Q0 Measurement Using dLL/dt (Rate of Change of Liquid Helium Level)

Overview

We're indirectly measuring the Q0 by figuring out how much heat a cavity is generating at a given gradient. We figure out that heat load by looking at how quickly the helium evaporates inside the cryomodule (more heat, faster evaporation).

Note that the relationship between the liquid level sensor readback and the volume of LHE held in the cryomodule is not consistently linear. This is due to the interior shape of the helium vessel. Imagine you were pouring liquid into the following vessel at a constant rate:

     ----------------       D
    |                |
    |~^~^~^~^~^~^~^~^|
     ----        ----       C
         |      |
         |      |
     ----        ----       B
    |                |
    |                |
     ----------------       A
    

If there were a level sensor in the vessel, it would increase at one rate in the [A,B] and [C,D] regions and at another (higher) rate in the [B,C] region.

We only take data with the downstream liquid level sensor reading 90-95% in our calculations. Everything is nice and linear in this region, just as it would be in region [C,D] (crossing 90 is analogous to crossing into region [B,C])

Cryomodule Calibration

Run the handy dandy script! It makes a calibration curve that maps heat load to dLL/dt (change in liquid level over time) for the cryomodule in question. But as to what the script actually does:

1. Scans the past 24 hours to see if there's a 1.5 hour chunk of time where the liquid level was stable. The idea is to find the JT valve position that lets helium in at the same rate at which it's being boiled off by the static heat on the cryomodule (where the static heat is the heat leaking into the cryomodule from the outside world plus a contribution from the electric heaters at their default settings). After locking the valve at this setting we know that any change in liquid level is due *only* to the heat load that we've added.

   1. If it finds a stable period, it finds the average JT valve position over that time span.

   2. If not, prompts the user to ask cryo to fill to 95% on the downstream sensor and wait 1.75 hours for the liquid level to stabilize, then average the JT position over the last 30 minutes.

2. Prompts the user to ask the cryo group to lock the JT Valve at the position found in step 1.

3. Increases the heat load on the cryomodule by 8 W using the heaters (distributed evenly across them)

4. Waits for the liquid level to drop by 2.5% (The amount we've empirically determined is necessary for a good linear fit).

5. Prompts the user to ask cryo to refill to 95% if the current liquid level is below 92.5%.

6. Repeats steps 2 through 5 10 more times, except with a heat load increment of 1.6 W instead of 8.

Cavity Q0 Measurement

Run the other handy dandy script! But as to what the script does per cavity:

1. Determines a new JT Valve position if necessary (using the same method as in Calibration step 1)

2. Checks that the downstream liquid level is above 92.5% and that the valve is locked to the correct value

3. Makes sure that all the waveform acquisition controls are enabled/at the correct values

4. Turns the SSA on

5. Characterizes the cavity

6. Turns the RF on in pulsed mode

7. Checks that the On Time in 70ms (and sets it if not)

8. Increases the drive until it's at least 15 OR the gradient is at least 1 MV/m

9. Phases the cavity by getting the "valley" of the reverse waveform as close to 0 as possible

10. Goes to CW mode

11. Walks the gradient up to the requested value (usually 16 MV/m) and holds it there until the liquid level drops 2.5% (with some quench detection built in that triggers an abort)

12. Powers down the RF

13. Launches a heater run to be used for error correction during analysis (in order to find an offset for the RF heat load)

    1. Increases each heater by 2 W
    2. Holds until the downstream liquid level drops by 2.5%
    3. Decreases each heater by 2 W

Analysis

Input

A CSV file (input.csv) where each row follows the header format:

SLAC Cryomodule Number	Cavity 1 Gradient	Cavity 2 Gradient	Cavity 3 Gradient	Cavity 4 Gradient	Cavity 5 Gradient	Cavity 6 Gradient	Cavity 7 Gradient	Cavity 8 Gradient

As currently written, the program reads *every* row after the header and runs a separate analysis on each. Per row, the script will:

1. Read the first cell to determine the desired cryomodule
2. Look through an internal record of previous calibrations and present them as options, along with an option to run a brand new calibration
3. Analyze that data to generate a calibration curve (mapping dLL/dt to heat load)
4. Iterate through the remaining cells, reading the desired gradient per cavity (a blank cell will simply skip that cavity)
5. Look through an internal record of previous Q0 measurements for each desired cavity and present them as options, along with an option to run a brand new measurement

Calculation

After pulling all the required data, the script parses it into data runs based on heater and/or RF settings.

For the calibration, it fits the 10 data points (one per heater setting) to a line (heat load vs. dLL).

For the Q0 measurement, it:

1. Fits the liquid level to a line and finds that dLL/dt for both the RF and heater runs

2. Plugs the dLL/dt from the heater run into the calibration curve to get a heat load

   • If that back-calculated heat load is not equal to the heat added to the heaters during the run, finds that offset

3. Plugs the dLL/dt from the RF run into the calibration curve to get a heat load

4. Adjusts the RF heat load by the amount determined in step 2

5. Plugs that adjusted RF heat load into a magic formula to calculate Q0!