PACS. Optimum detector bias settings for Ge:Ga detectors, Time constant: bias change spectrometer IMT 509

Transcription

1 Test Analysis Report FM-ILT/IST Page 1 Optimum detector bias settings for Ge:Ga detectors, Time constant: bias change spectrometer IMT 509 J. Schreiber 1, U. Klaas 1, H. Dannerbauer 1, M. Nielbock 1, J. Bouwman 1 1 Max-Planck-Institut für Astronomie, Königstuhl 17, D Heidelberg, Germany

2 Test Analysis Report FM-ILT/IST Page 2 Req. IMT 509 Optimum detector bias settings for Ge:Ga detectors, Time constant: bias change spectrometer IMT A. History Version Date Author(s) Change description 1.0 June 1, 2007 Juergen Schreiber, Ulrich Klaas, First issue Helmut Dannerbauer, Markus Nielbock, Jeroen Bouwman IMT B. Summary The minimum NEP and therefore the corresponding optimum bias is reached between 60 and 70 mv bias voltage for the red array and 169 and 199 mv for the blue array which is slightly lower than found in the module level tests. Some more pixels than the already known bad pixels show an extraordinarily deviating NEP behaviour and are identified. For some bias steps signal transient behaviour was found with decay times between 1/2 to 4 minutes. The decay times setting the optimum bias of the red array were determined to be between 1 and 2 minutes. The recommendation is therefore to insert wait times of about 8 minutes should a bias change be necessary. No transient behaviour could be detected for the optimum bias voltage setting of the blue array. Of course, these statements are only valid for the used constant bias steps of 10 mv for the red and 30 mv for the blue array, respectively. IMT C. Data Reference Sheet Ref Date Archive filename TM file 1 10/04/07 FILT IST SPEC detector imt509 OBS 45.tm RD-1 13/02/06 DEC/MEC User Manual, -CL-SR-002 RD-2 01/03/05 CQM-ILT Analysis Report, Part II, PICC-KL-TR-001 RD-3 22/02/06 Test Analysis Report EQM-IMT: Optimum Bias Settings for Ge:Ga Detectors (Test ID: 523) RD-4 01/03/07 PTD RD-5 05/01/07 Calibration Document -MA-GS-001, draft 8 RD-6 10/08/06 Cold Performance Tests on FM High Stress Ge:Ga Detector Modules, -ME-TR-063 RD-7 16/03/06 Summary of the cold performance tests on LS-FM Ge:Ga detector modules at MPIA, -MA-TR-030 RD-8 12/02/07 Wavelength Calibration of the Spectrometer FM ILT part I -Calibration Analysis Report PTD 4.2.1

3 Test Analysis Report FM-ILT/IST Page 3 IMT D. Test Description IMT D.1. Introduction This FM-ILT/IST test was a preparation and reference test for the FM-IST tests at Astrium/Friedrichshafen. The goal of this test is to verify the optimum bias voltage at which the Ge:Ga detectors work stably (e.g. do not show spiking pixels) and the NEP shows a minimum. Transient effects of the signal after a bias change should be quantified. The resulting values should be compared to the corresponding results of the module level tests. IMT D.2. Test Overview The FM-IL/IST bias scan measurements were carried out on 10 April 2007 during the final phase of the FM ILT. The temperature of the blue array was set to 2.5 K which is for thermal budget reasons the maximum temperature close to the optimum derived during the nominal FM-ILT Ge:Ga optimization tests. The chopper was held at a constant position of staring to the internal calibration source 2 which was at a temperature of 60 K. The grating position was set constant at corresponding to a wavelength of µmfor the red array and 87.8 µmfor the blue array. The smallest capacitance of 0.14 pf (effective) was used. Data acquisition was done in default mode, i.e. averaging 32 readouts each in 1/2 s ramps, consequently this results in 4 samples per ramp that were fitted linearily to get the signal in readouts per second. The whole 16 bit ADC range was assumed to correspond to 6.26 V. For the red array we scanned through a bias voltage range between 20 and 90 mv in steps of 10 mv, while for the blue array a range between 80 and 290 mv in steps of 30 mv was sampled with a time length of about 120 s for each step. In Fig. 1 the course of the signal of one pixel of the red and blue array is shown overplotted by the applied bias voltage and the errors of the fit of the averaged ramps, which is a measure of the noise. The increase in signal with increasing bias voltage is clearly visible (bias voltages are negative against ground). The blue pixel clearly shows a transient behaviour at high bias voltages after a bias change. Also the increase of noise with increasing bias can be recognized. The known dead pixels described in RD-8 (4 in each array), the open and the dummy channels are not considered in the following calculations.

4 Test Analysis Report FM-ILT/IST Page 4 Figure 1: signals, fit errors, and bias voltages of red and blue pixel 10, 10

5 Test Analysis Report FM-ILT/IST Page 5 IMT E. Results IMT E.1. Optimum Bias Voltage Since the absolute flux of the calibration sources is not accurately known at the moment, the standard deviation (stddev) divided by the signal is used as a representative measure for the NEP (which is actually proportional to this measure). In Fig. 2 the mean and median NEP measures of all pixels of the red array is plotted against the bias, the error bars are the calculated standard deviation of this measure over all pixels hence reflect the dispersion. The minimum of the mean and median NEP measure is at a bias voltage of 70 mv, where also the dispersion shows a minimum. The difference of the NEP measure at 60 mv and 70 mv bias voltage is very small and the dispersion of the 60 mv NEP measure includes the whole 70 mv range. Consequently, the whole bias range between 60 mv and 70 mv could be considered as the optimum bias voltage. In Fig. 3 the NEP measures of the whole red array at 70 mv are displayed except the open and dummy channel. The 4 known dead pixels ([11,3], [11,5], [10, 19], [5, 22]) show up in black, while the 2 known low response pixels ([12, 1], [7, 12], see RD-8) are in white. But there are 2 more deviating pixels with high NEP at the positions [11, 2] and [16, 2] which were not conspicuous at module level tests. The boarder modules at y-position 4, 9, 14, 24 show up brighter on average than the other modules. This is very likely due to a remaining mis-alignment which leads to a reduced flux falling onto these boarder modules, thus increasing the NEP. Figure 2: mean (left panel) and median (right panel) stddev/signal vs. bias for all pixels of the red array. The error bars reflect the dispersion of the measurement. The mean and median NEP measures of the blue array and their dependence on the bias voltage are plotted in Fig. 4. The minimum of the mean NEP measure is at a bias voltage of 169 mv (measured housekeeping value while 170 mv were commanded). For the median NEP measure the minimum is found for the bias voltage of 199 mv (measured housekeeping value while 200 mv were commanded), but this is only slightly lower than the median NEP measure at 169 mv. For both measures the error bars are minimum for the 169 mv bias voltage. The differences of the NEP measures at 169 and 199 mv are very small, consequently, the whole bias range between 169 and 199 mv could be considered as the optimum bias voltage. In Fig. 5 the NEP measures of the whole blue array at 169 mv are displayed except for the open and dummy channels. The 4 known dead pixels ([6, 15], [3, 16], [5, 22], [2, 23]) show up in black. In the blue array all boarder modules at y-position 4, 9, 14, 19 and 24 are affected by poor alignment and show up brighter than the other modules. Additionally, there are some deviating pixels showing up exceedingly bright in the left-most module and in module 23. Also the pixels [1, 19] and [9, 23] show up extraordinarily bright.

6 Test Analysis Report FM-ILT/IST Page 6 Figure 3: mean stddev/signal at 70 mv bias of the red array. Modules (pixel columns in y-direction) are counted from 0 to 24, pixel rows are counted from 0 to 15. Figure 4: mean (left panel) and median (right panel) stddev/signal vs. bias for all pixels of the blue array. The error bars reflect the dispersion of the measurement.

7 Test Analysis Report FM-ILT/IST Page 7 Figure 5: mean stddev/signal at 169 mv bias of the blue array. Modules (pixel columns in y-direction) are counted from 0 to 24, pixel rows are counted from 0 to 15.

8 Test Analysis Report FM-ILT/IST Page 8 IMT E.2. Signal Transient Behaviour after Bias Change In order to estimate the stabilisation times of the signals after a bias change we fitted the signal course of each bias setting by a simple exponential function after subtracting the mean of the last few signals of the plateaux (see fit examples in Fig. 6). Practically, a linear fit was applied to the logarithm of the signal values. Therefore, the exponential function is offset-free, but has a multiplicative offset as free fit parameter. The other free fit parameter was the decay time of the exponential function. Figure 6: Examples of exponential fits to signals after bias change, left panel: red pixel, right panel: blue pixel. The resulting median exponential decay times of all pixels of both arrays are depicted in Fig. 7. The error bars show the dispersion of the decay times over all pixels. The resulting median multiplicative factor of all pixels is shown in Fig. 8, where the error bars indicate the standard deviation. The red array has only negative offsets for all biases (although not significant for the 2 highest bias voltages), i.e. the signals rise with time up to the final plateau value. The resulting decay times are not significant for 30 and 40 mv bias voltage and the multiplicative offsets are very small, i.e for these biases a transient behaviour cannot be confirmed. The fastest decays (decay time about 1 minute) was found at the smallest bias voltage level (20 mv). There is a minimum of decay times at the optimum bias voltage of 69 mv with decay times of about 1 to 2 minutes. The significant decay times span a range between 1 minute to 4 minutes. Therefore, we recommend to include a wait time of about 8 minutes after each bias change before restarting data acquisition with the red array. The transient behaviour of the blue array looks a bit different. The multiplicative offset is negative for the smallest bias and changes to positive values for the highest biases. This indicates rising signals with time up to the plateau value for the smallest bias (80 mv) and decreasing signals down to the plateau values for the highest biases (200 to 290 mv, although not significant for 290 mv). The resulting multiplicative offsets of the intermediate biases (110, 140 and 170 mv) are close to zero and therefore not significant. Additionally, the decay times at these biases are very high and not significant due to the high dispersion, i.e. a transient behaviour at these bias voltages cannot be really confirmed. The significant decay times span a range between 1/2 minute to 4 minutes, therefore, we recommend to include a wait time of about 8 minutes after each bias change before restarting data acquisition with the blue array.

9 Test Analysis Report FM-ILT/IST Page 9 Figure 7: Median decay time of signals after a bias change from the next lower level for all pixels, left panel: red arry, right panel: blue array. The error bars indicate the dispersion over the whole array. Figure 8: Median fit result of the offset factor to the exponential fit function for all pixels, left panel: red array, right panel: blue array. The error bars indicate the dispersion over the whole array.

10 Test Analysis Report FM-ILT/IST Page 10 IMT F. Conclusions The bias scan tests were carried out successfully and allow a verification of the optimum bias setting for both arrays. The minimum NEP and therefore the optimum bias is reached between 60 and 70 mv bias voltage for the red array and 169 and 199 mv for the blue array. This result differs from the bias of minimum NEP found in the module level tests where slightly higher optimum bias voltages were established (which is between 70 and 80 mv for the red array (see RD-6) and around 200 mv for the blue array (see RD-7)). Some more pixels than the already known bad pixels show extraordinarily deviating high NEP behaviour. Future tests of this type should provide a check whether the number of affected pixels remains stable. For some bias change steps a transient behaviour was found with decay times between 1/2 to 4 minutes. The decay times at the optimum bias of the red array were determined to be between 1 and 2 minutes while no transient behaviour could be detected at the optimum bias voltage of the blue array. However, from the general transient behaviour wait times of 8 minutes after a bias change are recommended. IMT G. IA scripts used / remarks on PCSS CAP ist.py spg java function fitramps() to fit all default mode ramps. IMT H. Lessons learned for IMT/IST/PV The measurement time after the bias changes should be extended from 2 minutes to about 8 minutes (if possible) to be sure to reach a stable signal plateau. This would also allow a more accurate determination of the transient behaviour. The right-most modules at the boarder of the field of view seem to receive less radiation than the others. It is known that the alignment was not completely perfect by the repair before ILT period 2. Another alignment improvement was done after ILT period 2, so that changes are expected for the upcoming IST tests.