Turn down that noise

This may sound like a familiar command to some of us. Not only does this word apply to some forms of music, but it also can be a reference to electromagnetic interference. As an Instrumentation Engineer you may be called upon to turn down the noise, but that won't mean adjusting the volume. An Instrumentation Engineers job during a wind tunnel test is to insure the customer gets the best quality data attainable. Our secondary goal is to ensure that the instrumentation measuring the phenomena operates reliably and consistently. Immediately after the installation of the TCA model in the test section, problems arose. A preliminary look at the data reveals that the Rolling Moment gage is showing a standard deviation of 140 counts over a 0.7 sec span of acquired data. This is the equivalent of 37 ft-lbs for a gage with a maximum capacity of 2000 ft-lbs, or a 1.85% error band. This is not music to the researchers ears. With errors this large it would be pointless to continue. The Instrumentation team springs into action.

We commence the search for the source of our problems. Instrumentation engineers get to use lots of electronic equipment. We used a spectrum analyzer in an attempt to characterize the noise. Again we see the rolling moment gage of the balance, only this time it's in the frequency domain. The balance is used to measure the forces and moments the model is experiencing. What we see from the graph is 165 mVp @ every 10 Khz. This is riding atop a channel that has a maximum output of 1.6 mV/V. So, the noise has a magnitude of a hundred of the desired signal. Now we eliminate the institutional data system, and continue the search with our own toys. As delve deeper for the source of the noise, reaching into the belly of the tunnel, the wiring looks like intestines down here.

We have isolated the culprit, a digital stepper motor used to control the model position. We need the motor to position the model, but we can't tolerate this level of EMI. So, we unleash the weapons of Instrumentation: by-pass capacitors, an external power source, and shielding and grounding. The intermediate results are shown below, the capacitors alone reduce the level of noise to 60 mVp. The test team test is happy again, and the test can continue.