Instructions for tutors performing the basic GOLEM laboratory experiment

The basic GOLEM laboratory experiment is performed by 2nd-year students of FJFI ČVUT. Its goal is to replicate basic tokamak diagnostics using an oscilloscope and a bit of twisted wire, and subsequently to gauge the dependence of the energy confinement time on the magnitude of the toroidal magnetic field.

The lab is so hard for students that we give them an extra week for writing the report. Sometimes not even that is enough. But it’s hard also for the tutors, who have to keep track of a lot of things while having (sometimes) no idea what’s wrong with the measurements and still appearing competent to the students. To the tutors (and myself), I dedicate this page.

1. Read the manual

As of January 2021, the current version of the manual is stored on the Overleaf platform. If you can’t access the link and download the PDF, write to me. A tutor must know the manual well before commencing the measurement, if only to gauge if the students have read and understood it.

2. Make sure the students read the manual in advance and appoint one per group to write data processing scripts

Demand and see to it that the students receive the manual at least a week before the measurement. Let them know that this document contains most of the information they’ll ever need and threaten that if they don’t read it, they will be thrown out of the laboratory. (They won’t be thrown out in reality, but it delivers the gravity of the command.)

While you’re at it, tell the students to appoint one person per group to write data processing scripts. Ideally, they will be able to download data from the website and plot them. Give them freedom to choose which programme they’ll use, but recommend Python strongly if they have only weak preference. Most students give in and turn to Python. Then, you can help them out during data processing by directly copy-pasting from your prepared script (see below).

3. (optional) Gauge the student level of knowledge

Students can be divided into three tiers.

1. Beginner, who hardly know what Maxwell equations are. All they know about plasma is that it’s cool. Typically encountered in high school.
2. Intermediate, who have had at least one course on plasma physics. They lack specific tokamak knowledge, but don’t need explanations for everything. Typically encountered in bachelor’s and master’s studies.
3. Advanced, who have experience with tokamak experiments. They seek challenge and often specific topics. Typically encountered in PhD studies.

Adapt the depth (and length) of your explanation to these tiers.

4. (optional) Agree on a mode of communication with the students

You will be provided with videoconferencing space by Vojtech Svoboda, but you might prefer your own mean of communication, such as Discord.

5. Contact other tutors

Make sure you have enough tutors and establish a channel where you can discuss problems with the measurements. MS Teams is sufficient.

6. Prepare your own data processing scripts

This is very important. You must have a complete data processing script up and working prior to the laboratory start. For a given shot, the script should be able to plot the standard diagnostics data, the raw oscilloscope data and the processed oscilloscope data. You should run this script for every discharge that your students make and make sure the data was collected fine. This is to avoid errors which are not apparent at the time of measurement, but compromise the data processing later. They include:

• Bad oscilloscope axis X setting. Either it’s too short (and not the entire shot is collected) or it’s too long (and time resolution is lost).
• Bad oscilloscope axis Y setting. Either it’s too short (and the signal saturates, which is extremely fatal for \(B_t\) and \(I_p\) measurement) or it’s too tall (and voltage resolution is lost).
• Bad placement of the measuring wires. This can distort the signal and/or introduce large peaks on it.
• Signal fails to be recorded by the data acquisition system. This goes both for the standard diagnostics and for the oscilloscope data.
• The oscilloscopes are not connected. Then you must figure out where else to take data from (like raw signals of the basic diagnostics).

7. Listen to Vojta’s speech

Before every measurement, Vojta gives a short speech, welcoming the students to the tokamak. Resist the temptation to tune it out and lend it an ear. You’ll avoid repeating the things which Vojta has already explained (such as the remote control room).

8. Get on a private videoconference with the students and introduce them to the laboratory

For every group, there should be one tutor and one videoconference. Turn on your camera and make sure your microphone works. Welcome the students. Introduce yourself. Tell them to turn on their cameras. Tell them to interrupt you with any question they might have.

Recap what you’re going to do during the laboratory. Explain the underlying question (is plasma confinement affected by \(B_t\) magnitude). Make sure they know that the energy confinement time is not the discharge duration. State that the lab will take about 4 hours and that you’ll make a break in the middle. Ask who has prepared the data processing scripts.

9. Show the shot homepage and explain data access

Send the students a link to the shot homepage. Make one of them share their screen and explain the shot homepage to them - where they’ll find the comment, the discharge parameters, what the basic diagnostics mean and where they’ll find their data.

Explain how the diagnostics look, how they are connected and how we collect the data. Explain that different signals require different post-processing. Tell the students that they’ll prioritize making discharges for the time being, but if there’s time at the end, you’ll do some of the data processing together.

Make the student who has prepared data processing scripts (or is the most willing to do it on the spot) to load and plot oscilloscope data from the latest discharge. (There’s a large chance Vojta will do a test discharge before the laboratory.) This will ensure that they know where to get the data.

10. Take them through the control room and let them make discharges

Designate one student (not the one who has been the most active so far) to open the remote control room and share their screen. Then, take them through the discharge submission step by step. When that’s done, make them agree on a particular parameter scan and let them shoot. If there is a lot of groups and little time, you can agree on a central parameter scan with each group participating only with a few shots.

11. Control the data quality

For every discharge they make, check the data quality with your script. If something is wrong, address it immediately.

12. Answer their questions, solve their problems

There is a new problem almost in every instance of this laboratory. Don’t worry about it. In the worst case, the students can use data from some previous lab. Mistakes provide room for growth.

13. Begin the data processing

There is a fine line between supporting the students in their efforts and doing the work for them. It depends on how active the students are, how much Python they know, how much they know about tokamaks and coil signal integration… It is up to you to judge how much you should help them. Take comfort in the fact that small errors make little difference in the end.

14. Say goodbye and (optional) offer your contact

It’s nice if you promise to stay in touch should the students need any help with data processing. However, it isn’t compulsory. You might be too busy.