<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body bgcolor="#FFFFFF" text="#000000">
We've seen another instance of the 1024-week rollover, similar to
the K2/Altus issue discussed last summer. This time, it was Garmin
clocks on Webtronics digitizers in the MB(Montana) network. I
changed the PsnAdSend Earthworm module to include a similar patch to
what I used for the K2/Altus k2ew module. If you're having issues
with data being timestamped in 2000, it's probably the 1024-week
issue. There is a firmware patch for the Garmin clocks that might
fix it, but that requires physical access which may not always be
easy/possible.<br>
<br>
<br>
Code snippet & context for the change to PsnAdSend.cpp:<br>
<br>
/* Make the trace buffer header */<br>
void MakeTraceHeader( SCNL *pSCNL, double tm, DataHeader *hdr )<br>
{<br>
<b>/* check for 1024 week rollover glitch 2018/2019 jr */</b><b><br>
</b><b> if (tm > 911143558 && tm< 1530403744 )</b><b><br>
</b><b> {</b><b><br>
</b><b> tm += 619315200; /* add 1024 weeks of seconds */</b><b><br>
</b><b><br>
</b><b> }</b><br>
<br>
traceHead2->nsamp = ChanRate;
// Number of samples in message<br>
traceHead2->samprate = ChanRate;
// Sample rate; nominal<br>
<br>
<br>
<br>
<br>
<pre class="moz-signature" cols="72">--
===================================
Jon Rusho
Seismic Network Engineer
University of Utah Seismograph Stations
<a class="moz-txt-link-abbreviated" href="mailto:jon@seis.utah.edu">jon@seis.utah.edu</a>
801-585-5523
</pre>
</body>
</html>