CC BY
24
Simulation Of Reliability For Electronic Means With Regard To Temperature Fields
Artyukhova M., Polesskiy S., Linetskiy B., Ivanov I.
National Research University Higher School of Economics, Moscow, Russia mayaartyukhova@gmail.com, spolessky@hse.ru, blinetskiy@hse.ru, i.ivanov@hse.ru
Abstract
The paper considers the technique of modeling of electronic reliability based on modeling electrical components environment temperature. As experience of the simulation and exploitation of electronic shows, one of the main factors that significantly affect the reliability characteristics is the thermal effect. This is confirmed by the statistics of a number of companies. In the paper for the simulation were used systems ASONIKA-K and ASONIKA-TM. On the example of a real electronic mean proved the need for a point temperature estimate for each electrical component and the account of these temperatures, instead of the average values in predicting the reliability indices. Such approach will significantly improve (20% - 40%) the accuracy of estimates of the mean time to failure. Developed engineering method to predict reliability, built on the "downward" hierarchical circuit simulation.
The reported study was supported by RFBR, research project No. 14-07-00422 a. Keywords: reliability, printed circuit boards, thermal analysis
I. Introduction
Reliability is a complex electronic device property, which, depending on the purpose and conditions of its application consists of a combination of properties: dependability, durability, maintainability and conservation.
Today the actual direction of the reliability theory is the prediction of indicators of reliability in the early stages of design. The direction uses different approaches, one of the key is a methodology for the synthesis of highly reliable electronic means on the criteria of reliability.
Practice of design and operation shows that the greatest impact on the reliability by climatic, mechanical and electrical effects [1]. General failure rate model of the printing assembly of the electronic means in the mode of operation is as follows [2]:
m n
Apa EM = Ka • ^ ^ ^eij (1)
j=1 ¿=1
where: K - quality factor of production equipment, relative units; Aeij - operational failure rate of the i-th type of product j-th group (see model below), 1/h; n - the number of products j-th group, items; m - number of product groups, items.
The model Aeij in general for standard electronic components (chip resistors, chip capacitors, etc.) is as follows [2]:
n
Ien = h(h.g.) • Kr(Kt) • Ke • n Ki (2)
¿=1
where: Ab(Ab.g.) - basic failure rate of type (group) of electrical components, calculated according to the results of tests on the electrical component reliability, durability, life, 1/hr; Kr(Kt) - mode coefficient (temperature) takes into account the magnitude of the electrical load and (or) the ambient temperature (the product's enclosure), relative units; Ke - operating factor takes into account the severity of operating conditions, relative units; Ki - coefficient taking into account changes in operational failure rate depending on various other factors, relative units; n - number of factors taken into account, items.
Affecting electronic factors can be divided into four types of effects, as shown in Table 1.
Table 1: List of external influencing factors
№ Effects Factor name
1 Climatic • high pressure air or gas • reduced atmospheric pressure • changes in atmospheric pressure • Low ambient temperature • Increased ambient temperature • high humidity • atmospheric condensed precipitation • low air humidity • salt mist • solar radiation
2 Mechanical • Broadband random vibration • acoustic noise • linear acceleration • seismic shock • Mechanical shock of single action • Mechanical shock of repeated action
3 Biological • mold fungi • insects • rodents
4 Other • static dust • Dynamic dust • aggressive environment (ozone, ammonia, nitrogen dioxide, sulfur dioxide, hydrogen sulfide)
Objective factors are determined by the time and conditions of use and include the operation time; climatic factors; mechanical factors; biological factors; operating modes. The typical distribution of electrical component failure due to objective reasons shown in Figure 1.
Law temperature
24%
Shock and
2%
Figure 1: The electrical component failure rate of various objective factors
As seen from the model (2) and the real statistical failure (see. Figure 2) for each electrical component makes the largest contribution Kr(Kt), and it, in turn, is determined by the point modeling of ambient temperature (or shell) of the element or of experimental investigations [3]. As shown in Figure 2 for a typical printing assembly change in the ambient temperature of +25 ° C to +80 ° C leads to a change in failure rate of more than 1.66 times.
3.39E-5 ™ 7.64E-5
Sm
<u
^ 7.0SE-5
"S 5.6EE-5 d _o
2 6.03E-E-<V
5,44E-5 5, (J5E-E-
Figure 3: Graph of operational failure rate of a typical printing assembly from ambient temperature
II. Thermal analysis of electronic equipment
As shown in [3] thermal modeling reveals the weakness of the development, to correct them and protect from heat. This approach also allows one to get a more accurate value of time to failure under the conditions of use of the object prior to disposal. Thermal analysis will allow at a stage of simulations increase the value of system reliability indicators in the possible reduction of cost and the geometric dimensions. The fact that the higher the temperature, the lower the reliability. In the presence of operational data can be predicted the mean time to failure for a newly developed product.
Thermal modeling relevant because:
1. Analysis of temperature fields of electronic means - is rapidly expanding area of research;
2. Thermal analysis is applicable to many areas of design;
25 SO SB 40 45 50 55 60 65 70 75 SO ambient temperature
3. Thermal analysis is very important for engineering research.
Model of the failure rate of the temperature Ad. The failure rate for any reference temperature Tr can be calculated using the following equation, and with known Tb and Ab:
The failure rate is doubled by raising to 10 °C ambient temperature (K=293 °C) for AE=0.53. Most electronic solid state components have AE=0.4, and failure rate is doubled when the temperature rises to 13,5 °C.
Cooling systems [3] should be designed to control the temperature of the components. By varying the cooling systems in board electronic means in some cases could increase by 500% the average time to failure.
Implementation of the requirements of the thermal analysis leads to an additional increase in the cost of the design. However, the average cost of a heat-resistant electronic means compensated by saving operating costs.
A thermal analysis of the electronic means should be performed at the system level. Without it, it can happen that parts and components will continue to refuse. Components can be designed to work in normal conditions, but due to the low heat transmission from different heat generators, they can not work at increased temperatures.
There are two main areas in the thermal analysis of electronic means: 1. Knowing electrical component temperature and therefore to quantify the degradation of electrical parameters; 2. Reduce the temperature of the electronic components that improve system reliability. The first may predict "hot" spots in the development through detailed analytical prediction or through direct measurement of heat. The second allows local cooling of these areas that will significantly increase the component life time.
To select the mathematical models for calculating the reliability of foreign and national reference books were analyzed. For the basics reference [6] was taken as the most used and reliable.
Thermal modeling was carried out on the example of a typical printed board assembly of electronic means.
The task is this: to calculate the printing assembly for given thermal actions. Based on the analysis of the thermal characteristics of printed assembly conclude that the technical requirements for electrical components for thermal characteristics performed.
Data for calculation.
The initial data for the calculation of blueprints printed board assembly and output PCAD system files have been received, as well as maps electrical component operating modes. Design printed board assembly subsystem ASONIKA-TM, is shown in Fig. 3 (first side) and Fig. 4 (second side).
The capacity of heat generation electrical component in the PCA: B2 - 0,6 mW; R7 - 30 mW; R8 - 40 mW; R14 - 200 mW; R15 - 200 mW; R17 - 30 mW; R20 - 110 mW; R21 - 10 mW; R23 - 20 mW; R24 - 110 mW; R25 - 10 mW; R26 - 20 mW; R27 - 110 mW; R28 - 80 mW; R29 - 30 mW; D5 -1500 mW; D6 - 1500 mW; D7 - 1500 mW; D20 - 157 mW; VT1 ... VT3 - 40 mW; Total 5777,6 mW.
According to the results of thermal calculation unit in the subsystem ASONIKA-T obtained the following air temperature inside the unit:
• for natural convection 100,2 °C;
• with forced convection blowing speed of 1 m/s 53 °C.
Use the data the temperature values as the boundary conditions for the thermal design of printed assembly.
(3)
Figure 3: Design of printed assembly in the subsystem ASONIKA-TM (side 1)
jY 21.0 I 19 0 I 17D I 15 0 I 13.0 I 11 0 I 90 I 7D I 50 I i'o I 10 I J_22.0 20.0 13.0 160 140 12.0 100 S.O 60 40 2.0 0.0
Figure 4: Design of printed assembly in the subsystem ASONIKA-TM (Side 2)
Results of thermal analysis.
Calculation of thermal characteristics of printed board assembly was held in an automated subsystem ASONIKA-TM. Fig. 5 and Fig. 6 shows obtained thermal characteristics for printed board assembly mode 1 in operation (the air inside the unit for natural convection 100.2 °C) and mode 2 (air inside the unit in a forced convection blowing speed of 1 m/s 53 °C). Maps of thermal modes of electrical component are presented in tables 2 and 3.
Figure 5: Temperature Field for printed board assembly in mode 1 Table 2: Section of the map of thermal modes of electrical component (when stationary thermal action) for of
№ Symbol of electrical components side The temperature of electrical components Coefficient of thermal load, [relative units] Overheat, [°C]
Estimated, [°C] Maximum permissible, [°C]
1 R1 1 111.222 100.000 1.112 11.222
2 R17 1 105.574 100.000 1.056 5.574
3 R18 1 105.445 100.000 1.054 5.445
4 R19 1 105.418 100.000 1.054 5.418
5 R2 1 107.819 100.000 1.078 7.819
6 R20 1 106.025 100.000 1.060 6.025
7 R21 1 105.487 100.000 1.055 5.487
8 R22 1 105.418 100.000 1.054 5.418
9 R23 1 105.479 100.000 1.055 5.479
10 R24 1 106.004 100.000 1.060 6.004
20 C1 1 108.730 85.000 1.279 23.730
179 C95 2 105.153 100.000 1.052 5.153
Table 3: Section of the map of thermal modes of electrical component (when stationary thermal action) for the _printing unit in mode 2_
№ Symbol of electrical components side The temperature of electrical components Coefficient of thermal load, [relative units] Overheat, [°C]
Estimated, [°C] Maximum permissible, [°C]
1 R1 1 64.652 100.000 0.647
2 R17 1 58.946 100.000 0.589
3 R18 1 58.820 100.000 0.588
4 R19 1 58.790 100.000 0.588
5 R2 1 61.206 100.000 0.612
6 R20 1 59.401 100.000 0.594
7 R21 1 58.863 100.000 0.589
8 R22 1 58.790 100.000 0.588
9 R23 1 58.855 100.000 0.589
10 R24 1 59.376 100.000 0.594
20 C1 1 62.169 85.000 0.731
179 C95 2 58.530 100.000 0.585
III. Calculation of reliability printed board assembly
The first and one of the main steps of calculating the reliability of the printed board assembly is to identify the electrical component parameters.
Under the parameter identification should be understand the process of determining the parameters of a mathematical model of reliability calculation, for each specific type of electronic components. The process of identification of electronic components can be represented schematically in the form of an algorithm, illustrated in Figure 7. When performing the
identification of electronic components, according to the algorithm, some points should be noted. Secondly, found in prior specifications, not always this information is sufficient for the calculation of reliability, in such cases, according to the block 9, was searched averaged parameters of technological groups and subgroups in the directory of the reliability of foreign-made product when checking the adequacy of the information. As a result of the identification of all part types from the list, we were assigned to a particular class of electrical component, and in line with the previously selected mathematical models, all the necessary parameters have been found.
Calculation of reliability of the printed board assembly.
An indicator of reliability of printed board assembly is its mean time to failure with no recovery in the process. Reliability of printed assembly is characterized by a set of failure rates of its components (electrical component). The scheme of calculating the reliability of printed board assembly corresponding to a predetermined criterion of failure, is a serial connection of a technologically and functionally combined electrical component groups.
Figure 8 shows the sequence diagram for calculating the reliability of the device included in printed assembly on the level of technology combined electrical component groups.
Operational electrical component failure rate was calculated according to the corresponding reference books [7-9] and a set of maps the correct application of electrical component.
Figure 7: Algorithm for electrical component identification process 91
Figure 9 shows the window ASONIKA-K system with the results of printed assembly calculation (estimation).
- crystal msÜ1 ater Capacitors Integfated circuits JsiiiuciOfi
W w
Resistors Semiconductors Ccruiectors -
Figure 8: The scheme of calculating the reliability of printed assembly
Figure 9: ASONIKA-K system: the results of printed assembly calculation (estimation)
As can be seen from Fig. 9, obtained by calculating the value of the average operating time of printed assembly is ~21,364 thousand [hours] (Electric Load coefficients varying depending on the type of electric components from 0.1 to 0.7 at a temperature of 65 [°C]) that does not satisfy the technical requirements (To=150 thousand [hours]).
Figure 10: System ASONIKA-K: The results of printed assembly calculation (adjusted calculation)
Adjusted calculation printed board assembly reliability.
Adjusted calculation operating electrical component failure rate was based on electrical component temperature, the resulting heat-transfer simulation using subsystem ASONIKA-TM, and other data about the electrical component of printed assembly were taken from the set of maps of the correct application of electrical component.
Fig. 10 shows the window ASONIKA-K system with the results of printed assembly calculation (adjusted calculation).
As can be seen from Fig. 10 obtained by calculating the average value of use of printed assembly is ~ 19,802 thousand [hours] (Load for electric coefficients varying depending on the type of electrical component from 0.1 to 0.7 at temperatures electro obtained by subsystem ASONIKA-TM), which does not meet the technical requirements (T0=150 thousand. [hours]).
Analysis of the results of calculations.
To assess the influence of ambient temperature environment was constructed operational temperature dependence of printed board assembly failure rate in the temperature range +25 ... + 85 [°C] for given values of electrical load coefficient depending on the type of electrical component from 0.1 to 0.7 (see Fig. 11).
ambient temperature Figure 11: Dependence of operational failure rate of printed assembly of temperature
As can be seen from Fig. 11, a simultaneous change of electrical component temperature in the range +25 ... + 85 [°C] causes a change in the intensity of printed board assembly failures in 2 times.
Assessing the impact of the specific characteristics of reliability of electrical component on operational intensity printed board assembly failures carried out directly during the calculation.
Figure 12 shows the contribution classes printed board assembly electrical component to the total failure rate.
Figure 12: Contributions of electrical component classes to the total intensity of the printed board assembly
failure
As shown in Figure 12 of the most unreliable class electrical component is a class "Integrated circuits" and "connectors".
Figure 13 shows the contribution of electrical component class "Integrated circuits" to the total failure rate.
Figure 13: Contributions of class "Integrated circuits" electrical components to the total failure rate
As it follows from Fig. 13 unreliable chips are chips D1, D2 type TMS320VC5416PGE160 and D3-D5 type TPS73HD301PWPR, TPS73HD325PWPR.
Figure 14 shows the contribution of electrical component class "connectors" to the total failure rate. As it follows from Fig. 14, the connectors are unreliable connectors X1, X2 type C 6921 03164.
Figure 14: The electrical component class contributions of "connectors" to the total failure rate
IV. Conclusion
The calculation of printed assembly reliability has shown that:
- At a temperature of 65 [° C] average time to failure is not less than 21.364 thousand [hours.]. Electric load factor depending on the type of electrical component that varies in the range from 0.1 to 0.7;
- At temperatures of electrical component derived from simulations using subsystem ASONIKA-TM, the average time to failure is not less than 19.802 thousand [hours.], For the electric load factor depending on the type of electrical component, varying in the range of 0.1 to 0.7.
Options considered analysis printed board assembly reliability showed that the reliability of the product does not meet the requirements (mean time to failure is to be not less than 150000 hours). The most unreliable electrical component classes are the class of "Integrated circuits" and "connectors". To improve reliability, we can recommend the following measures:
- change the type of electrical component (use electrical components with less Ah);
- to facilitate the operation of the electronic components (lower operating thermal and electrical load);
- reduce the number of electrical component (use the chip higher degree of integration);
- use electrical components with a high level of quality;
- reduce the ambient temperature (to increase the efficiency of the cooling system).
Using the concept of mathematical modeling of complex heterogeneous physical processes in the development of printed board assemblies within systems ASONIKA-K and ASONIKA-TM allows one to improve the accuracy of reliability parameters modeling;
In this paper: 1) proved by the example of printed board assembly need for differential evaluation of the temperature of each electronic components and their integration in predicting reliability, rather than the averaged temperature values; 2) developed a technique of mathematical modeling of reliability for thermal printed board assemblies that the example has proved its effectiveness.
References
[1] GOST RV 20.39.302-98. KSOTT. Equipment, instruments, devices and equipment for military purposes. Requirements for the program to ensure reliability.
[2] RDV 319.01.10-98. KSOTT. Equipment, instruments, devices and equipment for military purposes. Methods of reliability based design and manufacturing of electronics.
[3] V.V. Goldin Information Support lifecycle of electronic / V.V. Goldin, V.G. Zhuravsky, A.V. Sarafanov etc. - M.: Radio and Communications, 2002. - 386 p.
[4] GOST RV 20.39.304-98. KSOTT. Equipment, instruments, devices and location-ment for military purposes. Requirements for resistance to external factors.
[5] OST 4G 0.012.242-84. Radio-electronic equipment. The methodology of calculation of reliability indicators.
[6] Reliability of electrical component: Handbook / 22 TSNIII Defense Ministry. - M.: Ministry of Defense of the Russian Federation, 2006. - 641 p.
[7] RDV 319.01.20-98. The position of the directory "Reliability of Electrical products".
[8] B.S. Sotskov. Fundamentals of the theory and calculation of reliability of elements and arrangements of automation and computer technology/ B.S. Sotskov. - M .: Higher School, 1970 - 270 p.
[9] A.S. Shalumov. ASONIKA automated system for designing highly reliable radio electronic means on the principles of CALS-technologies: Volume 1 / A.S. Shalumov, Y.N. Kofanov, V.V. Zhadnov and others// Ed. Y.N. Kofanov, N.V. Malyutin, A.S. Shalumov. - M.: Energoatomisdat, 2007. - 538 p.
[10] Pisarev, A.V. Test System - the basis of ensuring the reliability / B. Pisarev, Kritenko M., V. Postnov// "Electronica: NTB», № 5, 2002 - p. 32-35.
[11] Reliability prediction of electronic equipment: Military Handbook. MIL-HDBK-217F (Notice 1, 2), 1991 - 205 p.
