Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/calibration/documentation/epic_cal_meetings/200610/MartinM.pdf Äàòà èçìåíåíèÿ: Wed May 14 13:06:25 2014 Äàòà èíäåêñèðîâàíèÿ: Wed Apr 13 13:16:26 2016 Êîäèðîâêà: koi8-r Ïîèñêîâûå ñëîâà: cygnus loop

EPIC Cal Meeting 26-27.10.2006

First attempt to derive deadtime of EPIC-pn from FIFO reset counters
Michael Martin

E. Kendziorra, T. Schanz, A. Santangelo

26.10.2006

IAAC-Dn PeadtimETromtFiFo uiesteon EPI T p E d MOSF e f Da a Aq R si i ts

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

1


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

Overview
· · · · · · Introduction Data readout chain Test setup Measurements First results Summary and outlook

EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

2


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

Introduction to FIFO resets
FIFO reset leads to loss of data · In orbit:
­ possible explanation: maximum ionizing particle leads to saturation of on-chip amplifier stage Effect was seen by illumination with radioactive Am-source (5 MeV Alpha)

· · · · · W · · ·

On ground:
­

Complete column is above lower energy threshold FIFO in readout chain is filled with events from row additionally to normal events FIFO overflow can occur, causing reset How many frames are lost due to the reset? orking plan: Look for correlation between Ax_Pmode counter and number of lost frames Correction factor on time for each mode (FF, eFF, SW, LW, Timing) E s ti m a te d e a d ti m e correct effective time on source

EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

3


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

Data readout chain
EPIC-pn event data input event word 1 event word 2 time word 1 time word 2

FIFO reset counter

Reset FIFO

FIFO full

FIFO (1024 x 16 bit) max. 511 events & time word

FIFO reset to prevent data corruption

CPU & electronics

Time word is needed to trigger CPU processing

EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

4


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

Test Setup I
@ IAAT lab-model of EPIC-pn flight level hard and software revisions

BUT operating just one CCD

force FIFO overflow by illuminating pn-CCD with LED pulse

EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

5


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

Test Setup II
Comparison between frame times (FT) in flight and at IAAT (FT are given in ms)

Mode FF eFF LW SW TM BU

EPIC-pn IAAT-pn FT FT 73.4 199.2 47.7 5.7 6.0 72.0 170.0 43.0 5.7 6.0 Not yet tested

T/FT -1.4% - 15% - 10% 0% 0%

EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

6


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

Test Setup III
Recording of the FiFo status

EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

7


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

Measurements
LED pulsed with Agilent AWG Square pulse with external trigger Amplitude: Period: Pulse width: # of pulses: 2-6V 1 sec 0.5-1msec 100

1. HK-files processed with IDL-progamme (@go) 2. Extract time words and search for irregularities 3. Check if found # of time gaps equals # of pulses 4. Get the time difference of gaps 5. Compute # of lost frames 6. Get conversion factor from Ax_Pmode counter
EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s M.Martiin tn InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh 8


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

First Results
Mode FullFrame eFullFrame Large Win Small Win Timing # FIFO Resets (Ax_Pmode) 613 600 582 634 Missing Frames Conversion Factor 613 613 602 1241 1 1.0217 1.0344 1.9574

Dead time = # FIFO Resets * CF * frame time

EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

9


Introduction

DEPMOSFET

Da Aqu nti 26 Su6 EPIC CatlaMeetiisigon -27.10.200mmary

Summary and Outlook
· First attempt to estimate dead time looks promising · Larger data sets are needed · Investigate strange behavior of Small Window mode · Test with EPIC-pn flight spare model at MPE PANTER facility needed for verification

EAAT pnEPMdtSFEfrom taFoqResieiton I PIC- D dea Oime T Da Fi A ui s t s

M.Martiin tn

InsttiitutfÝÝr Atsoronoimie und tAopropikysik s tut f r As r t nom e und As r st hysh

10