| | Current setup: iZIP mode, DIB1/2/3 → cables → 3.2 MHz filters → DCRC55/33/50. |
| | Overnight: Taking background data (bias scan) |
| 0805 | Tower/MC(H35)/MC sensor/Cold Plate = 31.0/25.4/41.2/157 mK, Still = 1.36 K / 0.10 mbar, G1/G2 = 115/100 mbar, 1K Pot = 1.65 K / 8.5 mbar, Main bath = 69.2% |
| | 1K Pot pressure is still a bit low but it has been stable overnight (actually increased a little). Will keep an eye on it, if it gets much lower should open needle valve a little more. |
| 0810 | Fill LN cold trap Dewar |
| 1115 | In the overnight bias scan data, the higher biases are probably plagued with breakdown problems, causing high noise in the phonons. To investigate, I watched DIB1 pulses while changing biases. Starting at 0/0 V, increase to -8/0 V: no problem. Increase to -8/+8 V: breakdown symptoms - poor pulse shapes and elevated baselines (-0.78 → 0.5 V). The symptoms fall in severity after a few minutes but there is still clear shelving in low-amplitude pulses and elevated baselines (-0.5 V). Then increase to -8/+14 V for a few seconds, then back to -8/+8 V: this removes all breakdown symptoms. |
| 1120 | Repeat above for +/- polarity. Breakdown symptoms are seen when the +8V is set: so it seems on either side breakdown is much more significant for positive voltages (this was observed before). Breakdown symptoms are more mild this time (also observed before that Side 2 positive is worse than Side 1 positive). Again setting briefly to -8/+14 V and back to -8/+8 V seems to completely remove breakdown symptoms. |
| 1137 | After flashing start background data of 1000 randoms at -/+8V. Wait, stop data because Qbias was already at -/+8V when we started, and we want to test what the data looks like going from 0V to -/+8V. |
| 1139 | Set Qbias to 0V. Start data again – background, 1000 randoms at -/+8V, FG=5/2/5, OG=3, 2000 events/file, phonon thresholds set at values from yesterday. |
| 1144 | Start 1000 randoms, background, -/+6V with otherwise same settings as above. |
| 1150 | Start 1000 randoms, background, +/-8V with otherwise same settings as above. |
| 1155 | Start 1000 randoms, background, +/-6V with otherwise same settings as above. |
| 1201 | Set DIB2 Qbias to 14V, then start 1000 randoms, background, -/+8V with otherwise same settings as above. |
| 1206 | Set DIB1 to 14V, then start 1000 randoms, background, +/-8V with otherwise same settings as above. |
| 1212 | Set DIB2 Qbias to 14V, then start 1000 randoms, background, -/+6V with otherwise same settings as above. |
| 1217 | Set DIB1 to 14V, then start 1000 randoms, background, +/-6V with otherwise same settings as above. |
| | PSDs from above tests suggest that overshooting the positive voltage improves noise for +8/-8 V and -8/+8 V. Only slight improvement if any for -6/+6 and +6/-6. However overnight noise is much worse than in any of these tests. The difference must be flashing: running a high bias after flashing seems to make breakdown worse. Tests below to see how bad noise is without overshooting and after a flash: |
| 1224 | Flash 30s. |
| 1436 | Flash 30s. |
| 1501 | Start 1000 randoms, background, -/+6V no overshoot with otherwise same settings as above. |
| 1509 | Stop data. Flash 30s. |
| 1530 | Start 1000 randoms, background, -/+4V no overshoot with otherwise same settings as above. |
| 1538 | Stop data. Flash 30s. |
| 1644 | Start background data at -/+4V, 15 cycles of 30min series, FG=5/2/5, OG=3, with 1000 randoms, 2000 events/file, 30s flash and 25min cooldown in between. Use same thresholds as yesterday. |
| 1645 | Main bath 54.3% |
| 1708 | Fill LN2 cold trap full. |