Finally, We Have a Shutdown
We’ve come a long way this week. On Monday, we hadn’t even begun the main crux of our experiment, which involves studying the adverse effects that high frequency electromagnetic interference could have on power supplies that are found in every desktop computer. Konstantin has set up something called an oscillator which uses a special kind of amplifier to output high frequency signals. More specifically, he has configured this oscillator to oscillate chaotically, or nearly chaotically. The signal that it puts out is exists in a wide band of frequencies instead of a narrow band.
For the past three or four weeks, I have been studying a power supply, trying to understand the circuity and identify places where it might be most susceptible to interference. This week we finally brought together our two aspects of this research project: I installed an RF connection and soldered it to different points on the power supply’s circuitry and then we injected signals from Konstantin’s oscillator into the power supply. We used oscilloscopes to monitor the output voltages and varied parameters in order to see what it would take for the power supply to get disrupted.
We ended up finally injecting signals on Wednesday. By that point, I felt we were a week behind schedule. You see, our final poster for the MERIT Fair (on August 11) is due the following Thursday, and our final technical report (~20 pages) is due the following Monday. This gives us precious little time to analyze our results. Wednesday and most of Thursday was difficult, because we weren’t getting any real good results. We hammered that poor power supply with up to 5 Watts of peak power, and we were just able to get some output disturbances on the range of a few hundred millivolts. While this might enough to technically say the power supply is ‘disrupting’, I was hoping to shut the thing entirely down.
I had this hunch though, after noticing the behavior of the power supply as I was testing it earlier. I wanted to try injecting the signal into a particular cable, but it was met with some skepticism. Late Thursday night, around 9pm I think, I wanted to try it regardless so I soldered the connector on and gave it a go. To my delight, the power supply shut down! I spent the next two hours systematically testing it and grabbing a lot of data.
Today we further explored what was causing the power supply to shut itself down, and did some more tests using different setups and controls to get a better picture of what was going on. It was pretty good.
This weekend is going to be tough though. I have to first sort through and organize all the data we took these past three days, as well as analyze them for trends. Then I’m going to write as much of the final report as I can (covering my parts), because much of that will feed into the poster.
I just have to get to next Monday, after we send off the final technical report to all the judges. Then we can chill a bit before the MERIT Fair.


