2020/02/26 (Wednesday)
| 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. No shield. |
|---|---|
| overnight: tower temperature falling at constant rate, 0.05 mK/hour, 1.1 mK/day | |
| 0900 | As seen yesterday, HV not quite at 0V, actually -0.25V. Also Vtest unstable overnight, frequently jumping around -20 to 20 mV. Disconnect one HV lead from the Kepco voltage supply and short the leads together. Vtest is now stable, reading 0.06 mV. |
| 0910 | Tower/MC/Coldplate = 55.4/32.9/196 mK, Still = 1.64±0.05 K/ 0.220 mbar / 0 mW, G1/G2= 345/310 mbar, 1K Pot = 2.22 K / 36 mbar. MB = 31.0% |
| 0930 | Fill LN2 cold trap bath |
| 1015 | Close 1K Pot needle valve narrower, 40° → 30°. Opening it wider yesterday seemed to make the tower a little warmer. 1K Pot settling at 2.18K / 26 mbar. Circulation pressures falling a little. Nothing else seems strongly affected. |
Analysis of yesterday's data (180V drifted):
- The amplitude of the 60keV line dropped by 16% over the 95-minute run. This suggests we need to drift the bias faster: 180V*0.16/95min = 0.30 V/min = 5 mV/s. Yesterday we drifted at 41.5 mV/s so let's try 42 mV/s.
- With a time-corrected PTOF and a high-pF cut you can just barely see the 18 keV line and a hint of the 14 keV line. Comparing calibrations in this 180V data to 0V data, it looks like a Luke gain of 12.0. This implies an effective bias of only 33V, 18% of the applied 180V - why so low?
- The relationship between the fraction of the applied bias seen at the detector and the electrode-detector gap is Vdet/Velectrode = h/eps / (h/eps + gap), where h is detector height and eps is detector dielectric constant, presumably 16.2 for Ge. Given that, 18% implies a 9mm gap. That certainly is not correct.
- Noise sigma from randoms appears to be about 56 eV (increases a little over time, from around 50 keV to around 60 keV)
| 1030 | Use HVscan.py to generate a sequence HVdrift_250V_42mVps_pre10-2_60min.msl. This should bias detector to 250 V after a 2 minute 10% prebias, then drift by 42 mV/s and run for 60 minutes. Then flashes 30 seconds and waits 30 minutes. (This time I correctly set it to flash the DCRC1 LEDs.) Drift rate adjusted from yesterday's test. |
|---|---|
| 1034 | Start a series at HV = 250→401V, rate 42mV/s, 60 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_42mVps_pre10-2_60min.msl |
| 1046 | Prebias wait ends, now reducing voltage to 250. (Reads a little lower on the Kepco) |
| 1047 | Data starts |
| 1050 | Vtest ~ 4 mV with some unstable excursions. Kepco = 240 V, pulses on BatViewer look OK. |
| checking with the Fluke meter, the Kepco voltage seems to be a little higher than the Kepco dial reading. | |
| 1055 | Tower = 54.5 mK |
| 1115 | Vtest = 4 mV with some unstable excursions. Kepco = 300 V, pulses on BatViewer look OK. |
| 1125 | Vtest = 4.3 mV with some unstable excursions. Kepco = 320 V, pulses on BatViewer look OK. |
| 1135 | Vtest = 4.8 mV and stable. Kepco = 345 V, pulses on BatViewer look OK. |
| 1205 | Data series has ended, HV is dialing down. 65 dumps, so about a 5 minute total lag in the sequencer. |
| 1205 | Tower peaks at 800 mK during flash. Check Kepco with Fluke: 0.3 V. Disconnect one of the HV leads and connect it to the other to shore the HV input to the HV box. |
| 1234 | Started LHe transfer at MB = 26.2%. |
| 1314 | End LHe transfer @ MB = 91.4% |
| 1400 | Looked at today's data, looks like the drift rate was a little too slow, should bump it up by 3 mV/s |
| 1400 | Use HVscan.py to generate a sequence HVdrift_250V_45mVps_pre10-2_60min.msl. This should bias detector to 250 V after a 2 minute 10% prebias, then drift by 45 mV/s and run for 60 minutes (final HV = 412 V). Then flashes 30 seconds and waits 30 minutes. Drift rate adjusted from this morning's data series. |
| 1409 | Start a series at HV = 250→412V, rate 45mV/s, 60 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_45mVps_pre10-2_60min.msl |
| 1421 | Data starts |
| 1435 | Vtest = 3.3 mV. Kepco = 270 V. DCRC4pB railed… Adjust DCRC4pB driver offset to unrail it. |
| 1440 | Tower = 53.9 mK |
| 1440 | All channels look OK now. DCRC4pB was railed for first 15 dumps |
| 1450 | Vtest = 3.9 mV. Kepco = 305 V, pulses on BatViewer look OK. |
| 1520 | Vtest = 4.8 mV. Kepco = 370 V, pulses on BatViewer look OK. |
| 1527 | Data series ends with 65 dumps |
| 1550 | Tower peaked at 800 mK during flash. Check Kepco with Fluke: 0.23 V. Disconnect one of the HV leads and connect it to the other to shore the HV input to the HV box. |
| 1610 | 30 minutes after the flash, Tower = 56.3 mK and falling. 2 mK or so above base. Is 30 minutes a long enough wait before starting another series? Maybe should be longer. |
| 1640 | 60 minutes after the flash, Tower = 54.9 mK and falling. 1.2 mK above the pre-flash level. Seems like we have to wait at least this long. (Would it be better to do a shorter flash?) |
| 1705 | Use HVscan.py to generate a sequence HVdrift_300V_46mVps_pre10-2_55min.msl. This should bias detector to 3000 V after a 2 minute 10% prebias, then drift by 46 mV/s and run for 55 minutes (end HV = 452V). Then flashes 30 seconds and waits 30 minutes. Drift rate slightly adjusted. |
| 1714 | Start a series at HV = 300→452V, rate 46mV/s, 55 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_45mVps_pre10-2_60min.msl |
| 1730 | Fill LN2 cold trap bath |
| 1729 | Data starts |
| 1730 | Tower = 54.2 mK |
| 1732 | Vtest = 3.8 mV. Kepco = 290 V, pulses on BatViewer look OK. |
| 1740 | Vtest = 4.0 mV with some unstable excursions. Kepco = 310 V, pulses on BatViewer look OK. Tower = 53.9 mK |