QUALITY ASSURANCE FOR SPACE INSTRUMENTS BUILT WITH COTS

Transcription

1 QUALITY ASSURANCE FOR SPACE INSTRUMENTS BUILT WITH COTS Peter Buch Guldager, Gøsta G. Thuesen, John Leif Jørgensen ØRSTED*DTU, Measurement and Instrumentation Systems, Building 327, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark Phone: , Fax: , ABSTRACT Instruments for space can be built with COTS. However no radiation data are available for COTS, so the only way to ensure that the components can survive the space environment is to irradiate each component. Samples from each Lot have to be irradiated, because the manufacturing process can be changed at any time and have major consequences to the components ability to survive the space environment. A safe way to protect to components, which are not Latch-Up immune, is to protect the components with a Latch-Up protection circuit. A strict control has to be established, when procuring COTS component, testing and manufacturing the instrument before the instrument is qualified for space. By having a strict control with instrument built with COTS, it is possible to manufacture a reliable instrument as with Rad-Hard components. 1. INTRODUCTION There are several ways to build an instruments for space; you can manufacture the instruments by using components designed for space (Rad-Hard components) / military components or by using COTS. If you are using components qualified for space applications, the size of the instrument will be larger and more power consuming compared to instruments based on COTS SMD. The advantage of using COTS is the increased functionality, the speed of the instrument and less power consuming, but how do you ensure, that the instrument will survive the launch, the radiation aspects for EEE components in space and the lifetime in space? How do you control the production of an instrument, when using COTS? 2. SELECTION PROCEDURE FOR COTS There is not the same traceability with COTS compared to space/military components, but this does not imply that it is not possible to use COTS when designing an instrument for space. How do you control the quality of the instrument, when it is built with COTS? The way to do this is by testing at component level or at board level or at unit level. For instruments used in space, one of the major concerns is the radiation aspect for the EEE components. When using COTS, no radiation data are available for the EEE component and you have to ensure the component will survive the entire lifetime of the mission. A reliable way to do this is to test each component for TID (Total Ionizing Dose) and

2 monitoring different parameters for each component during the irradiation. Comparison of the test results from the irradiation test with the results from space shows that this is a very reliable way to ensure the radiation tolerance. When designing a new instrument, different candidates of the components have to chosen. Test samples of the components have to be irradiated to see if they can survive the radiation and the best candidate is selected. If the component has passed the radiation test, a new Lot is acquired and more samples are irradiated to verify the same result as from the test samples. When the new Lot has passed this radiation test, it can be used on the final instrument, if it does not fail any of the other tests later on (radiation, EMC, vibration, thermal cycling, outgassing, proton testing (30-300MeV), shock, accelerated life test etc.). The EEE components and the instrument have to pass all the other tests, before it has proven its reliability and quality for space. 3. VARIATIONS IN LOT-TO-LOT AND MANUFACTURING PROCESS FOR COTS. The manufacturing process can be changed at any time without any notice to the enduser of the EEE components. The changes can be minor or major, but the result of these changes can have major impact on the components capability to survive in space due to the radiation environment. The next section will show the consequences, when minor and major changes have been done on the manufacturing process, which will affect the EEE component resistance to radiation. 3.1 Variation due to changes in the manufacturing process. When using COTS for manufacturing instruments for use in space, the buyer of the EEE component have to be aware of the possibility, that the manufacturer of the EEE component can change the manufacturing process without any notice to the end-user of the components. The reason for changing the manufacturing TID testing of PWM controller chip (Lot 1997) Vout process can be many; it can 3,35 be to minimize the number of failures during the chip 3,33 manufacturing or changing rise/fall time for a parameter to increase the speed for 3,31 3,29 the component etc. The impact for instruments built with COTS is 3,27 very critical, if the manufacturing process changes. 3,25 Figures 1+2 shows a radiation test of a controller chip Figure 1. TID results for the new LOT after the manufacturing process have been changed. used in a DC/DC converter (3.3 V output). The TID (Total Ionizing Dose) test shows the -Mode PWM Controller for the DC/DC

3 Converter before (Figure 1) and after (Figure 2) the manufacturing process was changed. The two figures show the output voltage as Vout function of dose rate. For the 3,5 old LOT from 1997 the first 3 symptoms that the irradiation of the chip, starts to impact 2,5 the chip from 52 KRad (the 2 voltage drops), and at 54 1,5 KRad a failure in the output voltage (0.5 volt, can not be 1 seen on Figure 1). 0,5 For the same chip from the Lot 2002 Figure 2 shows that 0 the controller chip fails at 3.1 KRad. This shows, that major changes have been done on the manufacturing process, and it is not only variation from Lot to Lot, as described below. 3.2 Variation due to Lot-to-Lot manufacturing. TID testing of PWM controller chip (Lot 2002) When the EEE components are made at the factory, there are small variations from the dices in each wafer and there are also variations from the wafers manufactured in different weeks. Figure 3-4 shows FiFo's from different weeks and years. It can be seen from the tested samples, that the best FiFo is the one manufactured in 1998 and the poorest is the one from The only way to know if it is a good or a bad Lot is to test samples from each Lot. Test results will show, if any major changes have been done on the manufacturing process or if it is only small variations, due to wafer variations in the Lot-to-Lot production Figure 2. TID results for the old LOT before any changes in the manufacturing process Byte Errors FiFo Byte Errors (1995) FiFo, Bytes Errors for 2 Lots (1999 & 1998) Byte Errors Dose () Figure 3. A FiFo from 1995, showing when errors start to occur as a function of the dose rate Figure 4. Byte Errors for FiFo from 1998 (red) & 1999 (black). The manufacturers of COTS components, do not inform when they make any changes in the manufacturing process. Thus the only way to find at suitable candidate for the in-

4 strument is to irradiate each component to find those, which can tolerate the dose expected to be received during the lifetime of the mission, including a safety margin. 4. LATCH-UP PROTECTING OF THE COMPONENTS. COTS components are not built to be Latch-Up immune, therefore it is necessary to protect each component by a Latch-Up circuit, which monitor the current consumption. If a Latch-Up appears, the Latch-Up circuit will shut down the power line for a short period and then turn-on the instrument again. The monitoring of the current consumption for the instrument has to be divided into several branches to detect a Latch-Up in every component. It is necessary to divide the power distribution into several branches, branches for different voltages and branches for high, medium and low power consumption components. The Latch-Up circuit should be designed so the trigger level for each branch is approx % of the nominal current level. E.g. a high power component like a CPU should have its own power line, because it is difficult to detect a Latch-Up in a branch, if one component uses 700 ma and the rest of the components in the monitored branch only uses 100 ma. A safe way to design the Latch-Up circuit is to use bipolar transistors, which are immune to Latch-Up. The only change to the bipolar transistor caused by radiation damages to the chip is that the gain decreases with the irradiation. A simple circuit to protect a power line is to use two transistors and a few resistors, and then measure the voltage drop across a resistor, which is proportional to the current in the power line. The reliability for the instrument built with COTS will be improved significantly by using a Latch-Up circuit. 4.1 Increased junction temperature during at Latch-Up. For most low power components (COTS), the junction temperature is approx. 5 C above the ambient temperature. An important question is: How much will the junction temperature increase during a Latch-Up and consequently how long time should pass before the Latch-Up circuit triggers and shuts down the power line. The table below show measurement from a 5 volt power line with a nominal current of 350 ma, a Trip level for the Latch-Up circuit of 700 ma. During a Latch-Up, the current and voltage are measured before the Latch-Up circuit shuts down the power line after 2.2 msec. Latch-Up Power Line Nominal ma Trip ma A Voltage V Time msec Energy Joule 5V msec 4.4 mj = During Latch-Up. For Silicium: Cp=712 J/(kg*K), k=2.3 g/cm 3 If we assume that the Silicium chip inside each component is 1 mm X 1 mm X 0.5

5 mm, then the volume of the chip is 0.5 mm 3. The mass of the Si is: m=2.3 g/cm 3 * 0.5*10-3 cm 3 = 1.2*10-3 g?t = E/(m*Cp) = 4.4*10-3 J/(1.2*10-6 kg * 712 J/(kg*K)) = 5.1 K The increased temperature on chip level is approx 5 degrees during a Latch-Up or approx. 10 degrees higher than the ambient temperature, so with the selected Trip and Trigger Time a Latch-Up will not destroy the component. If the Silicium chip is bigger than the assumed size of the chip, then the temperature will increase with lees than 5 degrees. The junction temperature will increase during a Latch-Up, but the junction temperature will not be near the maximum junction temperature (approx. 125 C for COTS), if the ambient temperature is C. A Latch-Up will not induce any risk to the instrument or shorten its lifetime. 5. CONTROL OF THE PROCUREMENT, MANUFACTORING AND INSPECTION. The components procured for an instrument has to be from the same Lot as the ones, which have been radiation tested, to insure the quality of the components. When the components arrive from the manufacturer, they are inspected and controlled to verify, that it is the right component and that all the components have the same Date Code/Lot Number. Samples from the Lot have to be irradiated in order to qualify the Lot. When the instrument is to be built, manufacturing procedures have to be made. Certified personnel solders the PCB boards for the instrument and a Log Book follows the instrument for traceability. This will improve the quality of the instrument. When the instrument has been built, the testing phase starts in order to qualify the instrument for space (functional test, EMC, vibration, shock, thermal cycling etc). When the instrument has passed all the tests, it is ready to be used in space. To ensure that the instrument will survive the lifetime in space exact copies of the instrument are tested in an accelerated lifetime testing (FIT Test). Here several units are tested at an elevated temperature (above 100 C) for several weeks or months. During the test the instruments are regularly controlled to ensure that the units are working normally and that no failure has occurred. If designed and manufactured properly, the test will show that the lifetime of the instrument built with COTS is many years. When the instrument has passed all the tests and inspections, the instrument built with COTS is qualified for use in space. 6. CONCLUSION The space qualification of an instrument built with COTS requires strict control of the components and the manufacturing procedures. During the manufacturing several inspections have to be performed to control the quality of the workmanship, followed by several tests of the assembled instrument. This ensures, that the quality of an instrument built with COTS can be at least as reliable as instruments built with MIL/Space components.