2016/11/02 (Wednesday)
| Setup: DIB1/3/4 → 3.2 MHz filters → cables 1,2,3 → HV board inputs 1,3,2 → HV board outputs 1,2 → DCRC55(1)/50(3) | |
| 0845 | Tower/MC/Coldplate = 31.8/34.6/214 mK, Still = 1.19 K / 120 μbar, G1/G2 = 165/150 mbar, 1K Pot = 1.72 K / 36 mbar. MB = 49.7%. Boiloff rate: closer to slow-mode than fast-mode |
| 0845 | Return Cf source to safe. |
| 0855 | Fill LN2 cold trap bath. |
| 1100 | Begin re-measuring Rp and Rn for DCRC1pA. Bill Page has suggested we not use the calculation method from Monday, in which the assumption was that Rshunt has a parasitic resistance equal to the Rtes parasitic resistance. He says that there are fewer wirebonds on the Rshunt line so you can expect lower parasitic resistance. His suggestion is that, as a rule, you should assume the total resistance of shunt plus its parasitic is 24 mΩ. Bill is going to look over the old schematic tonight to confirm/deny that there are fewer millmax pins on the Rshunt leg. We can compare the two techniques using data taken on Monday: |
| Assuming equal parasitic resistance in each branch | Assuming 4 mΩ parasitic on the shunt branch |
|||
|---|---|---|---|---|
| Channel | Rpar [mΩ] | Rn [mΩ] | Rpar [mΩ] | Rn [mΩ] |
| HV1pA | 11.3 | 125 | 8.7 | 96 |
| HV1pB | 10.8 | 116 | 8.4 | 90 |
| HV1pC | 15.2 | 105 | 10.3 | 72 |
| HV2pB | 17.0 | 147 | 11.0 | 95 |
| HV2pC | 11.0 | 139 | 8.6 | 107 |
| HV2pD | 21.6 | 148 | 12.4 | 85 |
It's difficult to say which seems more consistent. We don't expect a correlation between Rparasitic (which is assumed to be at the SQUET) and Rnormal (which is at the detector). There is a slightly stronger correlation in the equal-parasitic-in-each-branch method, suggesting that it is more biased. At any rate we will for now adopt Bill's method and assume that the shunt resistance is 24 mΩ.
| 1100 | Remeasure Rparasitic of HV1pA using baseline-shifts (IbIs) at QET bias = 0 and 20 μA. The equal-parasitic method gives Rpar = 11.1 mΩ, the Rshunt=24 method gives Rpar = 8.6 mΩ. These are quite close to Monday's values. |
|---|---|
| 1105 | Use a test signal on the QET bias (18 mVpp, 100 Hz) to measure HV1pA Rpar. QET bias set to 20 μA. Using the Rshunt=24 assumption this gives Rpar = 10.5 mΩ, significantly higher than the IbIs method. That's strange - it may suggest that the assumption that Rbias = 10 kΩ (used to convert test signal voltage to current) is not quite correct. |
| 1110 | Measure Rn using a test signal as above but with QET bias set to 400 μA to drive the TES normal. After subtracting Rpar this yields Rn = 103 mΩ. |
| 1125 | Measure Rn by heating the fridge. Applied 40 mW to MC heater to heat fridge to 80 mK which is above the transition of HV1pA. Use the test signal as above with QET bias = 0 μA. We get nearly the same Rn, 102 mΩ. |
| 1245 | Measure Rn and Rbiased using the IbIs method. Find Rn = 117 mΩ with equal-parasitic assumption, Rn = 90 mΩ with Rshunt=24 assumption. This is about 6% smaller than Monday's values regardless of which assumption is used. It's pretty close to the Rn value we measured today using a test signal. At the new operating QET bias of 185 μA, we find Rbiased = 65/48 mΩ, 56/53 % of Rn, depending on the equal-parasitic/Rshunt=24 assumption. |
| 1420 | Send the detector shipping vessel and wood crate to TAMU so they can use it to ship the next HV detector |
| 1530 | Build half of the lead/poly shield. The two permanent walls are on the north- and south-sides of the fridge this time as opposed to the previous east- and west-side arrangement. |
| 1537 | Start zero-bias data series with QETbiases = 0 (superconducting), Frontend Gain = 5/5, OG = 2/2, all randoms at 20Hz, with no flashing, with all trace lengths set to 20 ms. Intent is to take long-trace randoms only for noise studies. OOPS - did not mean to have OG = 2/2 |
| 1544 | Start zero-bias data series with QETbiases = 0 (superconducting), Frontend Gain = 5/5, OG = 8/8, all randoms at 20Hz, with no flashing, with all trace lengths set to 20 ms. Intent is to take long-trace randoms only for noise studies. |
| 1552 | Stop data |
| 1556 | 07161102_1556 - Take another superconducting data set with 52.4288ms (65526 bin) traces. Notes: DCRC 1, channel C looks anomalously noisy |
| 1610 | Apply 50 mW to the MC heater to heat the TESs to their normal state. Should target 105 mK to insure we're above all the Tcs. |
| 1615 | Tower = 85 mK and rising slowly. Bump the MC power to 60 mW. |
| 1622 | 07161102_1621 - Normal-state data set with 52.4288ms (65526 bin) traces, with Tbase = 102mK. Notes: DCRC 1, channel C again looks anomalously noisy |
| 1631 | Normal-state data set with 20ms traces, to make Matt happy |
| 1636 | set MC heater back to 0mW now that normal data is acquired |
| 1720 | MB = 36.7% |
| 1725 | Begin LHe transfer @ 35.8% |
| 1752 | Finish LHe transfer @ MB = 91.3% |
| 1755 | Fill LN2 cold trap bath |
| 1800 | Place Cf source near cryostat at detector height for activation. |
| 1807 | Release pressure from main bath |
| 1853 | Release pressure from the main bath |
| 2000 | looked at SC and normal data above. Unfortunately channel DCRC1chC, as predicted above, was inhibitively noisy. A fair amount of SQUID zapping and DCRC power cycling was required to get DCRCs in reasonable order. DCRC3, in fact, became noisy on all channels before power cycling/SQUID zapping bought it's channels down to the level of DCRC1 (and still channel B on DCRC3 was the worste of all 6 channels) |
| 2105 | started SC series but had wrong trace length (only 50ms) |
| 2108 | started SC series with correct 52.4288ms traces … this is crap too because ch B railed at some point. |
| 2118 | 07161102_2117 finally started hopefully good SC series with 52.4288ms traces … DCRC3chB has a peak at 100Hz that we'll have to live with |
| 2138 | the DCRC1 test signal has been malfunctioning. turning it on automatically sends a signal down channel B, with no ability to turn it off. This is not a GUI bug because the ODB reflects the changes made from the GUI. Either this is a driver bug or a DCRC firmware bug. Also have a new series with normal data: 07161102_2332 |