Integration of GET system for SπRIT-TPC

T. Isobe,∗1 M. Kurata-Nishimura,∗1 H. Baba,∗1 N. Nakatsuka,∗2 T. Murakami,∗2 W.G. Lynch,∗3 J. Barney,∗3

J. Estee,∗3 S. Tangwancharoen,∗3 G. Jhang,∗4 W. Powell,∗5 and Y. Ayyad,∗6 for the SπRIT Collaboration

A Time Projection Chamber (TPC) has been pro-duced as the main detector of the SAMURAI-SπRITproject for the study of the nuclear equation of stateusing heavy ion collision experiments1). As the read-out system for the SπRIT-TPC, we have integrated theGeneric Electronics for TPC (GET) system2), whichwas developed mainly by a France and USA collabo-ration3). For the integration of GET electronics, de-velopment of interfaces in terms of both hardware andsoftware are necessary.

Since the GET system was developed for general us-age, the interface of the GET electronics to a TPCdepends on the specifications of each detector, such asconnector types and the signal characteristics of thedetector, and this interface has to be developed foreach project. We call such an interface a ZAP board,which serves as an adapting connector between theGET electronics and the TPC, as well as providingprotection for the electronics. Thus GET electronicscan be mounted on any TPC by just making the ZAPboard. There are several requirements for designingthe ZAP board. In the case of SπRIT-TPC, the boardwas designed to fit in the space on the TPC which issupposed to be installed in the SAMURAI chamber,to reduce the noise in order to increase the dynamicrange, and to reduce the distortion of gain among thedifferent channels. It is especially important to reducenoise since it strongly affects the electrical treatmentsof a signal. Here the electrical treatments of each signalare: 1. hit pattern register which is made by the dis-criminator on each signal after the pre-amplification,2. internal trigger which is made with the hit patternregister, and 3. zero-suppression of digitized signal.The quality of the internal trigger strongly depends onnoise level, which is strongly affected by the design ofthe ZAP board.

To satisfy our requirements, prototypes of sev-eral types of interface boards were made andthe above requirements were checked quantitativelyby using the minimum GET system composed ofCoBo+AsAd+ZAP boards. The AsAd board is forthe amplification and digitization of an analog sig-nal3). The AsAd must be installed near the TPC inorder to reduce noise. The CoBo board is for handlingthe trigger, controlling the AsAd and managing data3).

∗1 RIKEN Nishina Center∗2 Department of Physics, Kyoto University∗3 National Superconducting Cyclotron Laboratory, Michigan

State University∗4 Department of Physics, Korea University∗5 Department of Physics, University of Liverpool∗6 Research Center for Nuclear Physics, Osaka University

To fit the AsAd board on the SπRIT-TPC within thelimited vertical space of ∼17 cm, a prototype flexibleZAP made of a thin Cu plate (10 µm) and polymide(12.5 µm) was first made. The noise level of the flex-ible board, which can be estimated through RMS ofpedestal, is not small. This is most likely due to thelarge capacitor made by the noise shield and signalline (50∼100 pF/ch). The capacitance of each signalline can be roughly calculated as a plane parallel platecapacitor. Ultimately it was decided to use a ZAPboard made of conventional rigid electric board whichhas a short signal line and small capacitor on each sig-nal line of 10∼20 pF/ch. The achieved noise level is4 ADC under the configuration of a dynamic range of12 bit, 120 fC and a shaping time of 233 nsec. It canbe reduced to be 2∼3 ADC after the subtraction of afixed noise pattern (FPN) line which is not connectedto TPC pads. Without this ZAP, the GET electronicsexhibit and average noise level of 3 ADC. This noiselevel is small enough for the SπRIT experiment, and sothe production of ZAP has started and the mountingof GET electronics on the TPC is ongoing.

The user of the GET system is not only responsiblefor the interface to their TPC, but they must also in-tegrate or develop the DAQ. For the SπRIT project,NARVAL4) is planned to be employed as the DAQ sys-tem. By employing NARVAL, large amounts of dataof more than 100 MByte/sec from SπRIT-TPC can behandled. The development of an interface to send datafrom Babirl to NARVAL is being developed so that rawdata from Babirl can be merged with data from GETsystem by using NARVAL.

We have performed a test experiment using the GETsystem on another TPC with the final ZAP at HIMACduring November 2014. NARVAL was not used in thistest experiment. The analysis result of the test exper-iment is reported by G. Jhang in this APR. In the testexperiment, we took benchmark test data of the GETsystem. We read out 756 channels with a samplingrate of 25 MHz and 256 time buckets. A partial read-out mode using the hit pattern register was used toincrease the data acquisition rate as much as possible.A data acquisition rate of 600 Hz and a data rate of1 GByte/min were achieved, which is sufficient for thefirst experiment of SπRIT project.

References1) S. Rebecca et al.: Physics Procedia 37, 1799 (2012).2) T. Isobe et al.: RIKEN Accel. Prog. Rep. 46, 151

(2014).3) E. Pollacco et al.: Physics Procedia 37, 1799 (2012).4) X. Grave.: Proc. 14th IEEE NPSS RT Conf, 65(2005).

Performance evaluation of GET readout electronics for heavy ioncollision experiments at RIBF

G. Jhang,∗1,∗2 T. Isobe,∗2 M. Kurata-Nishimura,∗2 T. Murakami,∗3 M. B. Tsang,∗4 S. Hasegawa,∗5 K. Hosomi,∗5

S. Hwang,∗5 H. Sugimura,∗5 J. Brzychczyk,∗6 B. Hong,∗1 P. Lasko,∗6 J. Lukasik,∗7 W. G. Lynch,∗4

P. Pawlowski,∗7 Z. Sosin,∗6 and S. Tangwancharoen∗4 for the SπRIT Collaboration

GET(Generic Electronics for TPCs)1) will be usedas the front-end readout electronics for the newly con-structed SπRIT-TPC(SAMURAI pion ReconstructionIon-Tracker Time Projection Chamber)2). The GETelectronics has been recently developed by the GETcollaboration for the particle and nuclear physics ex-periments and its performance is yet to be evaluated.We performed a test experiment at the Heavy Ion Med-ical Accelerator in Chiba(HIMAC) facility using theBRAHMS-TPC(TPM1)3) to evaluate the performanceof the GET electronics. The TPC has the multi-wireconfiguration similar to the SπRIT-TPC: anode wiresfor signal multiplication, gating grid wires to reducethe background noise.

We used 300 AMeV 132Xe beam with 500 mg/cm2

CsI target surrounded with the multiplicity trigger ar-ray of 60 scintillator plastics. To evaluate the resolu-tion in both wire and drift directions, in the first setupthe BRAHMS-TPC(TPM1) was placed at 60 degreesoff the beam axis and we took data with different sam-pling rate of 5, 10, 25, 50, 100 and 200 MHz and shap-ing time of 70, 232, 502 and 1014 ns. Due to breakdownof the µ-TCA power supply, only one AsAd board(256channels) is used in this configuration.

In the second setup the BRAHMS-TPC(TPM1) waslocated at 0 deg. We replaced the broken power supplyand took data with three AsAd boards(768 channels)controlled by a full CoBo(Control Board for 4 AsAdboards). We tested the gating grid drivers with beamintensities varying from 102 to 105 pps. We placedfive 3 mm thick Al plates in front of the TPC to stopthe beam and heavy fragments so that only the lightcharged particles with Z < 3 can reach the TPC at theintensity 104 and 106 pps. Finally, with beam intensity104 and 105 pps we took data under the SπRIT-TPCexperimental condition: sampling rate of 25 MHz,shaping time of 232 ns, 256 time buckets(10.24 µs),and zero suppression mode.

Figure 1 shows the preliminary result in the firstsetup. The resolution is presented as a function of the

∗1 Department of Physics, Korea University∗2 RIKEN Nishina Center∗3 Department of Physics, Kyoto University∗4 National Superconducting Cyclotron Laboratory and De-

partment of physics and Astronomy, Michigan State Uni-versity

∗5 Advanced Science Research Center, Japan Atomic EnergyAgency

∗6 Jagiellonian University∗7 Institute of Nuclear Physics PAN

layer numbers, where one layer means a row of 96 padsperpendicular to the beam axis and the BRAHMS-TPC has 12 layers. To obtain the resolution we se-lected three successive layers and calculated the differ-ence between the hit position on the middle layer andthe average position of hits on the first and the thirdlayers assuming that the resolution of each pad is thesame as δx. For example, x = 0 means the resolu-tion is calculated with the layer number 0, 1, and 2by accumulating x1− (x0+x2)/2 values for all events.By the error propagation the RMS value of the dis-tribution, σ, and the resolution, δx, have the relation:σ =

√3/2δx.

Used Layer (x = [x, x+1, x+2])0 2 4 6 8 10

Res

olut

ion

(mm

)

0.2

0.25

0.3

0.35

0.4

0.45

Drift

Wire

Fig. 1. Resolution of the wire and drift direction as a func-

tion of the layer number x, which is the first layer num-

ber among three layers.

Table 1. Resolution from two collaborations

CollaborationDirection

wire (µm) drift (µm)

BRAHMS 310 427SπRIT 217±10 401±22

Comparison between the result from both collabo-rations is summarized in Table 1. Our result is betterbecause the GET electronics is 12 bit ADC while theBRAHMS collaboration used 10 bit ADC.

Further analyses in the second setup are ongoing.

References1) E. Pollacco et al.: Physics Procedia 37 (2012) 1799.2) R. Shane et al.: Nucl. Instr. and Meth. A 784 (2015)

512.3) M. Adamczyk et al.: Nucl. Instr. and Meth. A 499

(2003) 437.

- 204 - - 205 -

Ⅱ-9. Instrumentation RIKEN Accel. Prog. Rep. 48 (2015)RIKEN Accel. Prog. Rep. 48 (2015) Ⅱ-9. Instrumentation

完全版2014_本文.indd 204 15/10/16 17:53