MOS Timing mode spectra problem Darren Baskill Timing

MOS Timing mode spectra problem Darren Baskill Timing mode spectral correction

Rev 219: X-ray binary 1700+37 Cause: ADU channels are not uniform in size; some are narrower than others. Usually GATTI correction is applied, to smear out this effect. However, GATTI correction is not applied in Timing Mode as this would interrupt the timing flow A particularly narrow channel reoccurs every 16 channels. Energy (ADU) Darren Baskill Timing mode spectral correction

Energy (ADU) Darren Baskill Timing mode spectral correction

Method 1: Remove every 16 th bin Energy (ADU) Darren Baskill Timing mode spectral correction

But, it is not only the 16 th bin that is the problem. All channels have varying widths. Darren Baskill Timing mode spectral correction

Folding over 32 spectral channels Darren Baskill Timing mode spectral correction

Folding over 16 spectral channels Darren Baskill Timing mode spectral correction

Method 2: Spectral renormalisation Darren Baskill Timing mode spectral correction

Folding over 16 spectral channels Spectral Renormalisation Darren Baskill Timing mode spectral correction

Making the folded spectrum… -Must take spectral shape into account flux -Take the ratio of a channel flux to the average of the neighbouring 16 bins Energy Darren Baskill Timing mode spectral correction

Spectral renormalisation Energy (ADU) Darren Baskill Timing mode spectral correction

This fixes the spectrum well BUT… - Still noisy - Would be a new SAS task for users to run on a processed spectrum (not transparent to the user) Darren Baskill Timing mode spectral correction

Method 3: Spectral redistribution of the original ODF file Darren Baskill Timing mode spectral correction

Folding over 32 spectral bins Retain events, but redistribute them in energy space. All other characteristics of an event would be maintained. Darren Baskill Timing mode spectral correction

Method -Folds spectrum -Calculates “move from channel” probability for each event -Calculates a “move to channel” probability for each channel - Moves event in energy (adu) channel Darren Baskill Timing mode spectral correction

Energy (ADU) Darren Baskill Timing mode spectral correction

4 U 0543 (quasar) – rev 520 Darren Baskill Timing mode spectral correction

Conclusion Timing mode spectra need correcting Spectral Redistribution looks the best method Darren Baskill Timing mode spectral correction

http: //www. src. le. ac. uk/projects/xmm Darren Baskill Timing mode spectral correction