|
卫 星 上 的 抗 辐 射 电 路 的 研 究 A Study of Radiation Hardening Circuit for the Satellite Young Hwan Lho*, Sang Yong Lee+ Abstract :The electrical characteristics of solid state devices such as BJT (Bipolar Junction Transistor) and MOSFET, etc, are altered by impinging photon radiation and temperature in the space environment. In this paper, first of all, a BJT (Bipolar Junction Transistor) is tested to γ-radiation and compared with the specifications under the pre-irradiation and the post-irradiation. Second, the research is to design the circuit with the degraded device performance improvement in the switching time and the output voltage at the interface circuit, which is used in KOMPSAT (Korea Multipurpose Satellite), consisted of diodes, transistors, and resistors in order to accommodate the worst case design and radiation effect. Key words: Radiation effect Worst case design I. Introduction II. The radiation hardening parts in satellite should be used since there exist various kinds of radiation particles in space environment. For the past 40 years, the countries with advance of satellite technology have doing research in the field of radiation effects of passive and active components for electronic circuit for space and defence, and are at the level of stabilization. The researches in these countries have been sharing many reports that lead to exchanges of satellite technology. However, the level of the technology in Korea is far behind other commercial industries. In order to experiment the components and the circuits in the radiation environment, we need the radiation simulation facility. The types of radiation are generally divided into particle radiation and photon radiation. The particle radiation consists of the charged particles which have protons, electrons, α particles, ions, and neutral particles that are the neutrons. The particle radiation may also induce ionization so that excess carriers are generated within a semiconductor material and device. The photon radiation consists of g rays and/or x-rays. The units primarily used in radiation effects work with ionization induced by g-ray are rad. The rad is the amount of radiation which deposits 100 ergs of energy per gram of material. When dealing with particle radiation, the units are flux (number/cm2-sec) and fluence (number/cm2). II. Radiation Effects on Transistor The current gain (β) and breakdown voltage of the transistor are tested and compared with the specifications under the pre-irradiation and the irradiation of low dose rates of 4.97 and 9.55 [rad/sec], and maximum total dose of 30 [Krad]. The test procedure and method is carried out by Mil-Std-883 Method 1019. The g source using 60Co is used for the test of commercial transistor products. The sample size of testing transistors is 6 pieces. Many experimental equations[1] were suggested for estimating the relationship between the current gain (β) and the dose. The value of β for pre-irradiation is, in general, more than 150, the one of β has tendency of decreasing to 30 in NPN transistor, and to 20 in PNP transistor when exposed to the dose of 1 Mrad. When the transistor is irradiated under nutron, Messenger-Spratt[2] characterizes the eq. (1) as (1) where βpost the gain of post-irradiation, βpre the gain of pre-irradiation, fn the dose, fT is the unity gain frequency. As the quantity of dose increases for the specified fT , the current gain βdecreases in eq. (1) since the current of collector in transistor is much affected in the radiation III. Simulation of Interface Circuit The serial interface circuit in Fig. 1 changes the digital command data to the bi-level output (0 [V] or 5 [V]) for the other subsystems, and makes the clock by using of the data coming from other subsystems. The interface circuit is the part of the driving circuit, and supplies the initialized signal of VDE (Valve Drive Electronics). The diode of D2 at the input terminal of Q1 is used for improving the switching speed when the transistor of Q1 is ‘off’. The FPGA [4] for bi-level function is connected to the terminal to collector of Q1. It's value is 0.4 [V] when the transistor is ‘on’. The maximum current from FPGA is 2.5 [mA]. Also, when Q1 is ‘off’, the output of VVDEN should have values between 2.4 [V] and 5.5 [V], which is the input level recognized as ‘1’ in FPGA. The sum of saturation voltage between the collector and the emitter in Q1 and Q2 is less than and equal to 0.4 [V] for the maximum allowed collector current. The serial connection of Q1 and Q2 accomplishes the function of NAND gate. The darlington circuit for improving the current gain and reducing the switching time is designed and simulated through P-SPICE simulation. The output voltage measured at VVDEN in the darlington circuit is 303 [mV] which is satisfied with the specification of less than 0.4 [V]. IV. Experiment Results The Gummel plots of Fig. 2 and Fig. 3 for representing IB (base current) and IC (collector current) are measured by setting VCE as constant and VBE as variable.
Fig. 1. Fundamental structure of the interface circuit of satellite The current gain (β) of pre-irradiation in Fig. 2 shows approximately 185, and that of post-irradiation of 20 [Krad] with dose rate 9.55 [rad/sec] in Fig. 3 is decreased to 71. Fig. 2 shows the constant interval of IC and IB . As the voltage increases, the upper line of IC goes downward, but the lower line of IB goes up, and the value is decreased. The fact says that the minority carrier of the transistor under post-irradiation flows more than that under pre-irradiation. Fig. 4 shows the current gain, β, for the device when exposed to g rays. The current characteristics of irradiated transistor show based on quantity of dose and after annealing at 100 oC and 168 hours. The quantities of doses are 0, 5, 10, 15, 20, and 30 [Krad], respectively, and the dose rate is 4.97 [rad/sec]. As the quantity of dose increases, the value of beta decreases, up to 30 [Krad], but the current of beta is recovered to some degree at the level of 20 [Krad] after annealing at 100 oC and 168 hours. The breakdown voltage [3] shown in Fig. 5 are continuously increasing due to the charge trapping at the surface of p-n junction as the quantity of dose increases up to 30 [Krad], but the ones are recovered to the value of 10 [Krad] after annealing. The interface circuit using the darlington circuit and the simulation results are shown in Fig. 6 and Fig. 7 respectively. Reducing the radiation effect of the current gain, the delay time of tphl (propagation delay from high to low) and tplh (propagation delay from low to high) are 24 [nsec] and 15 [nsec], respectively, for 1x1014 [particles/cm2] of neutron dose. The switching time of tplh in Fig. 1 is 49 [nsec], but that in Fig. 7 is reduced to 15 [nsec], and the performance is improved. Also, the switching time of tphl in Fig. 1 and Fig. 7 are 41 [nsec] and 24 [nsec], respectively. The switching time is reduced by 17 [nsec]. Fig. 2. Current gain (β) characteristics of Gummel plot (pre-irradiation) Fig. 3. Current gain (β) characteristics of Gummel plot (total dose 20 Krad, dose rate 9.55 rad/sec)
Fig. 4. Current gain (β) characteristics of a irradiated transistor (dose rate 4.97 rad/sec)
Fig. 5. Breakdown voltage (VBR ) characteristics of a irradiated transistor (dose rate 4.97 rad/sec)
Fig. 6. Darlington circuit for reducing the radiation effect of the current gain Fig. 7. Simulation result in switching for the interface circuit V. Conclusion The results of the current gain obtained in the experiment for the commercial transistor are proved as shown in the eq. (1). The failure rate in measuring breakdown voltage and current gain to g-radiation of 30 [Krad] dose is about 30 %, i.e. 2 pieces out of 6 pieces. It is above the quality control specification for the commercial manufacturing line. It is suggested that the robust structure for sustaining the radiation effect should be designed to reduce the failure rate and prevent the degradation of the current gain. The darlington circuit proposed in this paper gives a comparable switching time, and the output voltage is less than that of the circuit being now operated in KOMPSAT. References: [1] The Technical Reports for Radiation Effects, 1988, Neamen, Univ. of New Mexico, ALBQ, NM, U.S.A. [2] Proceedings of the Workshop on Calorimetry for the Supercollider, Editors, Rene Donaldson and M.G.D. Gilchriese, pp. 575, March 13-17, 1989, University of Alabama, Tuscaloosa, Alabama. [3] S.Y.Lee, Y.H.Lho, “On the Design of Field Limiting Ring for Improving Corner Breakdown Voltage”, Proc. PEMC98, Prague, Czech, pp. 41-44, 1998. [4] TRW Civil & International Systems Division Space & Electronics Group, “KOMPSAT Equipment Specification for Remote Drive Unit”, April, 1996. |