⚡Electrical Tuesday
Tuesday, August 11, 2026
If your controller oscillates, lags, or quietly walks away from setpoint, the answer is in one of these three concepts. They've been the same since Bode worked at Bell Labs in 1938.
Term 1: Transfer Function
ElectricalThe Laplace-domain ratio of output to input for a linear, time-invariant system. It encodes every pole, zero, gain, and delay between your command signal and the thing the system actually does. The roots of the denominator (poles) tell you stability; the roots of the numerator (zeros) tell you where the system goes deaf to certain inputs.
Why it matters: Tune a PID loop without knowing the plant's transfer function and you're guessing. Identify the transfer function from a step test and you can stop guessing — which is the difference between a process that runs and a process that babysits the operator.
Term 2: Bode Plot
ElectricalThe transfer function evaluated along the imaginary axis (s = jω), plotted as magnitude in decibels and phase in degrees, both against log-scale frequency. Every control textbook has them. Every working controls engineer can sketch one for first-order, second-order, and integrator systems from memory.
Why it matters: The Bode plot is how you reason about loop stability without solving the closed-loop equations every time. If you can look at the open-loop Bode plot and know where the gain crossover sits, you know whether the closed loop will ring.
Term 3: Gain Margin & Phase Margin
ElectricalHow much extra gain (GM) or extra phase lag (PM) you can add to the open loop before the closed loop goes unstable. Aerospace standard practice in the cluster: at least 6 dB gain margin and 45 degrees of phase margin. Less than that and you're running a stability test every flight.
Why it matters: Every gimbaled engine, every flight control law, every motor-drive servo bus on the Arsenal lives or dies on gain and phase margin. A loop that meets specs in the lab and oscillates in the field almost always lost margin somewhere — usually to sensor lag nobody modeled.
📍 Huntsville Pulse
- •IEEE Control Systems Society Huntsville chapter is restarting after a quiet spring. First meeting back: model-predictive control case studies from the missile defense community.
- •NI's TestStand and FlexLogger workshops at UAH are back on the calendar this fall — registration opens late August.
- •Cullman Electric Cooperative is hiring SCADA engineers — control systems people with utility experience are in short supply across the valley.
- •Tuesday-night dinner play: Phil Sandoval's on Whitesburg is quietest between 7:30 and 9:00 if you want to actually have a conversation with the engineer you're trying to recruit.
“A controls engineer's spouse asks 'how was your day?' The controls engineer responds 'asymptotically stable, with about 8 dB of margin.' The spouse files for divergence.”