Methodology for the Selection of Cots Components in Small Satellite Projects and Short-Term Missions

— From a management point of view, considering increasingly lean and controlled budgets, and a restricted schedule there is a need to find cheaper and more viable alternatives for the present scenario related to commercial electronics parts. On the other hand, the increasing offer of COTS (Commercial Off The Shelf) and the higher quality manufacture are a good opportunity to use COTS components in electronic projects for small satellites in short and medium-term missions, through a coherent study that combines the restrictions and advantages mentioned above, a method that indicates the best COTS for the systems engineer and/or project can be of great usefulness.


I. INTRODUCTION
The current trend in the use of Commercial-Off-The-Shelf (COTS) components, due to its of cost management, development time, availability for purchase and higher quality and reliability achieved by its large scale utilization in automotive electronics, and mobile phone, generates a new approach allowing its use in short-term space projects.
On the other hand, COTS components does not follow the rigor established in the military standards in terms of tests, selection, documentation, and required quality levels , which makes it difficult to track the component since its manufacture and testing.
The main restrictions on the use of these components in the space area are due to the hostile environment to which they will be subject during their useful life since launch until orbit.
In the space environment, some of the main factors that degrade the components are: • Vibration (acceleration) at launch; • Thermal (during the life cycle of the satellite); • Ionizing radiation (total accumulated dose -TID and single-effect events -SEE's) from the trapped particle in the radiation belt around the Earth and solar activity. example, any type of assembly or component through a catalog without any additional testing at the component level. Delivery of the component by the manufacturer as it is.
3.2 COMPONENT SELECTION: Consists of a series of tests and inspections to remove non-conforming components and/or early failure also known as infant mortality (components with defects that are likely to result in initial failures) and thus increase the reliability of the components selected for use.

COMPONENT BURN IN TEST:
Test applied to the electrically polarized component (current or voltage) at an elevated temperature for a specified number of hours. It is a process of accelerated aging and it aims to make the component operate at a maximum nominal value of operating conditions, to reveal intrinsic failure in time and early defects during manufacturing (infant mortality: manufacturing defects).

COMPONENT CHARACTERIZATION:
The process of testing a sample of components in a controlled environment (temperature and acceleration levels) is done using applications and / or setups to measure the electrical parameters of the component. Component characterization results are often used as a basis for establishing batch qualification tests.
3.5 COMPONENT SCREENING: A series of intended component-level tests and inspections to remove nonconformities and child mortality (defective components) and increase the reliability of the component selected for use.

IV. CONTEXTUALIZATION
There are three main reasons for using COTS components in space projects: • Best performance; • The absence of list of qualified parts for space; • It has 1/10 of the cost of QPL equivalent for space The first two items are the main reasons for use in space projects, the lowest cost being the main driver for satellite launchers and constellations (http://wpoaltertechonology.com/accede) [2].
Given the possibility of using some options available in the market, but different manufacturers and unknown quality levels in terms of reliability, we propose to discuss the following approaches: 4.1 Develop a method of choice based on the probability of failure instead of the reliability approach; 4.2 Use the FIDES guide [3] to calculate component failure rate based on physical failure mechanisms (Overstress: thermal, mechanical, relative humidity, subassembly of plates and weld points) and manufacturers' quality factors (manufacturing processes and quality) considering the life stages of the component; 4.3 Introduce a cost based choice of COTS in specific cases, for example the intended COTS of the project does not have sufficient data to prove or demonstrate the required reliability through accelerated environment tests for MTTF inferences and burn-in for the general cases ( up screening).

V. METHODOLOGY
The selection of the appropriate COTS component is not a trivial task and was considered a decision-making process with several criteria. In our case, only two criteria: reliability and cost.
After allocating the reliability of the proposed electronic subsystems / modules, taking into account the minimum reliability established for the subsystem in question (our case study: power module -DC / DC converter), an analysis will be made using the FIDES method to find out the failure rate and a theoretical cost analysis related to the minimum tests necessary for screening in specific cases.

VI. FORMULATION
It has been formulated the selection model considering the following challenges associated with the problem: 5.

Note:
The FIDES Guide deals with manufacturer quality factors (Π_PM and Π_Process) together with the AHP method can be treated with the designers' preference functionalities of the component.

VIII. APPROACHES TO METHOD
In order to begin the discussion around the problem of choosing the appropriate COTS to meet the electrical and environmental functional requirements required in the project, we must stick to the sequence of steps necessary to achieve our goal starting with the allocation of reliability for each unit or subsystem of the mini-satellite so that the mission achieves its goal.
A reliability allocation study must be done previously on subsystem/equipment/ module level before the choosing of the COTS component can start.
Schematic diagrams of the problem and solution are shown in Figures 1, 2 and 3 as follow: Propose Solution:  Once the number or figure of reliability of the subsystem under study is obtained it will be treated on the suggested procedure.
The reliability allocation method is called the AGREE method which is based on the complexity of the unit or subsystem rather than the failure rate. The importance or essentiality of the unit quantitatively defines the relationship between the unit and the target system failure rate and is explicitly considered in the AGREE allocation formula.
The allocation formula is used to determine the minimum acceptable average time of each unit to satisfy the minimum acceptable system reliability. The premise is that the unit within the system has an independent failure rate and operates in series with respect to its effects on mission success.
Unit complexity is defined in terms of the number of modules and associated circuits where a module can be a valve, a transistor or a magnetic amplifier. The unit importance factor is defined as the probability of system / unit failure if a particular unit fails. If the factor of the importance of a unit is 1 the unit must operate satisfactorily for the system to operate satisfactorily otherwise if the factor of 0 then the failure of the considered unit does not interfere with satisfactory system operation.
The specific basis of allocation is to require each module to make an equal contribution to the success of the mission and the equivalent requirement would be for each module to have the same expected average life or failure rate.
The mathematical model for the method considering the approximation: Where: x is small and less than 1 The allocated failure rate of this unit is shown in AGREE.
Where: nj = number of modules (module = electronic component) of subsystem / unit, jth; N = total number of components in the system; EJ = Importance factor of jth unit, and tj = number of hours the jth unit will be required to operate in T system hours (mission time) (0 <tj (duty cycle) ≤ T) The allocated reliability for the jth unit (subsystem) for tj (duty cycle) unit operating hours, R (tj), is given by

IX. BIRNBAUM MEASURE
The importance of a component should depend on two factors [8]: • The location of the component in the PCA / Unit; here we are concerned with a good thermal design in order to reduce thermal stress, understanding that temperature is one of the main factors for component reliability; Birbaum's measurement is then obtained from the partial differentiation of system reliability with respect to pi (t). This approach is well known as a classical sensitivity analysis. If (i/t) is large, a small change in component reliability will result in a large variation in system reliability over time. Let's consider each independent component for analysis, this means that there is no independence between components (obviously this approach does not reflect the actual behavior of systems, this is the interdependence between modules or series components) but already points to a degree of importance. reasonable in its determination.
By noting the fault tree, Birnbaum's measurement [9] can be rewritten: Where: Birnbaum's measure is named after the Hungarian-American professor Zygmund William Birnbaum  Thus, the next step in this methodology would be the search for options for DC/DC converters in the component market that would meet the functional electrical and environmental requirements of the project. For this, we need to find out a failure rate related to DC/DC Converter prescribed and to check if that value is appropriate in our case it means if the value has not compromised the reliability allocated for the unit. Otherwise, it continues to choose another part that meets this requirement. This methodology for reliability evaluation in electronic components has two components:

Reliability Prediction
-Component reliability prediction guide, -Reliability process control and audit guide.
Although component prediction models allow component failure rates to be calculated based on component characteristics and application-related data (eg, applied thermal and electrical stress), the reliability process control and audit guide assess component manufacturing quality and the effects of all processes throughout the life cycle from the design specification phase to maintenance and support activities. The FIDES Guide aims to enable a realistic assessment of the reliability of electronic equipment, including systems operating in harsh environments (defense systems, aeronautics, industrial electronics, transportation, etc.). The general model of FIDES is expressed by the equation: In this case, our components for study: a DC / DC converter (hybrid), A / D Converter (IC) and a semiconductor we apply to formulas to find the failure rate, as follows: Integrated Circuit and Semiconductor Nota: All factors (sensitivity, location, technological, physical stress) and basic failure rate associated with the assembly will be requested in the algorithm of choice. The

XI. COMPLEXITY FACTOR
We will make an analogy with the Karmiol / Bracha [9] (the method used to determine effects factor weights to obtain unreliability and subsequently allocate reliability to subsystem/unit) with the complexity of a component understanding that the problem handling can be analogous. We introduce this factor to increase the stiffness in the reliability calculation since we cannot increase reliability in the usual ways as a redundancy. Where: Kb and Kp should be estimated at the beginning of the development stage. So the complexity factor is: Therefore, for a semiconductor type, we have for example:

Fig 4: Semiconductor Schematic Symbols
In other words, the complexity of the component in this case is low and therefore coherent with the semiconductor diode component. On the other hand, components that are more complex tend to zero.

Low Complexity= 1
High Complexity≈ 0 For an integrated circuit, we have:  /dx.doi.org/10.22161/ijaers.77.14  ISSN: 2349-6495(P) | 2456-1908(O) www.ijaers.com Page |136 After we have been able to determine a failure rate for the COTS via the FIDES method we multiply the result by C (complexity factor) and close the loop in the choice algorithm to see if the value still meets the allocated reliability for the module/circuit in question (one mix between two methods FIDES and HDBK 217). Another thing that must be observed in the case of less complex digital components is that the number of gates will be used as a stiffness factor in the failure rate. Otherwise, we start with a new choice from the available manufacturers. If not apply additional tests like Burn-in or (HALT or MTOL) for MTTF inference and cost analysis, optimizing Reliability versus Cost and having as a reference to MIL 883 or ECSS-Q-ST-60 -13C Class 3 The decision making in choosing the COTS would then be after the analysis of the component failure rate via FIDES Guide and its consistency with a decision based on functional component preferences by the designer through an analysis via AHP and the listed criteria.

XII. COTS RADIATION
Considering the cost of testing in a qualified laboratory in the order of USD1500 per hour and minimum test time required of 60 hours, one can have an idea of the final cost of one of these non-destructive tests since one would be curious to see the functional behavior (some electrical parameters) of a specific component under radiation levels that must be found in the environment provided for in the mission based on Software such as SPENVIS of ESA, OMERE and ANGEL [12] Some radiation of the type TID, SEU and SET can be mitigated by means of some known solutions, such as:

XIII. ENVIRONMENTAL TESTS FOR SPECIFIC COTS
We will start by treating Burn-in tests as the main test for the elimination of defective components (infant mortality), understanding that eliminating the components that may have manufacturing defects, the rest according to the bathtub curve (failure rate versus time curve). ) remains constant with constant failure rate during its "useful life" and "wear" at the end of the project's useful life.
It is understood that the rate of thermal variation predicted in the Burn-in tests (cycle: hot ↔ cold) will induce the mechanisms of physical failure of the component in addition to an acceleration in the aging of the component.

Page |137
Additional accelerated thermal tests in order to verify the MTTF and estimate a batch failure rate will also be carried out and inspections based and adapted according to the reference ECSS-Q-ST-60-13C Class 3

XIV. COST ANALYSIS
All of these tests generate costs so a balanced cost analysis of the type of optimization will be necessary and a risk analysis associated with the problem will be implemented.
Total cost of a Burn in test: Where: A: is the cost of Burn in per unit time L: is the cost of a failure during Burn in R (X) = Distribution curve Probability of Failure (eg: Weibull) [17]

XV. RISK ANALYSIS
This paper proposal aims to present a methodology for the selection of EEE COTS components in small satellites and short duration missions. Notion of cost / schedule and its impacts by developing tests at the level of Components, Cards (PCA) and Boxes [14] Figure 10 shows in the simplified representation that the cost to test decreases while the impact on cost and schedule for correction increases as component, board, and box level testing is performed. This occurs in components because the number of independent tests required decreases when moving to a higher level of testing. The cost of testing may be lower, but the cost and schedule consequences of a failure occurring increase dramatically. Total cost is lower if no problem or failure is detected at higher levels of testing.
We conclude that testing is important for minimizing future impacts and consequences. Therefore, there is a need to find a compromise or to measure and quantify the necessary tests in order to have a reasonable level of confidence for decision making when choosing the COTS to be used.

International Journal of Advanced Engineering Research and Science (IJAERS)
[  Figure 11, in a simplified representation, shows that testing at lower levels of integration improves the ability to detect component defects. Many partial defects are masked at higher levels of integration, but identifying these defects will increase system reliability, reducing the likelihood of latent failures. On the other hand, testing at higher levels of integration is more effective at detecting interactions between component manufacturing and assembly defects that affect reliability.

XVI. DECISION MAKING
As mentioned earlier AHP method [15] can be useful to decision making in the choice of COTS by experts The following is an AHP model for choosing the COTS according to the listed criteria:

XVII. CONCLUSION
This methodology approach is a convenient way to express the system reliability as a function of component reliability and the independence structure between the functional levels considered (subsystem/equipment/module/component) of course there is an interaction among levels but in this study, the values were negligible. Another important point that must be appointed is about the making decisions under many uncertainties considered in this model. By the other hand, these ways suggested give us a possibility to find out the best solution to the designers in the utilization of COTS in an electronic circuit searching a balance among Cost, Reliability, and Risk According to the article [16], we have a graph where it is observed that the power subsystem presents a number of failures very significant.

Fig 13:
Faults observed in the subsystems in Cubes Satellites after injection in orbit, (Dead On Arrival), 30 and 90 days.
Based on this, we will perform an analysis in the power circuit of the Tancredo I Satellite "Tubesat" as a case study according to the proposed method for its validation considering in this case only 3 steps, as follow: Consideration of the importance of reliability of each component (Birnbaum measure) based on the failure rate (λ) of an equivalent component MIL 883 or commercial by (HDBK-217).

Mapping of less critical components;
Using the FIDES Guide to determine the failure rate λ by treating the failure physics plus the Karmiol / Bracha Method (adapted) considering the complexity of the component according to the effect factors: Procedure step by step:  Electrical Scheme of the main power circuit of "Ubatubasat" [18].File TUBESAT_Power.sch

Importance of Reliability
For instance: U11 -DC/DC Converter could be the component of interest for the choice of a COTS considering some aspects related to its importance in the circuit in terms of sensitivity in the reliability of the power subsystem, since it is sensitive but not as much as the voltage regulator, some ceramic capacitors and current sensors as shown in table 4. 18.3 The next step (FIDES Guides) in this methodology would be the search for options for DC/DC converters in the component market that would meet the functional electrical and environmental requirements of the project. For that, it is necessary to find a failure rate of the DC / DC converter and divided by complexity factor in accord to suggest procedure and check if this value is appropriate in our case which means verifying that the value has not compromised the reliability allocated to the unit. In other wise, it continues to find another part that meets this requirement.
Calculation of the failure rate of the DC/DC converter using FIDES guide through the physical failure mechanisms: Failure Rate = λPhysical* πpm* πprocess  For the power subsystem, the reliability rate allocated considering reliability for the whole system of R (t) = 0.9, we would have: P (t = 8760h) = 0.029, that is, the DC / DC converter has around 3% probability of failing up to 8760h (one year) of use considering only one of the power cards (1/2). So it would serve this mission well, as the reliability allocated to the power unit was 95%.
Therefore, can we see that after a new failure rate calculated does not have a great impact on reliability