Instrumentation and Test Expert

Prototype Saturable Reactor for Automatic Test Equipment (page 6)

  V. PROTOTYPE SATURABLE REACTOR RESULTS
During the course of this project we developed a number of saturable reactor prototypes. We analyzed each of the prototypes looking for secondary effects such as inductance drift with temperature, sensitivity to differences in control current step sizes, hysteresis and disaccommodation. The following figures show the results of testing for one of the saturable reactor prototypes.

Figure 4. Inductance vs. coil bias current for a step size of 18.5 mA. In Figure 4 the bias current was stepped from 1.31382 A up to 1.38802 A, then down to 1.23969 A, then up to 1.33227 A, in approximately 0.0185 A steps. Each current step was held for at least five minutes to evaluate drift characteristics of the inductance and the inductance was measured once per thirty seconds. Note that there is a small hint of disaccommodation behavior in the data as seen by the initial negative drift after every step change of current. This drift varies from 0.1 to 1.0 micro Henries over the first thirty seconds following a step change of current.

Figure 5. Inductance vs. control coil current for a step size of 3.7 mA/ In Figure 5 the bias current was stepped from 1.32127 A up to 1.39171 A, then down to 1.25466 A, then up to 1.32122 A, in approximately 0.0037 A steps. During the first portion of the run, each current step was held for thirty seconds. The last current step was held for at least five minutes to evaluate drift characteristics of the inductance and the inductance was measured once per thirty seconds. Note that the inductance decreases uniformly with each step up in bias current, and increases uniformly with each step down in bias current.

Figure 6. Inductance vs. control coil current for a step size of 0.7 mA. In Figure 6 above the bias current was stepped from 1.31828 A up to 1.33236 A, then down to 1.30492 A, then up to 1.3183 A, in approximately 0.00074 A steps. During the first portion of the run, each current step was held for thirty seconds. The last current step was held for at least five minutes to evaluate drift characteristics of the inductance and the inductance was measured once per thirty seconds. Note that the inductance decreases uniformly with each step up in bias current, and increases uniformly with each step down in bias current.

Figure 7. Inductance vs. control coil current for various step sizes. In Figure 7 above the bias current was stepped alternately above and below nominal with increasing step size, while being returned to nominal after each step and the bias current was held at each step for approximately fifteen seconds while measuring inductance. This test was done to further evaluate hysteresis and disaccommodation characteristics.

Figure 9. Inductance stabilization parameters. Figure 9 shows the drift over time of the saturable reactor with a constant control current. This drift is entirely attributable to thermal effects as the inductor warms due to the resistive losses in the control winding. We minimized drift during the automated test of the PSEUs by imparting a constant control current to keep the inductor near its operating temperature when it was not involved in the automated test.

VI. INDUCTANCE CONTROL LOOP
The final step in creating our variable inductor was the design of a controller to precisely set and hold the inductance value. For this we implemented the Proportional Integral Derivative (PID) control loop shown in Figure 10 below. Extensive experimentation showed this control loop to be stable and robust for a wide range of control gains. 

VII. CONCLUSION
This paper discussed the design and development of a variable inductor for a PSEU ATS application. Testing showed the absolute accuracy of the variable inductor to be better than 2.0 micro Henries with a resolution better than 0.5 micro Henries. The entire ATS simulates the electrical, communication, and sensor interfaces to the aircraft as seen by the PSEU. The ATS is capable of testing PSEUs from over twenty different models of commercial aircraft and is currently deployed in production test.

© copyright SafeMachines, PLLC all rights reserved