Calibration of temperature sensors for ProtoDUNE-SP M. Antonova, A. - - PowerPoint PPT Presentation

Calibration of temperature sensors for ProtoDUNE-SP

Joint Cryogenic Instrumentation & Slow Controls meeting 18/04/2018

• M. Antonova, A. Cervera, P

. Fernández, M. García, A. Izmaylov, J. Llusar, M. Sorel, P . Novella, P . Bernabeu, J.V. Civera, P . León

IFIC-Valencia

• X. Pons

CERN

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

T-gradient monitor

• High precision (<5 mk) vertical temperature profiling

with sensors cross-calibrated in the lab

• 48 sensors separated 11.8 cm at top and bottom,

and 23.6 cm in the middle area

2

9 m

175 cm

LAr level

bottom cup

7.5 m

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Calibration setup

• Almost converged after many different setups, including a

profesional pressure vessel

• Home made cryostat: a polystyrene box with 12 cm thick walls

and a 3D printed PLA box with two independent volumes

• Sensors are put inside a cylindrical aluminum capsule to minimise

convection and avoid thermal shocks

3

5 cm

PLA Box made in a 3D printer PLA Box inside of the polystyrene box Aluminum capsule with 4 sensors inside Aluminum capsule inside PLA Box 35 cm 10 cm

12 cm

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Sensor support and cable

• Sensors are mounted on a PCB with an IDC-4 connector
• 4-wires readout to avoid lead resistance and effect of connectors
• Four sensors are put together inside the Al capsule using a PLA

support to hold the sensors and keep them at a fix distance in the middle of the capsule

4

PT102 with IDC-4 connector Put 4 sensors as close as possible PLA support for 4 sensors and cables

1.4 cm ~5 cm

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Readout

• An accurate current source for the excitation of

the temperature sensors, implemented by a compact electronic circuit using high a precision voltage reference from Texas Instruments.

• A multiplexing circuit based on an ADG707

Analog Device multiplexer electronic device

• A high resolution and accuracy voltage signal

readout module based on National Instruments NI9238, which has 24 bits resolution over 1 V range

• The current mode of operation averages over

2000 samples taken every second for each sensor

5

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Readout

6

PT100 current source and MUX card ADC MUX controller current source for MUX controller

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Calibration strategy

• Put several sensors as close as possible in LAr such that we can

assume they are at the same temperature

• Compute the temperature offset between them and refer those
• ffsets to a single reference value
• For the sensors calibrated at once (a “set”) we can compute all

possible combination of offsets

• But obviously the 48 sensors cannot be put together so:
• either we use a single sensor as reference for all sets
• or we do different sets and then correlate them cross-calibrating a

sensor from each set

• In fact we will do both, such that we can cross-check (next slide)

7

• Use a sensor as reference. keep that sensor in all sets
• Thus, three sensors are calibrated in each set
• 3 calibration rounds
• In the first round we calibrate 18 sets of 3 sensors

each (4 measurements for each set)

• In 2nd round we cross-calibrate the middle sensors
• f the 1st round (5 sets, 4 meas each)
• In 3rd round we cross-calibrate the middle sensors
• f the 2nd round (1 set, 4 meas)
• This is a total of 18x4+5x4+1= 93 measurements

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Calibration strategy

8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

1st round 2nd round 3rd round

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Single set

1. Cool down the cryostat 2. Put four sensors in the capsule (3 + reference) 3. Close the capsule and cool it down slowly until it reached 98 K 4. Submerge it into LAr and put the PLA cover 5. Fill the cryostat with LAr 6. Record temperatures for about 40’ 7. Take out the capsule and wait until it reaches 200 K 8. Repeat steps 4-7

9

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Results

• Dispersion of the measurements is below 2

mk for the middle sensor (closer to reference) and below 4 mk for outer sensors

• Reproducibility, defined as the RMS of the

mean offset between in the stable region (1000-2000 s) is below 1 mk for most sensors

10

stable region 1000-2000 s

1 2 3 (ref) 4 1 2 4

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

2nd and 3rd sets

11

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

4th and 5th sets

12

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

6th and 7th sets

13

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

8th and 9th sets

14

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

All sets together

15

39616 = 9.98 +- 0.43

sensor ID mean

• ffset

reproducibility

• Check the same sensor (middle one) after a month
• NO changes (< 1 mk) after a month
• Need more checks like this

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Stability over time

16

March 22nd

April 17th

March 22nd + April 17th

• In green sets already calibrated
• We will do the second round for the first 8 sets in the

next few days

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Calibration strategy

17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

1st round 2nd round 3rd round

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Conclusions

• We believe ~1 mk reproducibility can be achieved
• Steel need to understand variations of behaviour from one set to

another

• Improve some elements of the calibration setup, with better

insulation of the sensors

• Monitor pressure and ambient temperature
• There is still room for improvement:
• Improve cryostat
• Modify the sensor PCB such that all 4 sensors can be put even

closer

• Amplify the voltage signal such that we use the full range of the
• ADC. Currently using 20 mV, having a range of 1 V

18