The old joke is, “If it squirms, it’s a biology lab, if it stinks, it’s a chemistry lab and if it doesn’t work, it’s a physics lab”. My current job is to make a lie of the latter, by making physics apparatus that works for the students without needing a “resident expert” to maintain or explain it.

I work for TeachSpin, a small company that builds equipment that is used to teach students in advanced undergraduate physics labs. Some of the instruments are capable of making “research grade” measurements, and all are designed for open-ended investigations where the students can go beyond what is outlined in the manual. As TeachSpin’s senior scientist, I am in charge of designing new experiments, supervising their construction and testing them before they leave the workshop for new homes in labs around the world. The instruments the firm makes span a range of physics topics, from atomic physics and magnetic resonance to acoustics and optics; they can be as simple as magnetic-force apparatus or as complex as measuring the hyperfine splitting in the excited states of rubidium atoms using Doppler-free spectroscopy.

TeachSpin was founded in 1992 by Jonathan Reichert, who was then a physics professor at the University of Buffalo in the US. He was also my thesis advisor – proving once again that when it comes to careers, it is often who you know rather than what you know that counts. After I completed my PhD in solid-state physics in 1993, I worked at TeachSpin for several months, designing electronics for its first instrument — a pulsed nuclear magnetic resonance spectrometer. A few postdoc positions later, I found myself working as a staff scientist at the W M Keck Free Electron Laser at Vanderbilt University when the facility lost its funding for a year. As a recently married new father, I thought I should start looking for a job that did not depend on the three-year research-funding cycle. By that time, TeachSpin had grown and it was looking to hire a full-time physicist. It was a perfect fit for both of us.

From “care and feeding” to design

Part of my time is spent on what I call “care and feeding” of existing instruments. This can range from talking to potential customers about the apparatus, to helping existing users get their experiments up and running, and even to coaching students. Every once in a while, I get some good physics questions, but in the main, people want to know about set-up and specifications.

This part of my job also includes production-related work like answering questions from people in the workshop, finding replacement parts for something that is about to become obsolete or helping set up equipment for intermediate-level testing. My favourite part of production work is the final testing of the apparatus. First, I make sure all the mechanical parts and electronics are working. Then, I get to take the first pieces of data from this particular unit. Even though I have seen the same dips, bumps and/or wiggles of data hundreds of times before, it is still a bit of a thrill to see them in their latest incarnation. Once the data are recorded and a copy placed in the user manual, I wheel the instrument out to be packed up for shipping. At these times, I cannot help feeling a little like a proud father sending another of my “babies” out into the world, hoping the new owners will cherish it as much as I do. Note that “father” is a good analogy — it is the production people who do all the finicky assembly.

Most of my time, however, is devoted to designing new instruments. This could be in collaboration with an outside expert — the firm built both its diode-laser spectroscopy unit and its Fabry–Pérot cavity with Ken Libbrecht of the California Institute of Technology, for example — or entirely with in-house staff. My favourite part of the design process is starting on a new project. For example, my latest project is the study of “Johnson” and “shot” noise in electronic circuits; there are some beautiful correlation techniques using two identical amplifiers that can be used to remove the noise, but it looks like I will not be able to do this with a single instrument. As is often the case, I may at this point have to learn some new physics (a great excuse to go and buy some books) and, at this early stage, the sky is the limit. My colleagues and I always like to think about all the possible things the new piece of apparatus might do, and any crazy ideas can be explored.

Soon after this, reality sets in and we have to balance things like the costs and the time involved against the potential for increased sales. With a few more parts, for example, our optical-pumping apparatus could be made into a rubidium magnetometer or an atomic clock, but most students will still be struggling with the basics after several weeks of using the equipment. The challenge is to let the apparatus be used over as wide an experimental range as possible — giving the student lots of different knobs to turn (figuratively and literally) — yet keep the price low enough that physics departments can still afford to buy it. I sometimes feel like I am cheating the student by doing all the design work: I learn a great deal from all the mistakes I make, but the students do not get this opportunity. However, making mistakes takes time, and in undergraduate labs, this is often in short supply.

Making things work

A career in instrument building can involve knowledge from almost any area of science and technology. A list of what I use daily might start with practical topics like electronics, technical drawing and material properties, and continue on to entire fields like optics, atomic physics, electromagnetism or solid-state physics. One very useful trait for this job — common to many scientists and engineers — is a desire to understand how things work, and perhaps to make them work better or be produced more cheaply. Aside from this, I think one of my greatest assets is having the tenacity to stick with a problem until I understand it. A small glitch or wiggle in an experimental spectrum is not acceptable. Chasing down all the small noise sources that can crop up in a piece of equipment takes time, but the reward is a good instrument.

One of the drawbacks of working for a small firm is that there are few other physicists to help you bounce ideas around. E-mail helps, but it is not the same as standing at a whiteboard, drawing pictures and waving my hands. However, this changes for a few months each year when David Van Baak, a physicist from Calvin College, visits to collaborate on new projects. During this period, we have a marvellous time and the ideas just keep on coming — we brainstorm, argue, refine and continually think of more experiments to do with the apparatus we are designing until we finally have to stop and focus on getting it out the door.

As there is a large practical component to my work, my advice to anyone interested in a similar career is to get to know the technicians in your department or university workshop. If you are a graduate student designing equipment for an experiment, do not just submit a drawing to the technicians — take the time to talk to them about what you are doing. You may have to bribe them with pizza or other offerings, but this will be money and time well spent. You do not have to accept all of their suggestions, but they are bound to have some good ideas about how to make things work.

Obviously, a physics degree can be good preparation for this kind of work, but my first degree was actually in engineering, so I was not exposed to the classic advanced-physics labs. This naivety can be useful. First, I do not have preconceived notions of how the experiments should be done, so I may be able to think of a different way to show the desired effect. For example, you do not necessarily need to be able to sweep the frequency of your Fabry–Pérot cavity if you can tune the wavelength of your diode laser. And perhaps even more importantly, when I start a new project, I am approaching it much like the students: I am doing it for the first time.