CRTN NorCal Misconfiguration

[Jim Frame]

It was recently discovered that all of the northern California CRTN mountpoints have been casting incorrect base positions. The problem was fixed today (reportedly for all NorCal stations, I only checked UCD1), but the problem began sometime between January 14 and February 12. If you've relied on any CRTN positions recorded between January 14 and today (March 25), you may want to review them to determine if remedial action is necessary.

[Edward M Reading]

Yikes!

[land butcher]

Anyone know if this was limited to NorCal or is SoCal included?

[Jim Frame]

According to Yehuda Bock, only the NorCal stations were affected.

[Jim Frame]

CRTN sent this out this afternoon:

Alert (3/26/2019): It was discovered that the northern California CRTN server (132.239.152.175) was transmitting incorrect coordinates in the RTCM3 messages. This started on January 16th when we moved the server to another workstation. The problem was resolved in the late afternoon of March 26th. Anyone who relied on the CRTN coordinates over this time period will need to revisit their rover coordinates. Please contact us if you believe your surveys may have been affected by the erroneous coordinates. Note that the southern California server was not affected. Thanks to Jim Frame and Art Andrew for discovering the problem. We’re sorry for the inconvenience.

[Edward M Reading]

How much were they off Jim?

[Jim Frame]

How much were they off Jim?

It varied. The one I use the most -- UCD1 -- was off by about (but not exactly) a meter horizontal and a foot vertical. However, I did a quick check on another nearby station -- P268 -- and it was off so far that I couldn't do the comparison in my head.

[Edward M Reading]

How much were they off Jim?

It varied. The one I use the most -- UCD1 -- was off by about (but not exactly) a meter horizontal and a foot vertical. However, I did a quick check on another nearby station -- P268 -- and it was off so far that I couldn't do the comparison in my head.

A meter and a foot are errors that might cost some serious dough.
Good catch Jim.

[Jim Frame]

There wasn't any magic sleuthing on my part -- I was simply trying to check into known control and missing by a lot. The hard part was persistently bothering both Javad and CRTN to determine where the problem was, as neither wanted to own it. I'm just happy that it's been fixed!

[T. S. Higgins]

Just as a point of reference, we checked into several local projects performed during the "bad window" utilizing the CRTN TRLK base and were only missing in the 0.15' range horizontally, and less than 0.10' vertically.

Considering most of our projects are run from local control established in reference to the CRTN and then never published as CCS83 coordinates, it wasn't a great concern either way; nothing too concerning on the 4 projects that I spent the afternoon checking though.