Overview of IbIs automation technique

• IbIs scans should be done on one channel at a time, with all other channels biased in their transitions. It's known that a biased channel (compared to superconducting) heats its neighbor channels enough to make a noticeable difference in their IbIs curves.
• It is most efficient to enable only one DCRC in MIDAS, since data from other channels is generally not useful.
• Set Driver Gain = 1.
• Manually find the approximate peak in the IbIs curve by adjusting the QET bias. Then adjust the Driver Offset to move the baseline to close to the maximum (1.25 V is max - adjust to around 1 V).
• Decide on a range over which to scan - perhaps 150 above and below the peak.
• Stop the MIDAS run before beginning the IbIs scan.

The MIDAS sequence for automated IbIs does the following:

• Set to 20 Hz randoms
• Disable standard file handling
• Cycle through QET biases for one channel
  • start at highest bias and work down (because the transition to superconducting typically causes the SQUID to lose lock)
  • standard is a 300 uA range and step size of -2 uA
• For each QET bias, pause for 3 seconds then take a 3 second MIDAS run
• Name the file labeled by channel and by QET bias (bias part is QET bias times 10, so that 10ths of a uA are possible without introducing a decimal point)
• At end of loop reset QET bias to put channel back in transition

Currently we use a MIDAS sequence called IbIs_loop_Generic. To run it you first edit it by changing the DCRC name and channel number (in 5 places), the maximum QET bias for the scan, and the final reset value for the QET bias.

The sequencer will put the data in /data/chocula/k100/midasraw/livedata, which is a temporary location. At the end of a sequence, you should make a new directory in /byseries with the usual date-and-time naming convention, e.g. 07211215_1531, and move the data files (named QBscan_*) from /livedata to the new directory. You should first log in as sudo user spa-u-cdmssoft.

Compiled Root code has been written specifically to process IbIs scan data. This code uses the cdmsbats libraries to unpack raw trace data. The code creates a Root file with a Root TTree. There is one branch of the tree for each phonon channel. The code simply fills these branches with sequential sample values from each trace. Thus each branch is essentially many traces tacked end-to-end.

The code unpacks a single raw data file (representing one QET bias) and creates a single Root file. The default directory for the Root file is /data/chocula/fritts/data/K100pulses.

One could write a batch script or something to process all the files from an IbIs scan, but currently I simply handle it with Excel, and copy-paste commands to a terminal.