Back when I was an undergrad, we did a lab in the junior-level quantum class that involved making a dye laser. We had a small pulsed nitrogen laser in the lab, and were given a glass cell of dye and some optics and asked to make it lase in the visible range of the spectrum.
My partner and I worked on this for almost the entire lab period, and got nothing more than the occasional faint flicker of a green beam. We got the TA to help us, and he couldn’t do any better. The TA went to get the professor teaching the class, but he was helping other students with one of the other experiments (this class ran several experiments in parallel), so it took a while, and we kept beating on it to no avail.
Finally, the professor came in, and said “OK, what have we got?” We explained the situation, he did a couple of quick checks of the alignment, and agreed we had set it up right. Then he said “You know, sometimes the orientation of the cylindrical lens makes a difference,” and rotated one of the optics in its holder by about 10 degrees. At which point, in my memory at least, this brilliant green beam shot across the lab like in Real Genius. We had everything lined up about as well as it could possibly be lined up, but were thwarted by a factor we hadn’t even been aware was something we could adjust.
I try to avoid this sort of evil professorial magic trick as much as possible, because I remember the frustration of that afternoon trying to get the dye laser going. Sometimes, though, it can’t be avoided, and damn if I didn’t do it again yesterday.
We have an upper-level lab course in our curriculum, which is taken by junior and senior physics majors. We have this broken up into six different modules where students do some advanced experiments and write them up as formal reports. My turn for this has come around again, so I spent yesterday (which was a beautiful day, I should note) in a windowless basement lab running one of these experiments on the calibration of the free spectral range of a Fabry-Perot interferometer.
A couple of years ago, I got complaints that the lab was too cookbook-y, so I modified it to have the students do more of the alignment themselves, taking one of the mirrors off the Fabry-Perot, and very slightly misaligning the other. That worked pretty well last year, and I did the same thing again this year.
Unfortunately, the laser we’re using this year gives a less distinct pattern than the one we had last year (which dies last week, damn it), and I failed to make clear that the adjustments that needed making were small. So yesterday’s lab group spent an hour looking for a big and brilliant pattern to no avail, in the process working the interferometer off into a regime where it couldn’t possibly work well (more or less). After more than an hour of futility, they came and asked me to help.
Whereupon I did the Evil Professor Magic Trick, and got the thing working in less than half an hour. Which I can do, because I’m an optics guy going way back, and know exactly what I’m looking for, and a whole bunch of tricks to get there more quickly. But as I know from experience, it’s really frustrating to beat on an experiment for a long time, then have a faculty member come in and make it look easy.
This carries on longer than you might expect, by the way. One of the post-docs at NIST when I was there was struggling with a model 599 dye laser from Coherent one day, and went to ask the local expert on these lasers if there was a trick he was missing. The local expert came in, took hold of the input and output coupler mounts in his hands, wiggled them around for a minute and got it lasing, then eyeballed the position of the mirrors, let go of them, and used the screws to get them back into roughly the right position, from which it took only a minute to get the laser up and running.
Both of these guys had Ph.D.’s in physics from doing experiments with lasers, but the local expert had done his thesis work with the model 599 laser with serial number 001. He had forgotten more about how those lasers work than any of the rest of us ever knew.
So, I don’t know if there’s any way to avoid ever doing the Evil Professor Magic Trick– it’s really just a function of vastly different levels of expertise, and teaching is necessarily a situation involving vastly different levels of expertise. There will be times when something goes wrong that only an expert can fix, but that expertise will make it look easy. All you can really do is try to minimize the number of times that happens.
Of course, the real question here is what I should do with next Tuesday’s lab group. Should I give them better instructions to avoid the need for the Evil Professor Magic Trick, or should I try to ensure that their lab experience closely parallels that of yesterday’s group?
I note that the Evil Professor Magic Trick stuck in your memory. Maybe this is a good thing pedagogically.
I once set up the class with a problem that developed a transcendental equation and asked them to solve it. After about five minutes of struggle (and goading by the evil professor), I let them know that it couldn’t be solved algebraically. This kind of punchline can reinforce and in some ways deepen the learning.
There are times when deus ex machina is a useful device, and times when it is annoying. For students, most of the time it’s the latter. I’d say give clearer instructions to save the next group of students from going wrong in that fashion. Though of course they may surprise you by going wrong in some other fashion.
The Evil Professor Magic Trick I remember involved a problem from Jackson. I had gone through the problem and obtained an answer which I knew couldn’t possibly be right (it was discontinuous in a place where no discontinuity should have existed), but no matter how hard I looked at my work I couldn’t see the error in my work. So I asked a classmate for help. He went through the problem and explained how he obtained his answer, which I promptly informed him was the same wrong answer I had obtained. It turned out that only one student in the class, who not coincidentally was the professor’s advisee, had avoided the pitfall that snared the rest of us: in curvilinear coordinates, the unit vectors in the coordinate directions are functions of position, so when you differentiate a vector with respect to position you have to include the terms that arise from differentiating the unit vectors.
Well, it’s better than some of my electronics labs as an undergrad where you spend forever fiddling with it, the professor comes over and then spends just as long messing with it and still can’t tell you why it isn’t working. There was some vindication in that clearly it meant something was wrong with the lab write-up, but I’d rather it have been something he could have easily fixed, so we could have moved on, even if it was with help.
It’s only an Evil Professor Magic trick if you don’t use the opportunity to teach something. If you just walk up, make the equipment work and then walk away without explaining why, then you’re Evil.
Mike.
Its only an EPMT if you laugh evilly after doing it.
The laugh makes the evil.
The biochemistry/crystallography equivalent of the EPMT goes like this (possibly apocryphal). A graduate student was given the task (and the established protocol) of growing crystals of a certain protein for x-ray crystallography analysis. He followed the protocol exactly, without success. After weeks of being increasingly careful, he finally admitted to the EP that he couldn’t get the crystals to grow. EP reviewed the protocol, and agreed that it should work. EP then inspected the beaker containing the concentrated protein solution. He was deep in thought, looking at the beaker, stroking his beard, when he noticed that crystals were indeed starting to form. The graduate student, clean shaven, had neglected to properly “seed” crystal growth with “stuff” falling out of a properly stroked beard!
Two examples:
One, in an electronics lab: the dreaded feedback oscillator. We’d heard horror stories about this from previous classes, and were prepared to spend all eight hours of the lab fighting it. What we didn’t expect were the complete, abject failures we all had, six hours in. When we all started clamoring for the prof’s help, he couldn’t get it to work, either. At any bench.
Turns out he had changed the procedure and hadn’t confirmed experimentally that it worked. I won’t even pretend to remember what obscure thing was going wrong with it, although I do remember it was a really stupid way to build an oscillator circuit. It was a complete pedagogical failure.
Prof was so apologetic he generously rewrote the procedures and let us do it again on our own time when the lab wasn’t otherwise occupied by another class. So in this case, the magic trick was, “Here, do it this other way.”
Second, in a physics lab with the guy who is still my all time least favorite professor: measuring Boltzmann’s Constant from the IV curve of a diode. We never, ever, ever figured out what the hell we were doing wrong. We weren’t just wrong, we were way-the-fuck-wrong. Like 2300% error, or thereabouts. It’s a number that stuck in my brain more than Boltzmann’s Constant. Not that I’m bitter.
I don’t think there’s any way to avoid the EPMT– no one is perfect. God knows I’ve been in the position of having to train younger engineers, and even to write some documentation for posterity. It’s not easy. What matters is what you do after it. The above examples are not exactly role models.
My EPMT story took place during an undergraduate research internship, where I was supposed to put samples of green fluorescent protein in an automated fluorescence spectrometer, then expose them to ultraviolet light for varying amounts of time, then put them back in the spectrometer. On the first day of this, the spectrometer was producing garbage data, I tweaked all the knobs I knew about, the grad student tweaked all the knobs she knew about, we gave up and asked the professor…
… who looked the spectrometer over for a few minutes and declared that a small hole in the casing needed to be covered with black tape, so room light would not contaminate the measurements. (This was a hole put there by the manufacturer, to be clear.)
A novice was trying to fix a broken Lisp machine by turning the power off and on.
Knight [now Professor Knight at MIT], seeing what the student was doing, spoke sternly: “You cannot fix a machine by just power-cycling it with no understanding of what is going wrong.”
Knight turned the machine off and on.
The machine worked.
At least your physics students are taking labs! When I was an undergrad at Harvard in the 1970’s there were no labs with the freshmen advanced physics sequence and lab courses were optional and discouraged.
The Truly, Really Evil Professor Trick:
The head of the applied mathematics department lectured on mechanics. At a tutorial session he gave us a problem from the text book, somethng simple regarding the equation of motion of a chain falling off a table onto the floor. None of us could solve it. The tutors couldn’t solve it. A few days later the professor announced to the class that he couldn’t solve it either and had written to the author for help. A few weeks later he repeated this and added that he hadn’t heard from the author, but that a new edition of the book had come out and that problem wasn’t in it.
I’m with Mike at 4: It’s only Evil if you don’t explain. Otherwise, it’s merely EE: Extremely Experienced.