Cosmic Origins Spectrograph Instrument Handbook for Cycle 17

5.2 Non-linear Photon Counting Effects (Dead-time Correction)

---

The electronics that handle the COS detectors have a finite response time, and that limits the rate at which they can detect photons. This effect of non-linearity is sometimes known as the dead-time correction.

For the FUV channel, there are three factors that influence the detected count rate. The first is a Fast Event Counter (FEC) for each segment that has a dead time of 300 nsec. The FECs only matter at count rates well above what is usable, amounting to a 1% effect at a count rate of 33,500 per segment per second.

The second effect is due to digitization of the detected events, and that has been measured for the FUV XDL detector, with a dead-time constant of 7.4 µsec. For a given true count rate C, the detected count rate is given by:

where D is the detected count rate and t is the dead-time constant. For t = 7.4 µsec, the apparent count rate deviates from the true count rate by 1% when C = 1,350 counts sec^-1, and by 10% when C = 15,000 counts sec^-1. Note that when the effect is near the 10% level, then the FUV detector is near its global count rate limit (see Section 11.5) and so non-linear effects are small for the FUV detector.

Finally, for the FUV channel the count streams for the two separate segments must be combined in a single "round robin" Detector Interface Board (DIB) that takes signals from both the A and B segments and then stores them in the data buffer. The DIB interrogates the A and B segments alternately, and, because of this, a high count rate in one segment but not the other could lead to an additional effect. The DIB is limited to processing about 250,000 events sec^-1 in ACCUM mode and only 30,000 counts sec^-1 in TIME-TAG mode. Tests have shown that the DIB is lossless up to a combined rate for both segments of 20,000 sec^-1, and the loss is 100 sec^-1 at a rate of 40,000 sec^-1. Thus this effect is less than 0.3% at the highest allowable rates, and there is information in the engineering data that characterizes this effect.

For the NUV MAMA on COS, the dead-time is the same as for the STIS NUV MAMA, which is 280 nsec. Note that for the STIS MAMAs the 1% level of non-linearity is reached for C = 36,000 counts sec^-1. The MAMAs also show a local non-linear effect that is small and will be calibrated on orbit.

Dead-time corrections are automatically made in the calcos pipeline.