40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
 Cryo Lab eLog Not logged in
 Wed Aug 1 11:38:36 2018, aaron, mandy, DailyProgress, Temperature Sensing, progress on FSL Wed Aug 1 18:58:53 2018, aaron, DailyProgress, Temperature Sensing, progress on FSL
Message ID: 2117     Entry time: Wed Aug 1 18:58:53 2018     In reply to: 2116
 Author: aaron Type: DailyProgress Category: Temperature Sensing Subject: progress on FSL

I added blocks to the x1siq model to use the moderinger technique to readout loss and Q for four modes. On the fourth, presumably highest-frequency mode, I modified the moderinger to send the bandpassed signal to a lock in block that mixes the signal with an LO that we will give whatever frequency that mode has when we 'start' our measurement. I then take I and Q from that lock in and find the beat frequency by

$\Delta f=\frac{1}{2\pi}\frac{d}{dt}\tan^{-1}(Q/I)=\frac{1}{2\pi}\frac{1}{1+(Q/I)^2}\frac{d}{dt}\frac{Q}{I}$

I take the simplest possible time derivative of Q/I, by holding one sample of Q/I for one clock cycle, then multiplying the difference between subsequent Q/I by the sampling rate.

This beat frequency is sent to an EPICS output so I can pass it to our existing temperature control PID loop. We still need to:

• define new EPICS channels to hold all of our new variables
• modify our medm screens to make this all useful
• modify the PID script to switch between stabilizing the temperature with RTD sensors (used for the initial cooldown and 'ESD off' stabilization) and with a frequency sensor (used during measurement)
• this will require another indicator channel
• compile and test our model
• Measure Q!
• It would be nice if the LO was automatically seeded with the initial eigenfrequency, rather than having to enter it manually
• We could implement some automated temperature (really eigenfrequency) sweeping
• Add some calibration between beat frequency and temperature difference
• Instead of stabilizing to the initial eigenfrequency, we could instead use the mode Qs to stabilize the temperature to the zero CTE crossing. This could be the first step of some automated measurement, followed by a temperature sweep to produce plots of Q(T).
ELOG V3.1.3-