Generally poor baselines in one subband

A few spectra sometimes have had one particularly noisy subband or a subband with a large DC offset or slope. Repeat the scan if this occurs. If it is still present, run das_test_config from the operators terminal. If all is well, after seconds, a list of test results will be displayed and show zero (0) errors. However, if 1 or more errors are reported, this could have several causes, including:

• IVC LO lock problems; try reloading the configuration after running test_config. If there is still a problem, try a different configuration.

• "Zero delay counter error"; the zero delay counter effectively measures the DC offset of the baseline. If an error is reported, it usually means the total power level of one ADC module (corresponding to one of the subbands) has changed during the integration or is erratic. Most common cause is a bad connection from the IVC unit to the ADC sampler rack, probably around the sma connectors on the front of the sampler rack. Identifying which physical sampler module is causing the problem depends on the configuration. See section 11 to work out which module is faulty.

• IVC RF switch problems; an error may be reported from test_config even though the baselines and data are OK. This probably means that although the switch may be faulty, it is not required for the particular receiver being used, and you may continue. Otherwise if there are still problems, try loading a different configuration or see section 12.5.

• Correlation channel errors; the correlation function is now checked for bad channels. If a single ACF (AutoCorrelation Function) channel is found to be bad, the software will interpolate the data. This should correct some of the occasional ringing found in spectra, and will not noticably degrade the final data. A note is sent to the operator saying this has occured, as well as the offending channel. If the software has found more than one bad channel, it will not attempt to fix the data, but will continue and report to the operator that the problem exists. If these bad channels continue, and they appear to be interfering with the data, then see section 12.5.