Seismograph

I was wondering if people had ideas to help move the following forward...

If you can mount a massive pendulum from a "rock solid" support (the wall in a poured concrete basement, for instance), you have yourself a seismometer... If the ground moves, the pendulum will appear to, as it will stay still. Inertia, and all that.

If you mount an LED on the bottom of the pendulum, and put some kind of "webcam" under that, looking up, and record what the webcam sees, you will have a record of seismic activity.

I put "webcam" in quotes, because I doubt you'd want, literally, a webcam. But some device with a silicon "eye". Ask the eye which pixel is brightest from time to time, and that's all you need to record... you don't record the whole image again and again!

You'll want to read and re-read the sensor frequently. And record what is seen frequently... although you can probably put something in the software to say "if the LED hasn't moved, don't record the position AGAIN".

Ideas... especially in respect of the sensor? One of Sparkfuns little CCTV units?

How about an optical mouse?

What about bouncing a laser off of it so that the edge of the returned circle of light falls on the center of a photo sensitive component - so that the voltage passed by the component varies with how much of the bounced light covers the component. Like the 'laser listening device' that can be found on the internet?

Why not just use a Geophone?

'cos there's no slot for a SIM card?

The idea should work - its a variation on Foucaults pendulum. Formulas you might need - Pendulum - Wikipedia

You'll want to read and re-read the sensor frequently. And record what is seen frequently... although you can probably put something in the software to say "if the LED hasn't moved, don't record the position AGAIN".

That is known as runlength compression, in fact what you write is {time, value, count} tuples - Time is optional but easy for processing later

Q: what do you want to measure?

  • the fact that there is movement (true - false)
  • the amplitude of the movement (0...n)

Do you live in an EQ zone?

Thanks for the interesting comments... keep 'em coming!

Want to measure: As much as possible... i.e. the direction and amplitude of the pendulum's movement relative to basement.

Live in earthquake zone: They are rare and small, but do occur. Mostly a "can I do it?" exercise.

Geophone: I suspect the pendulum would be more sensitive, and less prone to pick up the effect of footsteps or big vehicles going by the house. Also: Doesn't give information on direction of movement.

Laser/ Amplitude... I guess I'm just a digital junkie.... No problems of compensation for accumulation of dust, etc. Sounds like a good sensing idea to implement alongside the digital "answer", to see how they corespond.

Optical mouse.... interesting idea! Might be hard to position mouse close enough to reliably "read" target on bottom of pendulum... but probably not impossible. Avoids need to feed power across the pendulum's suspension. I wonder what the resolution/ response time trade offs would look like. Main dis-incentive: I don't wan't to deal with talking across the mouse's USB interface.

====
Further new idea, prompted by the laser suggestion....

If two small mirrors were mounted at 90 deg to each other on outside of pendulum, at an to vertical, and each had a vertical beam of light bounce off it, the "spot" thrown off would move up and down the wall it hit. Two simple linear arrays of light detectors (probably something like the sensor in a scanner?) could determine the pendulum's position. Anyone know of the sort of sensor this would entail? One of the beauties of this solution is that by varying the angle the mirror was mounted at, you could change the sensitivity and range of the instrument.

idea:
you could place a horizontal line laser near the pendulum so a shadow is cast on the wall with sensitive detectors. If the laser is placed near the pendulum there is an angular magnification (I don't know how to call it otherwise). A slight movement of the pendulum gives a larger displacement on the wall. It is easy if the pendulum is a cilinder to detect X and Y movements. Then one can calculate in which direction the cilinder moves first, and this might be indicative for the direction of the epicentre.

idea ?
I can imagine a sphere floating in a magnetic field as pendulum. Then 3 dimensions might be possible?

robtillaart:
idea:
you could place a horizontal line laser near the pendulum so a shadow is cast on the wall with sensitive detectors. If the laser is placed near the pendulum there is an angular magnification (I don't know how to call it otherwise). A slight movement of the pendulum gives a larger displacement on the wall. It is easy if the pendulum is a cilinder to detect X and Y movements. Then one can calculate in which direction the cilinder moves first, and this might be indicative for the direction of the epicentre.

idea ?
I can imagine a sphere floating in a magnetic field as pendulum. Then 3 dimensions might be possible?

Rob:

I think that idea might go to he11 in a hand basket if there is too much motion in three dimensions. You can't even be sure that the wall be there to hold the sensors in a large quake... ]:slight_smile:

Imaginative though...

You can't even be sure that the wall be there to hold the sensors in a large quake...

If the quake is that large you don't wait for the Arduino to tell you there's a quake ...

robtillaart:

You can't even be sure that the wall be there to hold the sensors in a large quake...

If the quake is that large you don't wait for the Arduino to tell you there's a quake ...

Good point! INCOMING!!!!

I can imagine a sphere floating in a magnetic field as pendulum. Then 3 dimensions might be possible?

Didn't I see that in a sci-fi movie? Or perhaps on TV ("Alias" or "24"?)

Definitely from sci-fi
a quick google "gadget floating sphere magnetic" gave things like - Hacker Builds Floating Jedi-Training Remote Droid | WIRED

such a thing could become a nice EQ sensor :slight_smile:

Serendipity?

Over at the "Linear CCD" discussion in this forum, they are talking about the sort of linear sensor arrays that my "light beam" approach would need... $16 for something that might work...

http://www.taosinc.com/productdetails.aspx?ID=10

(I just mention it, in case others found the idea of the sensor array attractive, wondered where to get one.)

You can of course get fancier sensor arrays up to your willingness to $pend.

Mueller Device .... But back to a practical level, note that keeping something floating in mid-air involves positional feedback and modulating magnetic fields that render the object relatively insensitive to things like seismic movement.

Can't you create a magnetic bowl in which the ball floats with fixed magnets?

ThinkingOutLoud:

  • Put 6 or more magnets with North up in a circle - maybe with a small angle
  • cover the ball with coinlike magnets with North to the outside. ..

Nice device to detect a movement in the Z direction ==> the Shear waves

Just read an interesting story in the Scientific American April 2011 edition Page 54 - 59

You could try using a 2-axis accelerometer mounted to a pendulum, using very light guage wire to connect it to an arduino so it don't interfere with the movement of the pendulum.

using very light guage wire

Use a lead-acid battery as the pendulum mass, and use Bluetooth.

Great discussion guys.
I live on the outskirts of a recent earthquake (well actually two) in Christchurch, New Zealand
In the second quake we experienced a very large vertical component in the order of 2.2 times gravity.

The local Geonet organisation had multiple sensors around the city, and while they were able to read upwards accelerations of 2.2, the downward readings were less, until they worked out the large mass they fixed the sensor to, was freefalling at 1g.

The type (and in our case) location produces two very different effects.
The 7.1 in September and its aftershocks are more horizontal movement and you can hear them coming.
It is further away, and often the shaking intensifies before dying off.

The 6.3 in February and its aftershocks are more vertical, you hear and feel them at the same time, and they generally are almost like someone crashing into the house, and then the shaking dies off.

After the first quake, I do know of one person using a simple accelerometer to measure the quakes, with very good results.

One of the things affecting the drums is sensitivity. By this any local vibrations will reduce the sensitivity to detect new events. I understand that the Geonet devices reset every 24hrs, and they have trouble with a mining operation on the West Coast.

For those that are interested here is a link to the readings and comparison between quakes.
http://db.nzsee.org.nz:8080/documents/43301/43317/GroundMotionComparison+Mw7.1+vs+Mw6.3+-+VERSION+4

You can also see the sources and intensities here http://quake.crowe.co.nz/QuakeMap/Single/

Good luck
Mark

Here's an easier way of doing this, perhaps. Use any graphics editing software to make a black-to-white rectangular gradient and print it out. Attach the paper to the bottom of a pendulum with a flat bottom. Below the pendulum have a simple photodetector and LED light source. As the pendulum swings the reflected light off the paper will change due to the gradient and you'll see a sinusoidal output from the photodetector. Make two sets of these with the gradients at 90 degrees of each other and you have full x-y information.

A little research into how much ground motion to expect from a typical earthquake will help guide the dimensions of the gradient. Also keep in mind the resonant frequency of the pendulum. Ground oscillations faster than the resonant frequency (above resonance) will be much easier to detect than those below resonance.

Use any graphics editing software to make a black-to-white rectangular gradient and print it out.

If I understand your proposal you could have two colors too, from Red to Green on one axis and from 0-100% saturation on the other, with a color sensor you could detect movements in two (X and Y) directions as every pixel reflects another color. Thereby detecting directional information about the quake. Correct?

with a color sensor you could detect movements in two (X and Y) directions as every pixel reflects another color

In principle that works as well. I'm not familiar with color sensors, but I imagine it has separate output channels for red, green and blue? Then you could take (R+G) as the saturation axis and (R-G) and the hue axis.

One draw back of my design, which has two orthogonal saturation gradients side-by-side on the pendulum, is that pendulum twist (torsion mode) would mimic a 45o swing. Having the patterns co-axial as in the hue-saturation design would be immune to this. One would not expect, however, for an earthquake to induce twisting ground motion - or so I believe.