My Online Seismograph: View Near Real Time Seismic Data From Sydney, Australia

[davenn]

Hi to all

A couple of weekends ago I finallly achieved what I wanted to do for the last 10 years
That was to get my digital seismo online so people ( and I) can view it remotely in near real time

http://www.sydneystormcity.com/quakedrum.GIF

currently there is just a single channel showing
1) Each horizontal trace line is 30 minutes long and there is ~ 36 hours of recording per screen.
2) Timing is kept relatively accurate by using an online time server that takes into account internet delays.
3) The display is updated every 5 minutes, do a refresh of your browser st see update
4) The seismometer is a long period, 11 second "Lehman" style unit I built back in June 2011 seen here on my www site scroll down ~ 1/2 way to see an overall view of the unit

The sensor is currently located on the concrete floor of our laundry. This means that people moving near the unit cause quite large swings of the sensor pendulum, apart from the vibrations the washing machine cause ( most of this unwanted noise will be seen during my local friday nite and saturday mornings)

I am always open to others asking for help setting up a seismometer
so don't be shy in asking questions :)

I Also have a friend ~ 50 km west of my home ( I am in inner west Sydney, Australia)
Rob's system went online about the same time, as we were doing a combined effort in finding a decent program that will do auto screen grabs at a determined interval and ftp them to our respective www sites.

Located here is Rob's display

 

[SHISHKABOB]

why does it look "thicker" in the more recent parts?

 

[davenn]

there's some quirk with the software and the way it displays the image
I haven't got around to asking the programmer about it yet.

Am still at work at the time of writing this, all I need to is, in the program, click on redisplay, and the wriggles will be nice and thin again.
I will be able to do that in a couple of hrs when I get home, and you will see the difference.

The guy who did all this is in Redwood City, S of San Francisco, USA
A seriously bright cookie :)
He designed the multi-channel preamp circuits, designed the A to D hardware
and wrote the data logging and analysis packages
I have been using the various versions of his gear and software since the min 1990's
prior to that I was just using the classic drum recorder, with ink pen on paper

Dave

 

[dlgoff]

A couple of weekends ago I finallly achieved what I wanted to do for the last 10 years
That was to get my digital seismo online so people ( and I) can view it remotely in near real time...

Very nice Dave. I'm curious as to how your geophone operates.

Regards

 

[davenn]

Greetings Don

Thankyou :)

Thats a 3 axis set of geophones, they came out of the seismic surveying industry
There are 2 horizontal and 1 vertical unit ( the vertical one is at the top of the pic
they are aligned E = East/West, N = North/South and Z = Vertical. They are bolted to the lid of the casing and the whole lot is waterproofed and buried.

The geophone is basically a magnetic mass suspended on springs within a coil of wire. Resistance of the coil is ~ 2000 Ohms and the undampened natural period of these geophones are 4.5Hz.

They are great for recording local small to M4 events up to a few 100 km's, will easily record the larger events up to 1000km, but what happens is the further out they are the attenuation of the P and S waves starts becoming very apparent ... somethimes from the large distant events all that will be seen is a few cycles of P waves and that's it, no S or Surface waves

Dave

 

[dlgoff]

Thanks Dave. I wasn't sure if maybe MEMS weren't being used somehow.

What a fascinating undertaking. I sometimes check out the USGS earthquake page when there are tremors here in the US midwest.

Regards

 

[billiards]

Great project Dave. Have you ever done anything with your data?

 

[davenn]

Great project Dave. Have you ever done anything with your data?

Not from giving it to a seismo institution point of view, no
none of them are interested unless you can prove you have a very stable system and accurate time recording down to ~ 0.2 of a second.

There is a large collection of us amateurs worldwide, that have a wide variety of commercial and home brew seismometer sensors. Some of the guys have got incredibly creative when experimenting with and designing sensors.
We do have a archive that we can upload our quake data to

Have a look at http://psn.quake.net/
This is one of and one of the earlier Public Seismic Nets that were set up anywhere in the world. Started in the San Francisco Bay Area and within a few years quickly spread worldwide.

As far as recording goes, for me, its the thrill of recording those BIG events. doing work to improve the system, playing with different sensors. My 10 second period Lehman type sensor should really be capable of 15 - 20 seconds. Its been a battle of wills between the seismometer and myself trying to figure out why I can't get it past ~ 10 - 11 seconds.
I have another one I need to build to orientate perpendicular to the current one. There are lessons learned from the construction of the first one that should improve the operation of the second unit. The sensor also needs to be relocated out of the laundry to a shed at the back of the property that is much quieter as far as manmade noise goes

Dave

 

[marcusl]

What is a Lehman seismometer, and what does the period refer to?

You could perform a running cross-correlation of data from your seismometer and your friend's 50 km away. This should greatly reduce local noise (trucks driving by, e.g.). Only correlated signals from events far compared to 50 km would remain, so sensitivity to those 1000+ km events should increase. The signals must be sampled at the same rate, and the time lag between them adjusted for the difference in arrival times at the two stations for the event considered.

 

[davenn]

What is a Lehman seismometer, and what does the period refer to?

Hi Marcusl

The Lehman seismometer is named after and popularised by James Lehman, with an article he did in the July 1979 Scientific American magazine. Right under the heading of that link there's a figure1 you can click and see his original design.
It is based on what is called the garden gate pendulum, where the top and lower pivot points are offset from each other. This original design has evolved immensly over the years since 1979 with many many guys experimenting with different variations. Most of the experimenting has been around the 2 pivot points and particularly the bottom one trying to make it as frictionless as possible. Some of the guys have done beautiful engineering jobs and their units are truly a work of art.

http://jclahr.com/science/psn/chapman/2008%20lehman/lehman_prototype.jpg
https://sites.google.com/site/bobmcclure90/horiz

Mine is definitely NOT a work of art but is it well functional. Version 2 will look much nicer :)

The seismometer is a pendulum, so therefore it has a natural period of oscillation, my one is ~ 10 seconds ... that's 10 sec per cycle! ... 0.1 Hz. Dampening lowers that period a little more, but as I said in my earlier post I am still struggling to get the 15 - 20 second period that it should be capable of.
The sensor MUST be dampened, else when the seismic waves pass by, the seismometer will just start oscillating at its natural period and continue till it finally dies out. We don't want that, we want the seismometer to reproduce the ground motion only
if you have a look at the 2 seismograms near the top of my page, you can see the difference in response between an underdampened and critically dampened seismometer.

You could perform a running cross-correlation of data from your seismometer and your friend's 50 km away. This should greatly reduce local noise (trucks driving by, e.g.). Only correlated signals from events far compared to 50 km would remain, so sensitivity to those 1000+ km events should increase. The signals must be sampled at the same rate, and the time lag between them adjusted for the difference in arrival times at the two stations for the event considered.

would be interesting to combine the data and use that to filter out the crud. It would probably require the 2 seismo's to be the same style and operating parameters.
Rob's seismo is a vertical style that operates ~ 1 second period and the responses of his and my units is very different.

Dave

 

[AlephZero]

The sensor MUST be dampened, else when the seismic waves pass by, the seismometer will just start oscillating at its natural period and continue till it finally dies out. We don't want that, we want the seismometer to reproduce the ground motion only
if you have a look at the 2 seismograms near the top of my page, you can see the difference in response between an underdampened and critically dampened seismometer.

If you are gathering the data digitally you should be able to get rid of the dynamics of the seismograph, whatever the damping is. The general idea is that you measure the signal you get when you give the seismograph a "bump" (literally hit the ground with a large hammer!) and then work backwards from the actual response to see what sequence of "bumps" would have created it.

In practice, approximately critical damping is probably a good idea, as a compromise between limiting the amount of data you need to process (i.e. the time it takes for the response to a single bump to die away) and getting no response at all because the damping is too high.

Any good book on control theory or vibration measrement should explain the math.

 

[jim mcnamara]

~ 0.2 of a second.

You can surpass this by an order of magnitude, using the NTP internet protocol, or use a GPS handheld device and get more accuracy or use differential GPS and get even more.

I must be missing something, though. Expense? Extremely remote location? Excessive Muggles? (muggles are people who are not in on something, usually very techie, like seismographs; and consequently freak out about it or steal/damage parts of it.)

 

[davenn]

I am using NTP internet timing :)

speaking of quakes I am getting signal from a M7.7 Queen charolette Is, BC region rite now (03:30 UTC)

its not in a good direction for my sensor orientation so I won't be suprised if the amplitude is a bit small


Dave