2019/06/17 (Monday)
| Setup | DIB1/3/4 → 3.2 MHz filters 2/3/1 → cable 3/1/2 → gender changer ?/3/2 → HVConn 1/3/2 → HV 1/2 → DCRCX(4)/DCRC55(1) DCRC33(2) → 1kV power supply → HV input Three walls of the lead/poly shield are up. |
|---|---|
| 0910 | Tower/MC/Coldplate = 41.2/30.8/198 mK, Still = 1.3 K / 0.090 mbar, G1/G2 = 151/143 mbar, 1K Pot = 1.62 K / 22 mbar. MB = 69.7% |
| 1000 | Refill LN2 Dewar and restart Ebox purge |
| 1005 | Fill LN2 cold trap bath |
| 1030 | Move PuBe source from Tate to PAN 443. |
| 1210 | With HV at 0V setting, we have 1.8 V applied. This calibration has been drifting and continues to do so. Why?! Look in the ODB, don't change anything, but suddenly HV=0.01 V! Notice DCRC2 has been set to RevC.1 again somehow. Change it back. Not sure what's going on here but HV looks okay again. |
| 1215 | Insert homemade low-pass filter between HV multimeter and HV board. It initially has ~1V across it, but shorting the filter capacitor takes it back to ~0V. |
| 1230 | Channels look good except DCRC4 pB which keeps railing low. Tried relocking and snapping numerous times. |
| 1240 | Connect test signal to FB line and output to pre-amp in open-loop mode. Should be able to see the SQUID modulation curve like this. On DRCR4 pB it's still railed. On DCRC4 pC, the modulation curve is a complete mess, even though this channel showed pulses just fine. |
| 1242 | Zap SQUIDs. Tower rises to 46.9 mK before falling. |
| 1255 | After some coaxing, get all channels working again. |
| 1300 | Start a 60 min series at HV = 0 V, with low-pass RC filter CDMS detector: Driver Gains = 1x, DCRC4/1 QET biases = (490/400/490)/(420/480/450) uA, Front-End Gains = (2/2/3)/(5/5/5) 1000 BOR, 5Hz InRun, and 1000 EOR randoms, trigger on DCRC1pB/C/D threshold = 6/6/6. |
| 1409 | Start a 90 min series at HV = 180→400V, rate 2.418V/min, with low-pass RC filter with 10% prebias for 2 minutes, flash at end with 30 min wait. CDMS detector: Driver Gains = 1x, DCRC4/1 QET biases = (490/400/490)/(420/480/450) uA, Front-End Gains = (2/2/3)/(5/5/5) 1000 BOR, 5Hz InRun, and 1000 EOR randoms, trigger on DCRC1pB/C/D threshold = 6/6/6. Use sequence HVdrift_180V_cal0612_2p418VPM_prebias_10_2_dur_90min_flashDCRC_1_4.msl |
- First look at data with the HV filter does not suggest any improvement in noise.
| 1621 | Begin strong neutron activation of detector by hanging PuBe source in small bucket inside NW corner of shield. |
|---|---|
| 1647 | End strong neutron activation by returning PuBe source to its paraffin-filled transfer tub. Almost exactly 26 minutes of strong activation |
| notes: Tower temperature fell 42.3→41.1 mK after removing source. When source was present, took a quick look at the channels - clogged with pileup. | |
| 1650 | Return PuBe source to storage in PAN443. |
| 1655 | Check channels, all look OK. |
| We will need to re-address the drift rate. Activation will increase it, but maybe not by much. Apart from that, the weekend data looks like it had downward-sloped gains, so the drift rate was too low, by about 0.08 V/min. So let's boost it by that and see if it works. | |
| 1700 | Build sequence for 180V→400V with drift 2.52 V/min = 0.042 V/s, 87 minutes |
[fritts@jiajin /home/hep/cdmssoft/scdmsDAQ/MidasDAQ/online/scripts/pysequencer]: python HVscan.py -iDCRC 2 -HVs 180 -HVcalFile HV_calibrations/HV_DCRC2_control_cals_440V_190612.txt -HVdriftRate 0.045 -HVpreBias 10/2 -tSeries 87 -DCRCs2Flash 1,4 > ../../sequencer/HVdrift_180V_cal0612_2p52VPM_prebias_10_2_dur_87min_flashDCRC_1_4.msl
| 1708 | Start an 87 min series at HV = 180→400V, rate 2.52V/min, with low-pass RC filter with 10% prebias for 2 minutes, flash at end with 30 min wait. CDMS detector: Driver Gains = 1x, DCRC4/1 QET biases = (490/400/490)/(420/480/450) uA, Front-End Gains = (2/2/3)/(5/5/5) 1000 BOR, 5Hz InRun, and 1000 EOR randoms, trigger on DCRC1pB/C/D threshold = 6/6/6. Use sequence HVdrift_180V_cal0612_2p52VPM_prebias_10_2_dur_87min_flashDCRC_1_4.msl |
|---|
- Where are the activation peaks? First look shows no sign, though Am peaks are still prominent
- We tried to match the 1keV rate to the 14keV rate. But we neglected write efficiency. Maybe it's very low.
- Also we need to compete with the uniform background, not just the localized Am events. So we should look at this differently.
- In Run 66 we activated for 4 hours over two days. And the rate was not extreme.
- Event rate in recent data is 70 Hz, pretty high.
- Maybe let's try significantly higher thresholds, reduce the trigger rate.
| 2118 | Start an 87 min series at HV = 180→400V, rate 2.52V/min, with low-pass RC filter with 10% prebias for 2 minutes, flash at end with 30 min wait. CDMS detector: Driver Gains = 1x, DCRC4/1 QET biases = (490/400/490)/(420/480/450) uA, Front-End Gains = (2/2/3)/(5/5/5) 1000 BOR, 5Hz InRun, and 1000 EOR randoms, trigger on DCRC1pB/C/D threshold = 15/15/15. Use sequence HVdrift_180V_cal0612_2p52VPM_prebias_10_2_dur_87min_flashDCRC_1_4.msl |
|---|
- It's still difficult to see activation lines in this data.
- Estimating how long we should activate to make the 10keV line 10X the background: 138 min total (or 112 min in addition to the 26 min from today)
- Calculating event rate, and estimating write efficiency from write rate of in-run randoms (set to 5Hz):
- with trig thresholds 6/6/6 (standard): written event rate = 70 Hz, write efficiency 25%→23%, goes down over course of series, with dips corresponding to high noise event periods probably
- with trig thresholds 15/15/15: written event rate 58 Hz, write efficiency 90%
- The original exposure time estimate was based on our standard data, which we now see seems to have only a 25% efficiency. So our 26-minute T_exposure estimate should have been multiplied by 4, because the 14keV rate is 4X higher than we though.