Taking data

By trade, I am a physicist; specifically a condensed-matter experimentalist. More so than other scientists, physicists are rather sharply split into two categories: theorists and experimentalists. Einstein and Feynman were theorists, working with equations and making calculations. Experimentalists work in labs and conduct experiments and analyze data.

A day spent taking data, then, is usually a good day: data is, after all, the lifeblood of experimental physics, and if you’re taking data then that means your apparatus is working, which is always a good thing.

I spent today taking data. Even though the process is heavily automated and computer controlled, the apparatus needs constant supervision. As is the case with most of the experiments I run, there is a very short duty cycle for attention: I’ll need to make an adjustment, then let the experiment run for a few minutes, then make another adjustment, and wait a few minutes, and so forth. And although the taking of data is an inherently good and fulfilling process, the time scale that my experiments work on can be frustrating.

The intervals between which my attention is needed are long enough so that, after a while, it’s sort of boring just to sit there and wait. On the other hand, the intervals are too short to get anything meaningful done. I often think that I should be able to read journal articles in these time spans, and I usually have an article or two with me that I intend to read, but this always fails: reading journal articles must be done in large, uninterrupted time blocks. If I do try to read a journal article, and get in depth reading it, then I’d be likely to let the experiment languish, and in the end that’d be a terribly inefficient way to take data.

As it turns out, surfing the web is the sort of activity that can be broken into minute-long chunks, and since the data taking computer is usually also hooked up to the internet, it actually works out quite well to browse the web during the gaps when the computer is doing all the other work.

Most of the data I take measure some sort of quantity as a function of temperature, and typically the experiment cools to it’s lowest temperature, and then the temperature is raised in steps, with a datum taken at each temperature. A large part of the waiting around I do is waiting for the temperature to stabilize. In many experiments, the speed with which the temperature can be changed and stabilized sets the rate at which I can take data. Inasmuch as I’m browsing the web while I’m waiting for the temperature to stabilize, being lazy and being careful can sometimes look very much like one another. To be careful, you really want to make sure the temperature is stable, that it isn’t drifting or oscillating, and generally, this means waiting longer. But waiting longer also means more time browsing the web and less time doing real physics, so if you’re waiting longer than it takes for the temperature to become as stable as your experiment needs, then you’re being lazy.