Testing of the new VIPIC chip for X-photon counting applications Intern: Pietro Peronio 1 Supervisor: Grzegorz W. Deptuch 2 1 Dipartimento di Elettronica, Informazione e Bioingegneria, Politecnico di Milano, Italy 2 Particle Physics Division ASIC Development group, Fermilab, Illinois, USA September 24 th , 2014 September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 1 / 16
Summary Introduction 1 Preliminary analyses 2 Autocorrelation analyses 3 Future developments 4 September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 2 / 16
Introduction Training program The intern will be involved in the testing process of a new kind of readout circuit (VIPIC chip) designed in a novel 3D-IC technology by the Fermilab Application Specific Integrated Circuit (ASIC) Development group Tasks The intern will learn the structure and the working principle of the new Vertically Integrated Photon Imaging Chip (VIPIC) The intern will develop Matlab code in order to do in depth analyses of the data acquired with the VIPIC chip at the Argonne National Laboratory (ANL) Synchrotron By means of the analyses the intern will infere if the chip is working correctly and so will help the ASIC group in its improvement After the basic analyses the intern will collaborate with ANL researchers in order to perform the X-Ray Photon Correlation Spectroscopy (XPCS) using the acquired data September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 3 / 16
Introduction VIPIC basic block scheme Counting Readout 5-bit counter 256 5-bit counter group 1 serializer LVDS pixel 1 4096 DIODES 5-bit counter 5-bit counter 256 pixel 2 group 2 serializer LVDS 256 5-bit counter group 16 serializer LVDS 5-bit counter pixel 4096 September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 4 / 16
Introduction VIPIC 3D structure September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 5 / 16
Preliminary analyses Working principle 1 24 2 153 ns 23 3 22 4 21 5 Internal clock t 20 6 APS clock 19 7 t 18 8 Photons sequence t 153 ns 17 9 16 10 15 11 14 12 13 Electrons and photons are sorted in 24 bunches. In order to correctly sort them it is mandatory to have a correct triggering mechanism September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 6 / 16
Preliminary analyses Module 24 analysis Using the APS clock photons are correctly sorted in the bunches September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 7 / 16
Preliminary analyses Direct beam operating condition Beam line Slit Detector September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 8 / 16
Preliminary analyses Matrix counts 10 20 30 40 50 60 0 10 6 10 10 5 20 10 4 30 10 3 40 10 2 50 10 60 1 Remark Moving further from the center of the slit the light intensity is decreasing whereas some pixel on both left and right borders seems to slightly increase their counts September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 9 / 16
Preliminary analyses Debug of the reset mechanism Each pixel is connected to a counter Counters are reset after being readout During the analysis it was found that some counters have a value bigger than one. This is not consistent with physics Cause Only counters that have been readout are forced to reset whereas the other ones keep on counting. This is a problem if the number of hits per frame is too high. By means of this analysis a new reset mechanism will be designed in other to avoid that problem . September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 10 / 16
Autocorrelation analyses Autocorrelation software Matlab code provided by ANL’s researchers g ( 2 ) ( p i , τ ) = < I ( p i , t ) · I ( p i , t + τ ) > t < I ( p i , t ) > 2 t Implemented algorhitm is based on a logarithmic sampling of time delays followed by a time average in order to reduce the dispersion of g ( 2 ) function for higher delays due to the lack of samples September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 11 / 16
Autocorrelation analyses Autocorrelation – Direct beam Counts sequence 10 g (2) (t) 5 1 0 1 10 2 10 4 10 6 Time delay [ μ s] Autocorrelation of a poissonian sequence � 1 if m>0 g ( 2 ) ( p i , m ) = 1 + 1 if m=0 µ where µ is the mean value of the sequence September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 12 / 16
Autocorrelation analyses Direct beam Poissonian sequence Counts sequence 9 40 10 6 samples mean value 5x10 -4 8 35 mean value 0.001 10 7 samples 7 mean value 0.1 30 10 8 samples 6 10 6 samples mean value 5x10 -4 25 5 20 g (2) g (2) 4 15 3 10 2 5 1 1 0 1 10 10 2 10 3 10 4 10 5 10 6 1 10 10 2 10 3 10 4 10 5 10 6 10 7 10 8 Time delay (number of frames) Time delay (number of frames) Conclusions The bouncing at short time delays seems to be due to the finite length of the counts sequence September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 13 / 16
Autocorrelation analyses Colloid data Data collected illuminating a colloid sample and collecting photons coming from it. During the analysis it was found that some data have a wrong header 10 20 30 40 50 60 10 20 30 40 50 60 0 0 10 5 10 5 10 10 10 4 10 4 20 20 10 3 10 3 30 30 40 40 10 2 10 2 50 50 10 10 60 60 1 1 (a) (b) Conclusions The problem is related only to a specific readout group. Only the header seems to be corrupted September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 14 / 16
Autocorrelation analyses g ( 2 ) – Central anular ring g ( 2 ) was calculated and averaged on an area with quite uniform illumination. The shape is still poissonian increasing the number of frames. 10 20 30 40 50 60 0 1.02 1.01 10 10 6 1 0.99 20 10 5 0.98 30 g (2) 0.97 10 4 0.96 40 10 3 0.95 50 0.94 10 2 0.93 60 10 0.92 1 10 10 2 10 3 10 4 10 5 10 6 10 7 10 8 Delay ( μ s) (a) (b) Conclusions g ( 2 ) needs to be averaged on a smaller area in order not to change statistics September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 15 / 16
Future developments Next steps The autocorrelation analyses will be done on all the files acquired at ANL These analyses will help the designers team to debug errors A new VIPIC chip meant to be connected to a 1 million pixels matrix will be designed September 24 th , 2014 Pietro Peronio (Politecnico di Milano) Testing of the new VIPIC chip 16 / 16
Recommend
More recommend
