2020/03/05 (Thursday)
| Setup | Run 72. DIB 1/3/4 → Stripline 43/41/48 → Cable 3/2/1 → nothing/nothing/accessory_interposer → HV_Conn 1/3/2 → HV 1/2 → DCRCX(4)/DCRC55(1). DCRCs on overnight. Lead/poly shield erected (4 lead walls, 2 poly walls). |
|---|---|
| Overnight: After flash Tower fell to 52.7 mK by 3:45am, then rose a bit (to 53.0 mK) before starting to fall again - another early-morning excursion. . Tower was back down to 52.7 mK before the LN2 fill. | |
| 0815 | Fill LN2 cold trap bath |
| 0845 | Tower/MC/Coldplate = 56.7/41.2/196 mK, Still = 1.49±0.05 K / 0.190 mbar / 5 mW, G1/G2= 514/481 mbar, 1K Pot = 2.19 K / 24 mbar. MB = 33.7% Tower temperature is still rising after the LN2 fill. |
| Use HVscan.py to generate a sequence HVdrift_100V_31mVps_pre10-2_120min.msl. This should bias detector to 100 V after a 2 minute 10% prebias, then drift by 31 mV/s and run for 120 minutes (ending at about 323V). Then flashes 30 seconds and waits 30 minutes. | |
| 1127 | Start a series at HV = 100→323V, rate 31mV/s, 120 minutes, with 10% prebias for 2 minutes, flash at end with 30 min wait. CDMS detector: Driver Gains = 1x, DCRC4/1 QET biases = (na/360/430)/(360/420/440) uA, Front-End Gains = (na/3/3)/(5/5/5) 1000 BOR and 5Hz in-run randoms, trigger on DCRC1pD threshold = 9 Use sequence HVdrift_100V_31mVps_pre10-2_120min.msl |
| 1130 | Tower = 56.4 mK |
| 1133 | Data starts |
| 1135 | Vtest = 1.1 mV. Kepco dial = 95V. Pulses from BatViewer look OK. |
| 1155 | Vtest = 1.3 mV. Kepco dial = 125V. Pulses from BatViewer look OK. |
| 1300 | Vtest = 2.6 mV. Kepco dial = 225V. Pulses from BatViewer look OK. Tower = 56.1 mK |
| 1330 | Vtest = 3.5 mV. Kepco dial = 275V. Pulses from BatViewer look OK. |
| 1352 | Begin LHe transfer @ MB = 26.0% |
| 1428 | End LHe transfer @ MB = 92.4% |
| 1540 | Tower has fallen to 59.7 mK since the flash. That's still kind of high, and it looks like it's slowing down. Meanwhile the circulation pressures and the 1K Pot temperature are significantly lower than before, Still pressure down a little too. Maybe we should try upping the still heat. |
| 1540 | Increase Still Heater 5→8 mW Tower starts cooling a little faster, not dramatic. |
| Swap the HV box leads on the HV power supply, so we can try some negative bias data | |
| 1600 | Tower = 57.7 mK, cooling more quickly now |
| 1615 | Tower = 56.8 mK, cooling is slowing down. May be an OK time to start data. (Gain changes by 3% for each 1 mK change in temperature.) |
| 1618 | Start a series at HV = -250→399V, rate 31mV/s, 80 minutes, with 10% prebias for 2 minutes, flash at end with 30 min wait. CDMS detector: Driver Gains = 1x, DCRC4/1 QET biases = (na/360/430)/(360/420/440) uA, Front-End Gains = (na/3/3)/(5/5/5) 1000 BOR and 5Hz in-run randoms, trigger on DCRC1pD threshold = 9 Use sequence HVdrift_250V_31mVps_pre10-2_80min.msl |
| 1633 | Data starts |
| 1635 | Vtest = -3.1 mV. Kepco dial = 240 V. Pulses from BatViewer look OK. Tower = 56.2 mK |
| 1650 | Vtest = -3.4 mV. Kepco dial = 260 V. Pulses from BatViewer look OK. Tower = 56.1 mK |
| 1650 | Vtest = -3.8 mV. Kepco dial = 285 V. Pulses from BatViewer look OK. Tower = 55.9 mK |
| 92 dumps in the series | |
| 2000 | Fill LN2 cold trap bath. I didn't remove the trap, unlike SNOLAB…. |
- Lab Updates for today's group meeting
- 1K Pot running warmer than usual this run, probably some ice in it but not enough to prevent cooling to base
- base 52-55 mK, not as good as usual but still can take data
- Built the lead shield, which makes it much easier to see the 14keV and 18keV peaks
- Taking drifted HV data. Nominal best noise width (PTOF) is around 34 eV, no clear improvement over Run 70 (before etching)
- One mystery solved: this run and Run 70, relative calibration showed most channels very unbalanced (factors of 7 and 15). Turns out the OF time window was too restrictive for most channels. This has been corrected. (Should reprocess Run 70.)
- One outstanding mystery: when we estimate Luke gain from calibrated data, the implied bias is much less than what we calculate the effective bias at the detector should be – from applied HV and corrections due to leakage current (<1%), detector charge buildup during biasing (~10%), and loss across the gap (~33%). The apparent Luke gain suggests a bias only 40% of what we calculate it should be. Nominal calculation assumes 1mm gap, it's actually smaller so bias should be even higher. It would make sense if the gap were 3 mm but no way it's that big. Calculation uses Ge dielectric constant of 16.2, it may be lower at sub-K temperature but that would predict a somewhat higher bias.
- Currently scanning over biases (positive polarity, will switch to negative later).
- After the bias scan data we'll try moving the source.