expensive) op amp. The good thing about the design, is that it is simple, and only requires one op-amp. The down side, is that the resistance of the wire running to the analog to digital converter affects the reading slightly (just offset a little, depending on lead wire resistance). I used Eagle CAD to layout the PCB, then used the pcb-gcode plugin in order to create the gcode files to mill the circuit boards. Next I milled the boards, using a 0.2mm bit for the traces, and a 1/32” bit for the holes.


I supply the pre-amp board with 12V, and find that they don’t output more than 20mA.

A/D Converter
The only requirements for this are, multiple channels, 0-5V measurement, 300Hz sample rate. Since those are easy to meet, there are tons of devices that would work. What I ended up settling with is an MCP3008 hooked up to a raspberry PI. Which, I have been super happy with. Some places to find out how to wire the mcp3008 into the raspberry pi can be found here. Here is what my setup looks like.

The circuit to wire the amplifiers into the analog digital converter is very simple. It is just converting the current from the amplifier to a voltage, and the MCP3008 measures the voltage. All that is required for this is a resistor, so much easier than converting voltage to current. . I used a 250Ohm resistor, which is a very standard size, because 4-20mA running through it will produce 1-5V. Originally I used the raspberry pi power supply for the MCP 3008, but found I had lots of noise in the signal. To get rid of the noise, I supplied the MCP3008 with an LM7805, which is supplied from the same 12V supply as the geophone pre-amps. I also put everything (raspberry pi, and 12V wall wart power supply), on the same ground. There was still some noise after that, so I added some capacitors in the inputs to the mcp3008, which fixed everything. Although now that there are capacitors, the response from the geophones has been damped a little.

I packaged the power supply, 250Ohm resistors, MCP3008, and raspberry PI up in a cardboard box. It may not be the neatest, but it works. I added electrical tape to the cables coming in from the pre-amps so that they cannot fit through the holes in the box they are in, and strain the connections to the breadboard. For the pre-amps I put those onto pieces of plywood, and nailed the cables onto the wood, so they could not put stress onto the connections to the circuit board. This has so far proved very portable. I am able to take the setup anywhere, lay the geophones out, create a wave, and see the results right away. When I take it out, I power the geophones with a LiPo battery, and the raspberry pi I power off of my laptop.

Recording Device
For the recording device, I chose the raspberry pi for several reasons. It runs linux, and therefore works with almost any library or programming language available. It has alot of processing power, low power consumption, low cost, and is very portable. So far I have been very happy with it.

I used python to create a script which constantly checks the MCP3008 inputs, and compares one of them to a trigger. When the trigger point has been reached, it starts recording a CSV file with a column for time since recording started, and columns for the voltage measurement from the geophones. When generating sound waves myself, I found that I cannot hear/feel reflections greater than 1 second after the wave is generated, so I only record for one second. If I were to use these for earthquake measurement, I would need to record for a much greater time.

Next step is to present the results to me with a nice interface. To do this I created a php script which reads all the generated CSV files, and creates graphs of the contents, as you can see below. When I go out and use them, to get started all that needs to be done is to log into the pi with ssh, and run the script which scans the analog inputs. Then create the source wave near the trigger geophone, and refresh the web browser to see the results.

That’s about it, if you have any questions let me know. I’ll be coming back and updating this with the gcode files to mill the boards, once I find them. I’ll make my next blog post about geophone placement, post-processing of the data, and source generation :).