Resolutions on Dealing with Uncertainities

As my 8 weeks here at Mazin Lab draw to a close, I figure I should tell you guys about my learning experiences here. As hinted at in my previous blog, I would like to share some tips that I have found are helpful when one needs to deal with uncertainties in the lab.

1. Start with what you know and “perturb” out.

Now, beyond just a suggestion to understand the “experimental” side of physics, this is a fundamental principle in physics. To understand the world around us, we first make simplifed models(i.e what we know), and then gradually build up more advance problems(“perturbing out”). In fact, this could even be considered a general problem solving tactic-If you can’t solve the problem at hand, solve a easier one.

2. If possible approach a “critical point” with a real-time display

Often, there are certain phenomena that happens past a certain threshold. For example, in the photoelectric effect, if the energy of a photon does not exceed a certain threshold, no electrons will pop off your metal. Or when you hit the resonance frequency of a circuit, you get a massive increase in power. So, try to “sweep” through your control parameters in as small increments as possible.

3. Try to “Double Count” measurements

Now, you cannot always trust the reading of electronics. So by double-counting, I mean measure the same thing in two different ways. For example, if I have a frequency generator hooked up to my computer, the interface tells me the frequency that is being generated. TO check that, I hook up the generated signal to a oscilloscope of spectrum analyzer.

4. Run a “control” experiment to make sure your equipment is calibrated everyday

It is inevitable that the data you collect on one day will at least be slightly different that the data collected on another day. This is just the chaotic way our world works. So before starting your data collection, run a simple experiment to check if measurements are still the same, and if not, calibrate them. Calibration is one of the most important parts of performing a good experiment, because often, we do not care so much about absolute measurements, but relative ones. And to get relative measurements, we need a “control”, and hence the calibration. So in a sense, one can think of calibration as the control group of a biology and psychology experiment.

5. Have a goal when reading electronic manuals

I would say that 80 percent of the manual will be irrelevant for what your task is. For example, if I only care about t he frequency of a electronic signal, there is no need to learn how to change the phase. Furthermore, at least for me, reading manuals just wears my brain down. So the less time you spend reading it, the better. But, note that in some cases, you may miss important instructions.

6. Record what you did each day

Now, this may be common in some areas of research and not so much in others, but I am advocating for the use of journals in a somewhat different manner. I am not saying record your data and procedure. Rather, write now steps in which you were unsure of, and why you were unsure of them. Or, note that there are multiple paths that you can take, and write that for the time being, you chose one particular one. This is all so you have somewhere to start when things go wrong. Because right in front of you, you basically have an outline of places to look where the error most likely occurred.

This all I have to say for this blog. Hopefully these tips can help you guys in your labs too!

Reading the Manual

I have always like to face intellectual challenges. Given an interesting problem, I would often pursue it for days before dropping it from my mind, just to pick it back up a few days later. Rarely do I give up on a problem, because once I do and get the answer, in a sense I have not learned anything. To be more precise, what I have gained in knowledge, I have lost in understanding. In this blog and upcoming one, I would like to discuss a particular group of challenges I have faced in the lab, which are all challenges dealing with uncertainty.

When I walked into Professor Ben Mazin’s Lab, I thought that I would not need to ask for help from my mentor. Seeing as my project was to build a analog readout, which for the most part seemed self-contained, it seemed, at least at the time, perfectly reasonable. I thought that I could, for the most part, use logic to figure out what I needed to do.

Needless to say, that ideal was wrong. In fact, for one entire week, my mentor and I could not figure out how a certain piece of equipment(an IQ demodulation device to be exact) would not work, just to find out the reason was because one little jumper(removable component of a circuit that acts like a switch) need to be removed.

Problem

Solution

Now, here is my excuse for such a blatant error. The instructions for what these jumpers do are in the “Advanced RF Testing” section of the manual. Seeing as my “Basic RF Testing” was unsuccessful, I made the seemingly logical choice and did not read that particular section.

What my mentor and I did instead was go through a series of “tests” in an attempt to isolate the problem.

First, we thought that perhaps the signal was being sent to some other port on the IQ mixer. So, we hooked up all the ouput ports to an oscilloscope, hoping that one signal would give a sinusoidal waveform. That didn’t work.

Next, we decided to upgrade our testing device from a oscilloscope to a frequency analyzer. Our thinking was that the oscilloscope was too insensitive to the signal that was being read out. In essence, we decided to upgrade the equipment. This also didn’t work.

Then, we tried to measure current and resistances on various parts of the IQ mixer to see if a component was broken. Everything seemed ok.

After a few other things that my mentor tried to do also failed, he told me that we should read the manual from front to back and see if it tells us anything. The next day, he found the “Advanced RF Testing” section, and making the adjustment shown in the pictures above, we finally got the IQ mixers to work.

The immediate thing I learned from this is that knowing which parts of the maual are relvant for my task is a good skill to have. I read most of the manual, and none of it helped for our problem, and in some cases, (because I am a physics major and not an electrical engineering major), confused me. But, wanting to turn this into a learning experience for me, in my next blog I will talk about how to deal with situations like this if they arise again.