This is an example time line for dark testing wafers.

Day	Action	Notes
1	Kam mails next wafer
2 - 3	We install wafer and start cooldown	Usually a wafer arrives the day after Kam ships. With current wafer holder we can put directly in, but new holders usually require some machining.
4	Pump down	Pumping down usually only take one day.
5 - 8	Cool down	Cool down can take anywhere from 2 to 4 days depending on the thermal mass of the holder.
9 - 12	Take measurements	This timing is for a quick turn around. If we are not doing a quick turn I take more time to try and understand the system more and get complete measurements.
12 (evening)	Start warm up	Kill cycle and turn on heaters. We are usually warm enough to open in the morning.
13	Open up and pack wafer
14	ship to Berkeley	If everything is ready the we can open, pack and ship in one day.

The following tests are standard (most example documents are for wafer KT7):

• Network analysis: Network Analysis
• Bolo IVs: IVs
• ETF: Electro-thermal time constants
• Tcs: Critical temperatures are now taken with a new PID control. I have not implemented an automatic pdf generation for them yet. Wafer 8.2.0's wiki page has a table with Tc values in mK.
• Bolo noise: Francois did some in depth noise measurements with KT7 this is an example of the data for one state where all the bolometers in the comb are biased 80% into the transition. KT7's wiki page has a large number of documents with noise taken in various configurations.

What we will do particularly for 8.2.1

• If there are ADC overloads with 8.2.1 we will take noise levels in arrangement where there are no overloads and also with the overloads occurring.
  • We will also try to troubleshoot the problem to reduce or eliminate the issue. Are there particular approaches that Mike and Ziggy have found effective for this?
  • We find ADC overloads in the following condition:
    1. Stage warm (700-800 mK)
    2. SQUIDs tuned
    3. No carriers
  • This is a scan of demodulator input in this condition with demodulator gain 1. (ADC overload flag true)
  • This noise was taken for comb A in the following configuration:
    1. Stage cold (230 mK)
    2. Bolometers overbiased
    3. Demod gain 2
    4. ADC overloads
  • PB_8_2_1_20100118-M1.pdf: Full data taken at demod gain = 2 with bolometers overbiased
  • PB_8_2_1_20100118-M1_processed.pdf: Overview of all the bolometers from M1
  • PB_8_2_1_20100118-M6.pdf: Full data taken at demod gain = 2 with bolometers 0.8 of R_ob. Noise for combs B (153-1-*) and V (157-1-*) should be ignored as they have bolometers that latched earlier.
  • PB_8_2_1_20100118-M6_processed.pdf: Overview of all the bolometers from M6
  • PB_8.2.1_Combs__FJ.pdf: Noise data taken without any bad squids or combs. demod gain = 2, all bolos 0.8 R_ob, there were ADC overloads on both combs.
  • PB_8.2.1_Combs_FJ_Summary.pdf: Summary of noise levels for combs F and J.

• We will take entire combs down to various depths into their transitions to investigate stability.
  • I took combs A and J down into their transitions. I dropped them to 0.9 of starting resistance each time so the total fraction of R_ob would be 0.9, 0.81, 0.729.
  • I found that J worked OK until 0.729 when bolometer 7 latched while tuning.
  • Comb A showed the ADC overloading at 0.81, however all bolometer tuned to that depth. Not surprisingly the entire comb had problems when I tried to take it to 0.729.
• Run NA with low carrier amplitude (still good S/N) (carrier gain = 0, amplitude = 0.2)
  • PB_8.2.1_netanal_LowAmp.pdf: NA taken with low carrier amplitude on normal comb
  • PB_8.2.1_netanal_LowAmpF.pdf: NA taken with low carrier amplitude on comb with suspicious peak.
• Run NA with high carrier amplitude (carrier gain = 3, amplitude =1)
  • PB_8.2.1_netanal_HighAmp.pdf: NA taken with high carrier amplitude on normal comb
  • PB_8.2.1_netanal_HighAmpF.pdf: NA taken with high carrier amplitude on comb with suspicious peak.
• Take 2-3 bolometers down to latching
  • Revisit those bolometers and see how low they can be nulled
• Take ETF tau measurements for a few bolometers at different depths in the transition - DONE
  • Time_Constants.txt: file with time constants in ms for the first bolometer in a comb.
• Do standard ETF out to 5 kHz for a couple bolometers
  • Make first measurement of this at overbias so we can separate the time constants
    • Please note I need to follow up on these measurements because they have been done with different bolometers.
    • PB_8.2.1_etf_GT5kHz_1.0.pdf: Data taken without lowering the bolometer
    • PB_8.2.1_etf_GT5kHz_0.8.pdf: Data taken with the bolometer at 0.8 R_ob
• Automate Tc document production
• Rewrite nulling to always take and return highest level of data and verbose only controls what's printed to the screen

Things of note

• I have been workin on a new ETF procedure however, we need to understand the analysis to use it's output
  • Check any discrepancy between calculations and analysis code
  • Determine why phase shift is in opposite direction from expected
  • Determine why phase shift is not as large as expected. *Tc and Bolo noise measurements are not automated at the moment.

• XB1 squids spec analysis:
  

• K06.pdf: These were taken on Sunday Jan 24th, they are not the same bolos as I showed on Thursday the 22nd and demod gain = 0. Data shown is for all squids turned on and with the first squid turned off.
• P1.pdf: Taken Monday Jan 25th. I showed data for SqCh7 on the 22nd. Both data sets have demod gain = 2. Data shown is for all squids tuned and on, and all squids untuned and turned on.
• A note on the units on these plots. The units are rms of the time stream data. Talking with Kam we come up with a noise floor of 7.5 pA/sqrt(Hz).
• P1_demod1.pdf: Take Thursday Jan 28th. demod gain = 1. Each squid is turned on and tuned by itself. Strangely this shows a 5K feedback with lower peaks on the 4Mz.
• P1_OpenVs10K.pdf: plots comparing openloop and 10K feedback.

-- PeterHyland - 21 Oct 2009

---

This topic: BolometerTesting > WebHome > PolarBear > PlanForWafers Topic revision: r14 - 2010-01-29 - PeterHyland

---