Our full technical support staff does not monitor this forum. If you need assistance from a member of our staff, please submit your question from the Ask a Question page.


Log in or register to post/reply in the forum.

SerialInRecord problem for 20Hz Eddy Covariance application


JBL Jan 27, 2015 01:17 PM

Hi,

I work on an Eddy Covariance data acquisition system with a CR3000 (27.04 OS version), a Gill Windmaster Pro anemometer and a Licor LI-7200 gas analyzer through the ethernet card NL115. Data acquisition is ideally fixed to 20Hz.

To save data from the analyzer, I use "SerialInRecord" instruction but some times, NAN data are saved in my table. I have no problems with "SerialIn" instruction but I can't use it correctly because there are no sync and terminal characters.

This problem is also present on the CR6...

Do you have an explanation? or alternative?

Thx

J-B


JDavis Feb 6, 2015 08:41 PM

There are two main causes for the occasional NAN values from an ultrasonic anemometer.

1) You will receive bad values from an ultrasonic wind sensor under a variety of transient environmental conditions such as heavy rain or a big bug flying through. These can't be avoided. You just need to watch out for and handle these missing data points.

2) When running any sensor in an asynchronous fashion, there is the chance for occasional missed values due to clock differences of the datalogger and the sensor. One way to compensate is setting the output of the sensor to double the rate that you need. You will occasional miss one out of two values, instead of one out of one. You could alternatively run the sensor in a polled fashion, then there will always be one output from the sensor per one scan of the datalogger.


JBL Feb 9, 2015 06:55 AM

Hi,

First, thanks for your answer.

1) I not receive bad values from the sonic anemometer due to stron,g meteorological conditions because he's in my office, and when I receive data from the RS232 and terminal, or Gill software, there is no missing 'NAN' data.

2) I have tested 20Hz anemometer output for 10Hz datalogger acquisition, but I have the same problem. In this configuration, when I switch from SerialInRecord to SerialIn, the problem is solved but I can't easily extract data with SplitStr when I have different data frame length: is there a big internal difference between these two instructions (other sync characters)?
And the problem is also present in polled mod.

Log in or register to post/reply in the forum.