Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stsci.edu/hst/fgs/documents/isrs/fgs_isr25.pdf Дата изменения: Mon Aug 23 20:24:58 2004 Дата индексирования: Sun Dec 23 00:12:24 2007 Кодировка: Поисковые слова: magnitude

FGS Instrument Report No. 25
1November 1993

Correcting FGS Counts for Deadtime in the PMTs
M. G. Lattanzi and L. G. Ta

1 Summary
In our calibration of the FGS photometric system, we are using stars in the magnitude range V = 8.3,13.8. As shown in this report, magnitudes brighter than V = 9.5 mag taken through the Clear Filter; F583W contribute an error of more than 0.05 mag if proper account for deadtime in the detector photomultipliers PMTs is not allowed for. The procedure to remove the e ect of deadtime is given in x 2. Since the deadtime constant is based on pre-launch tests no in- ight calibrations have been executed nor are any planned for the foreseeable future, the e ect of uncertainties in the deadtime constant provided have to be addressed; see x 3.

1


2 Deadtime Correction
The formula to correct from measured counts CM per unit of integration time to "true" counts CT is Astrometry Handbook 1986 CT = 1 1 CM 1 , CM TD=TI ; where TD is the deadtime constant 285 nanoseconds; Reed 1993 and TI is the integration unit time 0.025 sec for the FGS. Eq. 1 is expressed on a magnitude scale simply by taking the common logarithm of the ratio CT =CM . Fig. 1a shows the amount of correction in magnitudes owing to the PMT deadtime which needs to be applied to the instrumental magnitudes as derived from the measured counts as a function of visual Johnson magnitude. The absolute throughput counts is that for the F583W Clear lter. For example, for our standard single star Upgren 69 V = 9.6, the deadtime correction amounts to 0.09 mag.

3 Accuracy of the Correction
The deadtime constant given in the previous section is based on prelaunch measurements. There has been no in- ight calibration of the deadtime constant either during Orbital Veri cation or Science Veri cation. In Eq. 1, the measured counts are known to the limit of the photon noise, while TI is a given number. Thus, an error in the adopted value for the deadtime constant will result in an error in the predicted correction derived from Eq. 1. This error is in dex 2

E m= 2:5 Log e 100:4m CM TD =TI E TD=TD ;
2


where E m is the error of the correction m as derived from Eq. 1 after conversion to a magnitude scale and E TD is the error of the deadtime constant. In Fig. 1b is shown a plot of Eq. 2 as a function of magnitude for an assumed error in the deadtime constant of 25 worst case. From the two gures we conclude that the deadtime correction is important enough to be routinely removed from the data. The contribution from the assumed uncertainty in the deadtime constant is several times less than the sought for correction the actual errors being probably smaller than what is depicted in Fig. 1b.

4 References
L. Abramowich-Reed 1993, private communication. L. Abramowich-Reed and C. Ftaclas, 16 Sep., 1993, Bright Object Observations with the FGS, EM: MOSES 1024. HST Astrometry Operations Handbook 1 Oct., 1986, SMO-1040 pp. 3-62.

3