PSN-L Email List Message

Subject: Re: WinQuake
From: "Geoffrey" gmvoeth@...........
Date: Tue, 10 Mar 2009 13:38:10 -0700

Howdy Mr. Thomas Dick;

Answers inline below.

----- Original Message ----- 
From: "Thomas Dick" 
Sent: Tuesday, March 10, 2009 5:01 AM
Subject: Re: WinQuake

> Geoffrey wrote:
>> FYI:
>> There Is Only One Way I can Get Winquake to
>> see right times for my data.
>> It is necessary to crop the data to the nearest
>> second before giving it to Winquake.
>> No matter what I do WQ will not see fractions of seconds
>> for the start time.
>> No matter what the precision is of the numbers
>> WQ only will show the first sample to a whole second.
>> Like if I tell it 23:49: 20.4 for the first second
>> it will show only 23:49:20 for the first sample
>> SO you must run your data until it gets to
>> 23:49:21 then crop it from there to the end
>> before giving it to wq and tell wq
>> it is 23:49:21.0 and you can forget the fractions
>> your start time may be a couple milliseconds off
>> but WTHeck.
>> Otherwise if you do not do this things just aint right.
>> Also, the various filters in the WQ3 FFT used to work fine
>> but since installing the new WQ3  those filters dont work (all grayed 
>> out) ??
>> What am I doing wrong ?
>> cheers
>> geoff
> Geoff and all ....
> I know I don't belong in the same theater as most of you -- let along 
> the same stage. This is going to show my ignorance. I am confused about 
> the value of your pursuit into "It is necessary to crop the data to the 
> nearest second before giving it to Winquake".

Hey...thats why the academia split up the knowledge into celluraized groups
cause there is not enough human life for one to know everything.

> Here is where I am coming from;
> 1. when you try to get that close with the timing issue on P, aren't you 
> ignoring the timing errors caused by internal computer timing 
> inaccuracies and delays caused by getting the timing signal from the 
> satellite or whatever your  source is

My system is entirely slaved to a ring counter syncronized via
WWV at whatever freq works when tuned in via SW radio.
When the ring counter tells my computer it is exactly
HH:MM.00 then thats what the time is in the record.
The errors up to the ring counter are being entirely ignored
since I do not have the capability of dealing with
millisecond errors.
The computer clock only gets me to the correct minute.
The ring counter to tens of milliseconds and math does the rest.
I am totally unable to tell what the real time is except
my robot clocks (WWVB) all agree Im in the very same second.
But Winquake should simply do what i tell it to since it is no
master of time. If I tell it the first sample is whatever date and time
it should by all rights reflect this fact. It really should not guestimate
amy time between the samples unless it does so between the first and the last
as relating to the sample rate. The most impoirtant thing is that you give it
the right time in the first place as well as the proper sample rate.
not as easy to do as youd like to imagine.

> 2. the time of arrival tables for earthquake waves that are used are not 
> set in stone ---- they are averages ..... are you taking this into 
> consideration ...

Trust the masters since the regular person does not have
the ability to generate such tables. For ametures this is fun
and not life and death.

> 3. are you also pursuing the reaction time delays within the seismic 
> unit as well ... just how long does it take for you units to show the 
> first arrival of P

Are you talking about delays through the Electronics ??

I have a hard time seeing any delays with the test equipment
at my fingertips. I only know there is a phase delay meaning you put a signal
in at 0 degrees but get one shifted in phase by 45 or 180 or 270 or whatever.
If I DROP something on the ground it is like the ground first goes downward
so if my trace is upward I simply swap the leads intp the amplifier
to reflect the initial ground motion.
But for all testing I see no actual delays only phase shifting.
Even 50% the speed of light is nothing to 5 HZ or less.

> 4. and finally, what about the inaccuracies within the graphics display 
> of the computer .... what degree of definition are you using and will 
> this effect timing

I use no graphics for recording. You simply make sure that what you are looking at
is good enough to show each and every sample per line so if you look at
1000 samples you want like 1000 to 1024 pix on the horizontal.
Any other times than pixel times are simply an educated guess.
But what I have found is that sometimes my recorder may either miss or
add an extra sample you do not expect and its practically impossible
for me to tell what samples where are not right in the middle of everything.
If you cant find the errors you simply assume the thing works. BUT
you want your software to behave exactly as you expect it to work.
Computers are not human and they should never have a mind of their own.
They should only do whatever you tell them to do or something else is wrong too.
Possibly the programmer has made mistakes in the philosophy that
built the program in the first place.
(for time that is)
Amplitude is yet another story to examine.

> I am NOT attacking your efforts but trying to appreciate where you are 
> headed with your efforts.

Simply put, my times are all slaved to the ring counter which is synchronized to the minute
mark in WWV and the computer RTC gets me into the neighborhood ( DARE +  nearest minute after midnight )
rest to mathematically work everything else out to the millisecond ( do not round numbers till the last result is desired ).
Ta DAAAA  :-)

> Yours, TomD
> __________________________________________________________
> Public Seismic Network Mailing List (PSN-L)
> To leave this list email PSN-L-REQUEST@.............. with 
> the body of the message (first line only): unsubscribe
> See for more information.

Public Seismic Network Mailing List (PSN-L)

[ Top ] [ Back ] [ Home Page ]