PSN-L Email List Message

Subject: Re: Newbie trying to get started
From: ian ian@...........
Date: Sat, 21 Jun 2008 09:08:40 +0100


I'm running what might be considered to be the worst case scenario, 2 
pc's 1 windows box and 1 linux.  I have a National Instruments A/D and I 
couldn't be bothered installing the open source linux drivers, windows 
drivers are part of the standard labview install and it all works "right 
out of the box".  The machine runs windows 2000 and hasn't been rebooted 
for over a year nor has it crashed.  It's a 800 MHz pentium.  Everything 
is on a UPS.

 From there, the linux box (1 GHz pentium) makes a tcp socket connection 
to the windows box and "sucks" the time stamped data.  It then uses 
pgplot to graph and x screen grabs to keep the website updated.  All 
done from C and csh scripts.  One day I'll do something sensible and 
merge the 2 halves into 1 machine.

I'd recommend a hardware filter.  If you do it in software you need a 
fast processor and you need to up the sampling rate of the A/D.  Larry 
Cochrane sells filter boards through the psn website and they are pretty 
mainstream in the community.  I have a couple of dual channel Krohn Hite 
3322 programmable filters.  One is next to the sensor and set up as the 
high pass filter and the other, 100ft away, set up as the low pass 
filter (to filter computer noise etc)



Steinar Midtskogen wrote:
> [ChrisAtUpw@........
>>        You seem to be trying to do things the hard way!
> It is also a good way to learn how things work, so I'm willing to make
> a try.
>>        The SEP box is designed to plug directly into almost any old
>> PC running windows! You download the AmaSeis data logging and
>> analysis software from http:
>> // This gives you a WHOLE
>> SUITE of dedicated seismic programs, including both low and high
>> pass digital filters, FFT analysis and a lot more! The hard work has
>> already been done for you!
> I guess, but it is the x86 Windows part which is my problem.  True, I
> can probably get a 20 year old Windows PC for free somewhere, though
> it will still add to the electricity bill (wouldn't a PC with moving
> parts, i.e. harddisk and fans, nearby the seismometer cause some
> vibrations?) and it would take up even more space in the basement
> storeroom (where I have the seismometer) which I don't really have.
> My real concern is probably that I don't like to run software which
> I'm unable to modify to my needs.  No source code = no control.
> Anyway, the piece of code for reading the data was trivial, just a few
> lines of C code.  And gnuplot does the plotting job for me, at least
> the plotting that can be done automatically for a web page.  All this
> was done in a few hours.
> So I think the only thing I'm missing is the filter.  Does any seismic
> programs come with source code?
> But I'll have a look at AmaSeis anyway, which might be useful for
> manual analysis of seismic events.  If I make my logger program store
> raw data in a way AmaSeis can read it, I don't need to run AmaSeis on
> my datalogger.  My regular PC is an AMD64 running Linux, but chances
> that AmaSeis will run on it using Wine should be good.
> I did make a very simple filter:
>     output[i] = output[i-1] + alpha * (input[i] - output[i-1]);
> and experimented with different values for alpha, but I suspect that
> this filter is too soft.

Public Seismic Network Mailing List (PSN-L)

[ Top ] [ Back ] [ Home Page ]