Dual-beam measurement acquisition issues

[dlnorgaard]

The dual-beam acquisition of data does not appear to work in adaptive mode. Also, for some reason, Christoph is getting results of both channels into the same evaluation log files whereas I was getting them into two separate ones during testing.

On Sat, Jul 14, 2018 at 11:25 AM, Kern, Christoph ckern@usgs.gov wrote:
Hi Diana

I have made further tests with the dual beam Mobile DOAS software. You were right in that the fitting of the Slave channel does indeed work when using a fixed exposure time! In that case, the spetrum fits as well as the red and blue bars displayed in real time make total sense. However, the results of both spectrometer channels are still written into the same EvaluationLog. Not sure if this is intended.

Attached is some example data we collected on 7/12. These are wind speed measurements that we made while sitting under the plume with the master channel looking 10 deg towards the vent, the slave looking 10 deg away from the vent (20 deg separation of the 2 channels).

The cfg.xml is the same as before, with just the exposure time now fixed to 50 ms.

Thanks!

Christoph
Kilauea20_20180712_1223_windspeed.zip

[dlnorgaard]

Description of what Christoph was seeing in adaptive mode is below. I tried running it in office and do appear to get an odd fit for the second channel:

On Tue, Jul 10, 2018 at 7:35 PM, Kern, Christoph wrote:
Hi Diana and Matias

I went out and collected some dual beam MobileDOAS data today. The good news is that the spectral acquisition seemed to work really well, and I'm confident I can analyze the data later in the office.

What didn't work so well is the real time spectral analysis. I'm not sure why, but the fit for the second channel of the spectrometer was very strange. It didn't seem to really be fitting SO2. It immediately showed -2000 ppmm, then changed only by 2000 ppmm (just reaching 0) while the first channel went up to more than 10,000 ppmm. I also tried unplugging the channel 2 fiber, but the fit still gave numbers very close to 0 with hardly any noise. I'm not sure how this is possible for dark spectra.

Do you think you could check and see if you can find any obvious mistakes in the fit procedure for the second channel? I tried reproducing the fits in DOASIS, and they seem to work better there. Not sure what is going on...

I'm attaching some of the collected data. Kilauea 11 is a wind measurement attempt, Kilauea 13 contains 2 traverses.

Thanks

Christoph

[dlnorgaard]

The fit window names must differ in order to get separate evaluation logs for each channel. Some code changes were made to ensure that fit window names differ.

[dlnorgaard]

Issue in adaptive mode fixed. Also, there seems to be an agreement that we will have two eval logs, one for each channel. Changes in GUI has been made to ensure this.