it is very clear what the failure is, as i describerd earlier, info is simply missing in the seisan channel header. so i you can regenerate the error, you can look what happens to those variables.

jens

On Fri, 22 Feb 2013, Carpenter, Seth wrote:

Hi, Miguel et al.,

I ran into a similar problem with the SEISAN putaway routines (v7.6 and 
updated) and Paul and Jens have kindly offered their help to me also.  However, 
I still haven't figured out the problem.

Miguel, could the issue you're dealing with be related to having multiple SCNs 
that are distinguished only by a unique location code?  For example, with the 
issue I'm having, a data file with the following channels is OK in SEISAN:
STA1  BHZ XX 00
STA1  BHN XX 00
STA1  BHE XX 00

While a data file with the following channels appears to corrupted (notice, for 
example, there are two STA1.BHZ.XX channels, one with loc code 00 and the other 
with loc code 10):
STA1  BHZ XX 00
STA1  BHN XX 00
STA1  BHE XX 00
STA1  BHZ XX 10
STA1  BHN XX 10
STA1  BHE XX 10

The problem I'm having doesn't seem to be related to the number of channels, 
but with having data from SCNLs that are distinguished only by unique location 
codes.

I'm just curious if this could also be a problem you're having.

Thanks,
Seth

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of 
Paul Friberg
Sent: Friday, February 22, 2013 8:07 AM
To: Jens Havskov; Miguel Palma
Cc: [email protected]; [email protected]
Subject: Re: [seisan] SEISAN update to Earthworm putaway codes 
(trig2disk/waveman2disk)


Miguel and Jens,

Please tell me some more details about the case that fails and the case that 
succeeds. Is it just the number of channels being written? 9 vs 131 that causes 
this problem? This is a tough one to debug not knowing more details. What if 
anything is special about station FDF channel BHZ? Can  you store just this one 
to disk using waveman2disk and see if it works?

Miguel, could you do some tests increasing the number of channels in a 
waveman2disk and see if any one channel causes the program to produce a faulty 
SEISAn file?

Regards,

Paul


Jens Havskov wrote:
miguel

this file, as the previous i inspected after the upgrade, is corrupt. if you 
run seiasc, you see at which channel it stops and in a hex editor you see the 
problem. in this case, for channel FDF BHZ (following 3 channels with zeros and 
sample rate 1.0), the channel info about start time, sample rate and number of 
samples is missing. seem to remember, it was the same problem in previous 
corrups file. so there seems to be a new bug in the earthworm program.

jens

On Wed, 20 Feb 2013, Miguel Palma wrote:

Hello Jens,

Due to how big the file is to be sent by email, I already uploaded to my 
Dropbox. So, please feel free to get the data form this link:
http://dl.dropbox.com/u/9051471/2013-02-20-1314-00S.MAN___131

But I have to mention about a particular situation where waveman2disk works 
flawlessly with MULPLT. This happened when I got a small set of data (9 
channels instead of 131 of the previous example) and it works!!!. So, 
considering we have a data combination between IRIS, Colombia and Venezuela, 
maybe there are some seismic stations after using waveman2disk which MULPLT is 
not happy with! This small set of data can be retrieve from this link:
http://dl.dropbox.com/u/9051471/2013-02-20-0052-00S.MAN___009

Please let me know if you need more examples o any other test you might be 
interested in.

Regards,

Miguel.




El 20/02/13 12:02, Jens Havskov escribió:
miguel

please send the bad file and i can check. i earlier recived a corrupted seisan 
file from the new module so i tmigh be the same problem.

jens

On Wed, 20 Feb 2013, Miguel Palma wrote:

Hello Paul and Jens,

Sorry because I could not test waveman2disk before, but today I had
a chance and I found some fortran runtime error while opening a
seisan format file obtained by using the new waveman2disk executable.

I got the new waveman2disk.exe from the link provided by Paul; I
ran waveman2disk to get seisan format data to be analyzed by SEISAN
V9.1 (on Linux Ubuntu 64bits, but I also tried on Windows platform
with SEISAN V8.3 also 64bits); while using MULPLT to plot the data,
I got the following error
messages:
At line 620 of file wave.for
Fortran runtime error: Bad value during integer read

On windows a pop-out Visual Fortran run-time error appeared:
forrtl: severe (64): input conversion error, unit -5, file Internal
Formatted Read.

So, MULPLT could not read the new file as usual.

Any hints? 64 bits could be the problem?

Thank you in advance and best regards,

Miguel Palma.



El 11/01/13 14:59, Paul Friberg escribió:

      Hi Earthworm/SEISAN users,

      ISTI believes we have fixed a long standing problem with the
      SEISAN waveform putaway
      routines that was recently pointed out to us by Jens Havskov. We
      have gotten confirmation
      on our tests from Jens that our fix looked okay, but we are
      looking for more
      folks who use SEISAN and EW on a day to day basis to test this.
      The problem
      was that the channel name in the wave files was incorrectly
      formatted and that
      location code and network code were not used at all in the
      "event file channel" header.

      The fixed SEISAN putaway routines have been placed into EW
      Subversion, and we
      pre-built binaries for trig2disk and waveman2disk for Windows
      that you can
      grab from here:

http://www.earthwormcentral.org/distribution/v7.6_fixes/

      Please let us know if this works for you, or not. Thanks.

      Note that this EW contribution/fix was entirely funded by ISTI.

      Cheers,

      Paul Friberg
      ISTI == http://www.isti.com

      _______________________________________________
      seisan mailing list
      [email protected]
      http://mailman.uib.no/listinfo/seisan



--
Este mensaje ha sido analizado por MailScanner en busca de virus y
otros contenidos peligrosos, y se considera que está limpio.
FUNVISIS - Fundacion Venezolana de Investigaciones Sismologicas



--
Este mensaje ha sido analizado por MailScanner en busca de virus y
otros contenidos peligrosos, y se considera que está limpio.
Fundación Venezolana de Investigaciones Sismológicas (FUNVISIS)




_______________________________________________
seisan mailing list
[email protected]
http://mailman.uib.no/listinfo/seisan
_______________________________________________
seisan mailing list
[email protected]
http://mailman.uib.no/listinfo/seisan

Reply via email to