HV BOARD. Claudio Arnaboldi Tito Bellunato Gianluigi Pessina INFN, Milano. Thierry Gys, Didier Piedigrossi CERN. LHCb, 20/01/2005 Genova 1

Transcription

1 HV BOARD Claudio Arnaboldi Tito Bellunato Gianluigi Pessina INFN, Milano Thierry Gys, Didier Piedigrossi CERN LHCb, 20/01/2005 Genova 1

2 HV distribution Half Column (8 HPDs) Half Column (8 HPDs) 20 KV Input HPD 2 HPD 1 FILTER + DIVIDERS HPD 8 HPD 7 VOLTAGE MONITOR HPD 10 HPD 9 VOLTAGE MONITOR HPD 16 HPD 15 FILTER + DIVIDERS 20 KV Input PCB type A PCB type B PCB type B PCB type A 2 printed circuit boards, PCB: one for divider and splitting, the other for voltage monitoring. 3x10 GΩ resistors allow to monitor the 3 HV voltage of a semicolumn. A 300 MΩ divider (70 20 KV) generates the 3 needed HV voltages (20 KV, 19.7 KV and 16.4 KV), starting from the 20 KV input. LHCb, 20/01/2005 Genova 2

3 HV distribution Half Column (8 HPDs) 20 KV Input HPD 2 HPD 1 FILTER + DIVIDERS Now it is time to think also at the type of supply voltage to adopt. Is it better a few number of supplies having large current capability, or many units with a smaller current capability? PCB type A LHCb, 20/01/2005 Genova 3

4 HV distribution 20 KV Splitter 20 KV 4.57 M Ω 19.7 KV 50 M Ω 16.4 KV 250 M Ω 1 GΩ 1 GΩ 1 GΩ Protecting resistors One 1 GΩ-resistor set for every HPD We need only one HV connector per column or semicolumn. In principle different column may be supplied by the same HV power supply. LHCb, 20/01/2005 Genova 4

5 HV Monitoring 20 KV 19.7 KV 16.4 KV 10 GΩ 3 MΩ 10 GΩ 3 MΩ 10 GΩ 3 MΩ 10 GΩ= 1 % and less than 1 ppm/v of voltage coefficient. 100 ppm/ C stability. To the ADC (10 V range) The voltage monitoring does not load the line since the very large value of the resistors used for the check. LHCb, 20/01/2005 Genova 5

6 HV Monitoring To improve the accuracy of the reading, if 100 ppm/ C is not enough, a splitter may be used, at additional cost. 10 GΩ 3 MΩ Which kind of reading we would like? Differential and Optical coupled? Only Differential? LHCb, 20/01/2005 Genova 6

7 HV boards, lateral view The new mechanical supports will permit more space to allow the complete covering, with silicon rubber, of all the components that populate the board. HPDs HV cables HV board PCB HV board Board to board connections Cooling plate side Mec. frame Silicon rubber coating LHCb, 20/01/2005 Genova 7

8 The Latest mechanical support modification from Genoa The new version of the mechanical supports will have new characteristics. For what concerns the HV boards a large amount of space will be made available. LHCb, 20/01/2005 Genova 8

9 Mechanical details The larger space made available will allow easily to fit the dimension of the HV resistors and the divider and the protective rubber. HV LV L0 LHCb, 20/01/2005 Genova 9

10 Specification of PCBs used 2 mm PCB cross-section Copper FR4 0.8 mm The PCB measures 165 mm x 90 mm x 2 mm. The PCB has four layers. The HV tracks are buried under 0,8 mm of FR4 fiber glass to obtain very large insulation (FR4 has a dielectric strength of 80 KV/mm). We have provided cuts around the soldering pins. This way the upper and lower rubber coating join. The solution solved the problem of the shielding of the electric field parallel to the PCB surface, since the adhesion of the rubber is not able to assure perpetual shielding. LHCb, 20/01/2005 Genova 10

11 The principle of operation of the cuts: board with no cut Pin short circuited The Electric field parallel to the surface is not well shielded, depending on the adhesion of the silicon on the pcb LHCb, 20/01/2005 Genova 11

12 The principle of operation of the cuts: board with cuts Pin short circuited The cut is about 2 mm wide. Now the electric field parallel to the pcb is shielded by the silicon (having a dielectric strength of about 17 KV/mm) even in this direction, avoiding any discharge. LHCb, 20/01/2005 Genova 12

13 HV boards pictures Cooling plate side HPD cables side Even if not fundamental we will provide a conductive sheet embedded in the Silicon on both the bottom and top of the boards. LHCb, 20/01/2005 Genova 13

14 HV boards preparation The populated PCB. The populated PCB in the mould. The components and the cables are mounted on the PCB. The board is cleaned to remove moisture that degrades silicon rubber adhesion and reduces surface parasitic resistances. The board is closed in the mould. The silicon rubber is prepared and degassed in vacuum (to less than 0.5 mbar) for 10 min. The mould is filled with the silicon rubber and degassed again in vacuum. LHCb, 20/01/2005 Genova 14

15 HV boards preparation The board in the mould is pre-cured at 70 C. The mould is opened and the board removed. The board is finally cured (annealed) at 70 C for 4 days. We have observed that for the large thickness of rubber adopted (about 20 mm) any residual element that may give leakage is cured after a long period at moderate temperature. LHCb, 20/01/2005 Genova 15

16 The new HV cables Actual cable New proposed cable We searched for a new type of HV cable for the HV boards connections because the, so far, used ones are out of production. External diameter 3,81 mm (the old one is about 3.3 mm), internal copper 22 AWG. Working voltage 24 KV, breakdown voltage 48 KV. Composition: PVC protection (red), polyethylene insulator (transparent) and a conductive polyethylene as a Korona-guard (black). LHCb, 20/01/2005 Genova 16

17 Wiring connection The joining between the cables coming from the HPD and the cables coming from the HV boards may be made with the Didier method consisting in covering the soldered parts with 2 special term-retractable materials. LHCb, 20/01/2005 Genova 17

18 The test beam in November LHCb, 20/01/2005 Genova 18

19 News: results from radiation investigation We have tried to verify the rad-hard quality of the HV boards. The 1 GΩ HV resistors, The 10 GΩ HV resistors, The HV filtering capacitances, The HV board covered with 2 different type of silicon rubbers. Have been subjected to a fluence of: neutrons/cm 2 LHCb, 20/01/2005 Genova 19

20 Resistors Samples irradiated Voltage divider 10 GΩ Resistors 1 GΩ Resistors LHCb, 20/01/2005 Genova 20

21 The behavior of the 1 GΩ resistors Sample # Value Before-Irradiation (GΩ) Value After-Irradiation (GΩ) R1A 0,957 0,962 R2A 0,937 0,941 Results are compatible with the error given by the 2 different instruments used, the former in Milano Lab, the latter at CERN-Thierry Lab. The 1 GΩ resistors seem to be adequate for LHCb. LHCb, 20/01/2005 Genova 21

22 The behavior of the 10 GΩ resistors Sample # Value Before-Irradiation (GΩ) Value After-Irradiation (GΩ) R1B 9,978 9,957 R2B 9,978 9,961 Results are compatible with the error given by the 2 different instruments used, the former in Milano Lab, the latter at CERN-Thierry Lab. The 10 GΩ resistors seem to be adequate for LHCb. LHCb, 20/01/2005 Genova 22

23 Sample # S1.ABC S1.BC S1.C S1.B S2.ABC Value After- Irradiation (GΩ) Value Before- Irradiation (GΩ) 0,3040 0,2995 0,2495 N.A. 0,3030 The divider resistor 0,306 0,301 0,251 4,55 MΩ 0,304 A B C Results are compatible with the error given by the 2 different instruments used, the former in Milano Lab, the latter at CERN- Thierry Lab. S2.BC 0,2987 0,299 S2.C S2.B 0,2485 N.A. 0,251 4,54 MΩ The divider resistors seem to be adequate for LHCb. LHCb, 20/01/2005 Genova 23

24 Capacitances irradiated High voltage capacitances LHCb, 20/01/2005 Genova 24

25 Sample # The divider resistor Value Before-Irradiation (GΩ) Value After-Irradiation (GΩ) 1 KHz 10 KHz 1 MHz 1 KHz 10 KHz 0,1 MHz C1 805 p 791 p 772 p 810 p 805 p 796 p C2 813 p 795 p 774 p 805 p 799 p 788 p C3 2,030 n 1,990 n 1,936 n 1,951 n 1,942 n 1,925 n C4 1,878 n 1,856 n 1,824 n 1,841 n 1,831 n 1,814 n C5 9,730 n 9,565 n 9,473 n 10, 83 n 10,80 n 10,72 n C6 9,454 n 9,475 n 9,554 n 10,86 n 10,75 n 10,62 n Results are compatible with the error given by the 2 different instruments used, the former in Milano Lab, the latter at CERN-Thierry Lab. The filtering capacitances seem to be adequate for LHCb. LHCb, 20/01/2005 Genova 25

26 HV leakage test Current Monitor The test is under the worst condition of having 3 HPD short circuited to GND. -20 KV The current is expected to be negligible if all works fine. The 1 GΩ resistors are not assembled to avoid current absorption. LHCb, 20/01/2005 Genova 26

27 Leakage from 2 boards before irradiation Board #1 This board was covered with the so called DC Sylgard 184. The leakage current is very small, at the limit of the instrumentation set-up. 10 Board covered by Sylgard Note that the filtering capacitances are mounted on this board. Average Ambient conditions: 23±1 C and 50±5 % relative humidity. Leakage current (na) Days LHCb, 20/01/2005 Genova High Voltage (KV)

28 Leakage from 2 boards before irradiation Board #2 This board was covered with the so called DC The leakage current is very small, at the limit of the instrumentation set-up. In this test the HV voltage was cycled any 2 hours. Leakage current (na) Board covered by DC LHCb, 20/01/2005 Genova Days High Voltage (KV)

29 Leakage from 2 boards after irradiation Board #1 This board was covered with the so called DC Sylgard 184. The leakage current is very small, at the limit of the instrumentation set-up. Neutron irradiation gave no effects on the working condition of this board and its filtering capacitances. Average Ambient conditions: 20±2 C and 50±8 % relative humidity. Leakage current (na) Board covered by Sylgard 184 (after neutron irradiation) LHCb, 20/01/2005 Genova Days High Voltage (KV)

30 Leakage from 2 boards before irradiation Board #2 This board was covered with the so called DC The leakage current is very small, at the limit of the instrumentation set-up. Neutron irradiation gave no effects on the working condition of this board. Average Ambient conditions: 20±2 C and 50±8 % relative humidity. Leakage current (na) Board covered by DC (after neutron irradiation) Days LHCb, 20/01/2005 Genova High Voltage (KV)

31 Future prospective So far We have to make some discussion on the selection of the supply voltages. We need also to select the HV connectors. We have to decide which accuracy we would like to obtain in the HV monitoring. Which kind of reading we would like in the HV monitoring? LHCb, 20/01/2005 Genova 31