Iec Tr 62240-2005.Pdf

TECHNICAL REPORT IEC TR 62240 First edition 2005 06 Process management for avionics – Use of semiconductor devices outside manufacturers'''' specified temperature range Reference number IEC/TR 62240 2005[.]

Device selection, usage and alternatives

The equipment manufacturer must ensure that the design prevents any device from exceeding its absolute maximum value throughout its lifespan, even under the most challenging operating conditions This includes variations in supply voltage, equipment device characteristics, control adjustments, load changes, signal fluctuations, environmental factors, and the characteristics of all other electronic devices involved.

LICENSED TO MECON Limited - RANCHI/BANGALORE FOR INTERNAL USE AT THIS LOCATION ONLY, SUPPLIED BY BOOK SUPPLY BUREAU.

Before using a device beyond the manufacturer's specified temperature range, it is essential to conduct a review of alternatives If a reasonable and practical alternative is identified, it should be chosen Additionally, the evaluation results must be documented.

Examples of potential alternatives include:

– using a device specified over the required temperature range, with the identical function, but from a different manufacturer;

– using a device specified over the required temperature range, with the identical function, but a wider specified temperature range;

– using a device specified over the required temperature range, with the identical function, but a different package;

– using a device specified over the required temperature range, that has slightly different specified parameter limits, but which still meets the equipment design goals;

– using a device with the identical function, but a specified temperature range that still meets the application requirement;

– using a device specified over the required temperature range, but a different function, and compensating by making changes elsewhere in the equipment design;

– modifying the device’s local operating environment, for example, adding cooling, etc.;

– modifying the equipment specified ambient temperature requirement, in co-operation with the customer;

– modifying the equipment operating or maintenance procedures, in co-operation with the customer; and

– negotiating with the device manufacturer to provide assurance over the wider temperature range

For most applications, the preferred device for use in a wider temperature range should be the one for which the extension beyond the specified range is least

In situations where the required ambient temperature is 92 °C and no devices rated for above 85 °C are available, preference should be given to the device with a maximum ambient temperature of 85 °C over one rated for 70 °C, assuming no other factors are considered.

Understanding the technology of a device and its packaging is crucial for evaluating potential failure mechanisms It is advisable to consult the device manufacturer when considering usage beyond the specified temperature range If the manufacturer advises against the uprating process for technical reasons, users should weigh these concerns against their specific application needs.

5.1.3 Compliance with the Electronic Component Management Plan

Devices intended for broader temperature ranges must adhere to the equipment manufacturer's ECMP Compliance with ECMP is required solely for the specified temperature range of the device, as this technical report outlines the requirements for extended temperature ranges.

NOTE IEC 62239 is recommended as a resource for an ECMP

Using devices outside the manufacturer's specified temperature ranges is generally discouraged, but may be necessary when other options are impractical or unavailable Justifications for this usage can include factors such as availability and functionality However, it is important to note that such usage should not compromise the design integrity.

– requires the device to operate at an operating or environmental stress level that significantly increases the risk of unstable device operation or loss of equipment function; or

Operating a device beyond its maximum junction temperature or any other specified limiting temperature, as defined by the manufacturer, is essential for optimal performance.

Device capability assessment

Evaluating device capability requires ensuring that both device parameters and the functionality of the device and its associated circuits meet acceptable standards Consequently, it is advisable to conduct functional testing at the application or circuit level, as well as at higher levels.

5.2.1 Device package and internal construction capability assessment

Device qualification test data, along with any other relevant data, must be analyzed to ensure they validate the device's operation across the intended temperature range Additionally, it is crucial that the packaging and internal construction used during device qualification match those intended for the final application.

Device qualification test data, along with any relevant information, must be analyzed to ensure that the package and its internal structure can endure the stresses caused by broader temperature cycling ranges Additionally, it is crucial to verify that the materials used in the package do not experience harmful phase changes or alterations in their properties under these extended temperature conditions.

If this data is not available, then relevant testing based on the application should be considered

Conducting a preliminary risk assessment is essential for guiding decisions on the appropriate methods for capability assessment and determining their application timing By understanding risks specific to each application, organizations can make "risk-informed" decisions, allowing for accurate predictions of the impact of critical choices.

The process for assessing risks should consider applicable factors associated with the use of devices beyond the manufacturers specified temperature range Risk factors in this assessment may include:

– application criticality into which the device will be used;

– consequences of failure at device, circuit and system level;

– type or technology of device under consideration;

– manufacturer data available for the device;

– quality/reliability monitors employed by the manufacturer;

– comprehensiveness of production assembly-level screens performed at extended temperature;

– identification of both managed and unmanaged risks and cost models for each

It is essential to assess the likelihood and consequences of each identified risk, along with acceptable mitigation strategies Typically, these risks can be categorized into specific groups for better management and understanding.

Functionality risks refer to the potential threats that can lead to equipment loss, mission failure, or subpar performance These risks hinder a product's ability to meet customer specifications effectively.

– producibility risks – risks for which the consequences of occurrence are financial impacts

(reduction in profitability) Producibility risks determine the probability of successfully manufacturing/fabricating the product (where “successfully” refers to some combination of schedule, manufacturing yield, quantity and other factors)

Various strategies exist, each offering a distinct combination of risk mitigation factors A preliminary risk assessment's findings should guide the selection of effective methods for ensuring the functionality of devices operating beyond the manufacturer's designated temperature limits.

Device parameter re-characterisation involves evaluating device parameters across a broader temperature range than what the manufacturer specifies, leading to the adjustment of certain data sheet values or tolerances This process enables the device to be utilized in applications where the newly defined parameters meet the necessary functionality To accurately evaluate manufacturing variability, it is essential to consider multiple date codes, acknowledging that this may vary based on application and usage rates.

If device parameter re-characterisation is chosen for capability assessment, then the process described in Annex A shall be followed

If device parameter re-characterisation is chosen for capability assessment, it shall be used in conjunction with a quality assurance process that includes device testing, as described in 5.3.1

Device stress balancing involves operating a device at a temperature higher than the manufacturer's specifications while compensating by lowering at least one other operating parameter, such as power or speed, to ensure that the junction temperature stays below its maximum rating with an acceptable margin.

If device stress balancing is chosen for capability assessment, then the process described in

If device parameter conformance is chosen for capability assessment, then the devices shall be tested over the entire wider temperature range, according to the process described in

Sampling procedures and failure criteria for device testing should be according to Annex C

Where less than 100 % are sampled, then device testing also shall include testing at a higher level of assembly over the entire wider temperature range

5.2.2.4 Higher assembly level testing at temperature extremes

Higher assembly level testing at temperature extremes involves evaluating the device across the full ambient temperature range while it is integrated into a more complex assembly.

If higher assembly level testing is chosen for capability assessment, then the process described in Annex D shall be followed

NOTE 1 A higher level of assembly may include a module, a printed circuit card, another sub-assembly, or the end item

The purpose of sections 5.2.2.3 and 5.2.2.4 is to mandate that each device undergoes testing at least once across its full operating ambient temperature range if testing is utilized to evaluate device capability It is important to note that testing results from higher assembly levels are only relevant to the specific design revision of that assembly; therefore, additional testing or analysis is required for other assembly revisions.

To ensure assembly functionality across the target ambient temperature range, follow these steps: First, perform a Circuit Element Functional Mode Analysis to identify the device functions and parameters that need testing Next, review the assembly level test plan to confirm its ability to test the necessary parameters for successful operation; if it cannot be modified to meet these requirements, reject this uprating method After conducting the test, analyze the results and document your conclusions Finally, include instructions in the maintenance procedures for a full acceptance test across the target ambient temperature range after any maintenance action involving the replacement of an electronic device, unless the maintenance manual specifies adequate alternative procedures This testing should occur at the assembly level where the original capability assessment was conducted or at a higher level.

Device manufacturers typically assess the reliability of their products using consistent processes, irrespective of the specified temperature ranges They do not guarantee a product's lifetime across all applications due to uncertainty regarding usage conditions Therefore, caution is advised when extrapolating reliability from past experiences within the manufacturer's specified temperature range to conditions outside of it.

Assessing the application of each device and its impact on reliability is crucial It is important to identify new or accelerated failure mechanisms that may arise within a broader temperature range and to evaluate their effects on reliability If needed, further testing can be conducted to address any concerns regarding application reliability.

When evaluating a device's reliability, it is crucial to consider the time it operates beyond the manufacturer's specified temperature range Such uprating conditions typically arise in rare "corner conditions" or extreme environments Therefore, consulting device manufacturers is essential for assessing the reliability impacts associated with these scenarios.

The following steps shall be followed: a) Qualify the devices according the requirements of the user’s Electronic Component

Device quality assurance in wider temperature ranges

Regardless of the process used to assure device capability, the quality assurance processes documented in the equipment manufacturer’s ECMP shall be applied to the device

5.3.1 Device parameter re-characterisation testing

For capability assessment through device parameter re-characterisation, it is essential to ensure device quality by testing incoming devices based on a defined sampling plan and closely monitoring supplier change notices.

This guideline aims to ensure that, following the capability assurance activity, there are no alterations in the design or manufacturing processes of the device that could negatively impact its performance across a broader temperature range.

For capability assessment, if device parameter conformance assessment (5.2.2.3) or higher assembly level testing at temperature extremes (5.2.2.4) is employed, device quality must be ensured through device parameter conformance testing, higher level assembly testing (5.3.3), or a combination of both, based on the outcomes of the risk assessment outlined in section 5.2.2.

Refer to Figure 1 for a flow chart illustrating the process For quality assurance, the device assessment must begin with testing each individual device prior to its use in production equipment, or by conducting temperature tests on all production equipment at the extremes of ambient temperature.

Testing may be minimized or waived based on satisfactory test history and effective monitoring of supplier change notices The sampling rate, confidence limits, and decision criteria are outlined in Annex C.

For capability assessment, if higher assembly level testing at temperature extremes or device parameter conformance assessment is utilized, device quality must be ensured through device parameter conformance testing, higher level assembly testing, or both, based on the risk assessment results Refer to Figure 1 for a flow chart illustrating this process When selecting this subclause for quality assurance, a process akin to that in Annex D should be employed to evaluate the assembly test's capability to validate the uprated device at the specified temperature.

Assembly level tests assess the fundamental functional performance of an assembly or device by verifying all key characteristics at the sub-assembly or end-item level Unlike standard procedures, this process traces the device's role in these functions and confirms its capability through assembly testing across the specified temperature range.

Monitoring device data, including product change notices and manufacturer information, is essential to provide alerts regarding changes that could impact the device's ability to function effectively across the broader temperature range specified in section 5.2.

The requirement for monitoring component design and component manufacturing process change data is no different than the related requirement in the IEC 62239 ECMP specification

5.3.5 Failure data collection and analysis

Failure data should be collected for all uprated devices When clear trends are evident, the data should be analysed and corrective action taken

Failures of devices used in wider temperature range should be analysed to establish the root cause of the failure

When failure analysis is conducted, the results shall be documented.

Documentation

For each instance of device usage outside the manufacturers specified temperature range, relevant information shall be documented and stored in a controlled, retrievable format:

The documented information should include:

– equipment in which the device is used;

– manufacturer-specified operating temperature of the device;

– process for assuring device capability in the wider temperature range (including test and analysis results);

– process for assuring device quality in the wider temperature range (including test and analysis results);

NOTE 1 Required signatures include those of the responsible authorities within the equipment designer’s organisation and, if required, those of the customers

NOTE 2 The form of Figure 2 is recommended for use in documenting semiconductor device usage in wider temperature ranges.

Device identification

All device identification processes shall be consistent with other industry processes

Devices that meet the application's wider temperature range requirements through parameter re-characterisation or device testing will be identified as compliant with the design activity's uprating specification This identification must include the design activity's unique identifier, such as the CAGE code, logo, or acronym, along with the assigned part number Each uprated part must be distinctly marked to confirm it meets the application requirements The identification method should facilitate relevant activities, including spares and maintenance, to verify compliance with parameter re-characterisation or device testing.

Devices must have additional markings that complement the original manufacturer's markings, ensuring they remain readable when installed in their intended application All applied markings should be permanent and clearly legible.

Obtain customer approval (if required) Use component

5.2.2.4 Higher level assembly test at temperature range

5.3.1 Periodic re-characterization of parameters

5.1 Component selection, usage and alternatives

5.1.2 Understand component and package technology

Production functional test at temperature capable of validation

5.3.4 Monitor component change data 5.3.5 Collect and analyze failure data

Figure 1 – Flow chart for semiconductor devices in wider temperature ranges

WIDER TEMPERATURE RANGE DEVICE USAGE REPORT

Equipment ID no (if applicable)

Manufacturer's specified device temperature range:

Max Min (Ambient) Max Min (Ambient)

Is the device compliant to the equipment manufacturer’s electronic component management plan?

What alternate solutions were evaluated, and why were they rejected?

The device must operate effectively within the specified temperature range while minimizing the risk of catastrophic failure, unstable performance, and loss of functionality It is crucial to ensure that the device maintains its application-specific reliability under these conditions.

(Reference or attach capability assessment report)

Quality assurance process: Sample plan and monitoring _

Is the device’s quality assured? (reference or attach QA Plan)

Figure 2 – Report form for documenting device usage in wider temperature ranges

Glossary of Symbols

The following terms and definitions are used in this Annex

T rated-max : Maximum temperature at which part manufacturer guarantees operation of a part in accordance with the published data sheet

T rated-min : Minimum temperature at which part manufacturer guarantees operation of a part in accordance with the published data sheet

T req-max : Maximum temperature at which the part is required to operate in a system

T req-min : Minimum temperature at which the part is required to operate in a system

T test-max : Maximum temperature at which the part is tested, usually more than T req-max

T test-min : Minimum temperature at which the part is tested, usually less than T req-min

UL: Maximum limit of a parameter value in the part manufacturer specified temperature range (Specified by part manufacturer)

LL: Minimum limit of a parameter value in the part manufacturer specified temperature range (Specified by part manufacturer)

UL Max : Maximum allowable upper limit for a parameter for proper system operation

LL Min : Minimum allowable upper limit for a parameter for proper system operation

UL New : New maximum parameter value limit if the parameter limit is modified

LL New : New minimum parameter value limit if the parameter limit is modified

M UL : Margin of tested parameter value at extremes of target application temperature range with UL

M LL : Margin of tested parameter value at extremes of target application temperature range with LL

M UL-req : Required margin of tested parameter value at extremes of target application temperature range with parameter limit

M LL-req : Required margin of tested parameter value at extremes of target application temperature range with parameter limit

E: Precision of sampling for mean of a parameter

E A : Measurement inaccuracy σ: Population standard deviation s: Sample standard deviation n: multiplier for standard deviation (typically 3) à: Mean of a population

Part manufacturer specified temperature range

The operating temperature range over which the component specifications, based on the component data sheet, are guaranteed by the component manufacturer

The range between T rated-min and T rated-max

The operating temperature range of the part in its required application This temperature range may be wider than the part manufacturer specified temperature range

The range between T req-min and T req-max

The temperature range over which a part is tested for parameter re- characterisation

The range between T test-min and T test-max

Equation A.1 gives the relationship of the above temperatures: a) T test − min ≤T req − min ≤T rated − min ≤T room b) T room ≤T rated − max ≤T req − max ≤T test − max Equation A.1

It should be noted that for the conditions expressed in Equation A.1 above, parameter re- characterisation can apply for either condition a), condition b) or both.

Rationale for parameter re-characterisation

The re-characterisation process, performed by the device user or a designated test facility, involves measuring electrical parameters across a broader temperature range than specified by the manufacturer Test results may lead to the use of published data sheet parameters or necessitate modifications It is essential to maintain certain data sheet parameters for specific temperature ranges while adjusting others, although the new parameter limits may not apply universally Figure A.1 provides a visual representation of the rationale behind parameter re-characterisation.

Parameter distribution at manufacturer temperature limit

Parameter distribution at application temperature limit

P opul at ion o f de vi ce s

Before conducting any tests, it is essential to analyze data from various sources to assess the feasibility of parameter re-characterization Key data sources include users of comparable devices in similar applications, test laboratories, manufacturers, and industry organizations.

Device manufacturers implement processes to ensure that their devices maintain quality within designated temperature limits Accessing data from these processes can offer valuable insights into a device's anticipated performance across the intended temperature range For instance, simulation models like the BSIM3 model are utilized for analyzing short channel behavior.

MOSFETs, may be used to estimate the effects of temperature variation on device parameters, and therefore may be used to assess their “uprateability [1] 1 ”

Capability assurance

Parameter re-characterisation is a thermal uprating technique that involves characterising part parameters across a specified temperature range, employing methods akin to those utilized by the original device manufacturer When successful, this process allows the device to be utilized in applications where the re-characterised parameters meet the necessary standards.

Figure A.2 shows a flow diagram of the parameter re-characterisation process

It is essential to identify and re-characterize all critical electrical parameters across the entire target temperature range of the application Additionally, one must consider the potential interdependence of datasheet parameters, such as the dependence of logic voltage on supply voltage, when determining which parameters to include.

For effective parameter re-characterisation, it is crucial to use sufficiently large sample sizes to ensure that normal variations remain within the established limits Each instance of parameter re-characterisation requires a tailored sample size determination Additionally, to accurately evaluate device manufacturing variability, it is important to consider multiple date codes, acknowledging that this may vary based on application and usage rates Key factors must be taken into account during this process.

– number of devices available for testing,

– types of parameters to be tested,

– resources required to conduct the tests,

– desired confidence level for the results,

– other factors relevant to the device and the application

1 Figures in square brackets refer to the references given in Clause A.6

For each instance of parameter re-characterisation, the following information should be included in the uprating documentation:

– process used to determine the sample size,

– statistical distribution (assumed or known) of the parameters,

Check available information on part uprateability

Calculate sample size N (equation A.2) and electrically test over T req-min to T req-max + margin is discontinuous, non- monotonic or functional failures found

Establish T req-min , T req-max

Choose required confidence level, precision on parameter mean and population standard deviation

Calculate confidence interval X (equations A.3 or A.4)

Calculate test equipment measurement error E A

Functional test of LRU over operating temperature range

Increase sample size/improve test capability

Figure A.2 – Flow diagram of parameter re-characterisation capability assurance process

In most instances, a normal distribution is assumed For the normal distribution, the sample size, N, is [2][3]:

E is the required precision on the parameter mean,

Z α /2 is the value of standard normal variable at confidence level (1 – σ) ×100 %, and σ is the standard deviation of the population

In this equation, the sample mean is within ±E of the true mean, with a probability of (1 – σ)

Table A.1 shows the results of an example calculation (standard deviation values were obtained from typical device data)

Table A.1 – Example of sample size calculation

NOTE 1 If it is known that another distribution fits the data, then appropriate statistics should be used

In rigorous analysis, each parameter at different test temperatures has a unique distribution However, such detailed information is often not accessible in practice Therefore, unless proven otherwise, it is reasonable to assume that the same distribution parameters are applicable to all electrical test parameters across various test temperatures.

For large sample sizes, i.e., greater than thirty (case 1), the confidence interval estimate of the mean is:

For small sample sizes, i.e., less than thirty (case 2), then the Student’s t distribution should be used, and the confidence interval estimate of the mean is:

Where s is the sample standard deviation and t α /2;N-1 is the value of the Student's t distribution at the confidence level (1 – α) ×100 % and N – 1 degrees of freedom

Parameter re-characterisation tests must be performed across the full target temperature range, including margins above the maximum and below the minimum temperatures It is crucial to understand the device's absolute temperature limits during this process, and any exceedance of these limits should be carefully controlled to enhance understanding of device behavior Devices in practical applications must not surpass their absolute maximum ratings.

Testing should occur at multiple temperatures within the specified range, with the number of test temperatures and their intervals varying for each parameter re-characterization Key factors to consider when selecting these test temperatures include various influencing elements.

– device manufacturer’s specified temperature range;

– other thermal data obtained from the device manufacturer, for example, thermal conductivity, etc.;

– other uses of the test data, for example, performance derating; and

– previous relevant experience with the device

Additional test temperatures can be determined based on tests carried out during parameter re-characterization For instance, if a graph shows that the relationship between a specific parameter and temperature may not be linear, further testing is necessary to accurately define this relationship.

A device may satisfy its parameter specifications, but still fail to function in an application

Functional testing at both the application and circuit levels is essential for digital devices To achieve specific fault coverage, gate level design information is necessary for software development Without a complete set of test vectors, determining the percentage of fault coverage becomes challenging without in-depth knowledge of the device architecture Therefore, it is crucial to prioritize functional testing at these levels.

Testing may be performed in-house or at an external test house In either case, the equipment supplier is responsible for the tests and their results

Prior to parameter re-characterisation testing, a set of requirements and limitations on the electrical parameters should be developed The requirements and limitations depend on the application

Acceptable upper (M UL-req ) and lower (M LL-req ) margin limits should be established for each modified parameter

A.3.2.4 Assessment of electrical test results

A successful uprating process is indicated by test results showing no functional failures, the absence of discontinuities in parameter versus temperature plots, and the acceptance of modified parameter limits for the application.

A.3.2.5 Re-characterised parameter value calculation

Re-characterised parameter values encompass both the nominal values and their associated limits These limits are established by integrating variations arising from sampling, parameter values, and the accuracy of test equipment with the nominal values The method of this combination is depicted in Figure A.4.

The nominal value of a re-characterised electrical parameter is the designated value used for designing equipment with the re-characterised device This value can either remain constant across the target temperature range or vary predictably with temperature Typically, the nominal value is determined as the mean of test results for a specific parameter at a given temperature, although alternative values may be selected if justified.

Figure A.3 – Margin in electrical parameter measurement based on the results of sample test

Variation due to sampling is the confidence interval described in A.3.2.2 It is shown as 2×E in Figure A.3

Parameter variation, represented as \$n \times s\$ in Figure A.3, typically uses the standard deviation of the test sample as its measure, with the number of standard deviations, \$n\$, based on acceptable risk levels Additionally, the accuracy of test equipment, denoted as \$E_A\$ in Figure A.3, is calculated using standard methods from basic statistics and may fluctuate depending on the test temperature.

The parameter margin, M, is calculated by:

When the parameter margin is deemed insufficient (M < 0), it is necessary to adjust the data sheet parameter limits to establish new boundaries for equipment design This modification process starts with determining the appropriate margin for the specified temperature.

The potential variations outlined in sections A.3.2.5.1, A.3.2.5.2, and A.3.2.5.3 are applied to the nominal value of the parameter at the specified temperature If the resulting modified parameter values exceed the established maximum and minimum limits defined in A.3.2.5, the device cannot be uprated through parameter re-characterization.

Figure A.4 shows an example of the parameter limit modification process In this example, the new parameter limit is below the maximum allowable parameter limit, and thus acceptable

Table A.2 shows an example of re-characterising a 0 °C to 70 °C rated part to a –55 °C to

Maximum allowable parameter limit (UL Max )

New parameter limit (UL New ) Measurement accuracy (E A ) n × s spread

Precision - half of the confidence interval (E)

Figure A.4 – Schematic diagram of parameter limit modifications

Table A.2 – Parameter re-characterisation example: SN74ALS244 Octal Buffer/Driver

Parameter Commercial limit Military limit Measured value at military limit Derated limit

14,50 26,00 a) Assumes same degree of errors and standard deviation at all temperatures The margins at the commercial temperature limit are maintained at military temperature limit

To ensure the intended functionality of devices that have undergone parameter re-characterization, a representative sample of the assembly must be tested The success of the uprating process is confirmed only when the higher-level assembly operates correctly.

Quality assurance

To ensure the ongoing quality of successfully uprated devices, it is essential to monitor the process change notices (PCN) from the manufacturer or distributor Additionally, conducting equipment-level tests across the target temperature range, with an appropriate margin, is crucial Rigorous functional testing must be performed to verify that all system functional requirements are met.

LRU test over equipment operational temperature range

Functional and safety requirements verified

Component quality not assured NO

Figure A.5 – Parameter re-characterisation part quality assurance

Factors to be considered in parameter re-characterisation

Data used in initial uprateability assessments may not be an accurate indicator of expected future performance

Simulation models require careful use, especially when publicly shared, as they are often modified to conceal proprietary information, potentially compromising their accuracy in reflecting the actual behavior of devices The uprateability assessment process should solely serve to filter out unpromising candidate devices rather than replace essential electrical testing.

Data sheets do not always list all electrical parameters This is especially true for degradation type parameters for example, gate current, substrate current, trigger currents for latchup, etc

While manufacturer-specified temperature ranges may deem certain parameters unimportant, they can become significant at target temperature ranges In the absence of manufacturers' test procedures, measuring these parameters proves challenging, necessitating estimation instead.

To ensure accurate assessment of system variations, it is recommended to test systems beyond their specified temperature limits However, exceeding these temperature specifications can lead to overstressing other components, potentially causing invalid failures.

If the initial assessment indicates that any such parameters could be of concern at the target temperature, then the revised datasheet should include limiting values for these parameters

Some device lots may include outliers [5], which limits the efficacy of sample testing (see

Products subject to OIMS (Outlier Identification and Management System) disposition [5]

Figure A.6 – Schematic of outlier products that may invalidate sample testing

To effectively detect discontinuities in parameter versus temperature curves, the test temperature range must align with the target temperature range If the intervals between test temperatures are excessively wide, critical discontinuities may be overlooked In cases where test results reveal non-monotonic behavior, it may be necessary to incorporate additional temperature points An example of this is illustrated in Figure A.7, which depicts an intermediate peak of an electrical parameter.

V fb v ol tage f eed back th res hol d c hange m V

Figure A.7 – Example of intermediate peak of an electrical parameter: voltage feedback input threshold change for Motorola MC34261 power factor controller [4]

During initial characterisation, it is also necessary to check for hysteresis in electrical test data, but hysteresis tests can be eliminated if the testing does not reveal its existence

Hysteresis can occur during temperature characterization due to changes in part characteristics from extreme temperatures or failure to reach thermal equilibrium To address hysteresis effects, it is advisable to increase dwell times at the specified temperature If this approach does not yield results, further investigation into other potential damage causes is necessary If no alternative reasons for failure are identified, the device may not be suitable for uprating.

Inflection points in electrical parameters at extreme temperatures must be tested to identify whether the failure is hard or soft It is also essential to ascertain if these failures stem from changes in device characteristics due to device failure or from the impact of extreme temperatures on testing fixtures and equipment.

General

Stress balancing leverages the trade-off between power and temperature in specific applications, requiring less testing compared to parameter conformance assessment and re-characterization Testing is primarily conducted to validate analytical results within the application context For a detailed definition of stress balancing, refer to Clause 3.

Glossary of symbols

T A-Max : Manufacturer-specified maximum ambient temperature

T Up-Max : Maximum temperature up to which the device can be uprated

T App : Ambient temperature limit required for the application

T M : The margin by which the Iso-T J curve is derated

∆T A : Change in the ambient temperature

P Min : Minimum power dissipation at which the device can be operated in the system, calculated from maximum allowable limits on electrical parameters

P Max : Manufacturer specified maximum power dissipation at T A-Max

P App : Power dissipation of the device at application temperature limit, T App

P’ App : Power dissipation of the device at the application limit, without margins on the Iso-

P M : The derating achieved in power dissipation, as a result of the margin put on the Iso-

∆P: Change in the power dissipation

C PD : Power dissipation capacitance/buffer

C L : Load capacitance/buffer f: Frequency of operation of the device θ JA : Junction to ambient thermal resistance

Stress balancing

P is the power dissipation, and θJA is the junction to ambient thermal resistance

Maintaining a constant junction temperature in a semiconductor device ensures stable performance Power dissipation, influenced by electrical parameters such as operating voltage and frequency, allows for a trade-off between ambient temperature and these parameters According to Equation B.1, a higher ambient temperature can be tolerated if power dissipation is sufficiently minimized to maintain the junction temperature The stress balancing process is detailed in steps B.3.2 to B.3.7 and illustrated in Figure B.5.

B.3.2 Determine the ambient temperature extremes

B.3.3 Determine parameter relationship to power dissipation

The objective of this step is to identify the electrical parameters that can be derated and to quantify the extent of this derating Additionally, it involves calculating the necessary reduction in dissipated power for the proposed application to effectively uprate the device.

When selecting parameters to minimize power dissipation in devices, it is crucial to consider factors such as device technology, family, and electrical function For instance, CMOS devices exhibit a linear relationship between power dissipation and effective operating frequency, while lowering the operating voltage can also decrease power consumption Additional strategies include reducing output current, fan-out, or adjusting the device's duty cycle Achieving the desired reduction in power dissipation may require multiple parameters to be adjusted simultaneously.

B.3.4 Determine the dissipated power vs ambient temperature relationship

The objective of this step is to create a graphical representation that illustrates the relationship between device power dissipation and ambient temperature, as outlined in Equation B.1 This involves plotting power dissipation against ambient temperature while maintaining a constant junction temperature An Iso-T J plot, exemplified in Figure B.1, can be generated using one of the two methods detailed in sections B.3.4.2 and B.3.4.3.

A generalized Iso-T J curve, illustrated in Figure B.1, represents power dissipation values ranging from the device's minimum power dissipation (P Min) to the maximum specified power (P Max) To accommodate potential inaccuracies in data and calculations, the curve is adjusted downward towards the horizontal axis by an appropriate margin.

The junction temperature margin (T M) is crucial for determining the application power (P App), which is derived from the Iso-T J curve and includes a margin (PM) as shown in Figure B.1 This results in a defined temperature range above the maximum operating ambient temperature (T A-Max) The temperature associated with P Min represents the highest temperature at which the device can safely operate within the application.

This is denoted by T Up-Max As such, the area bounded by P Max – P Min – I – I’ is the uprated operating area

Derating involves reducing thermal, electrical, or mechanical stresses on electronic components to levels lower than the manufacturer's specified ratings This practice allows for the uprating of a component by adjusting one or more parameters accordingly.

Inaccuracies in power dissipation calculations, the assessment of thermal characteristics, and the lack of precise thermal resistance data can significantly impact performance evaluations.

Combinations of power and temperature values in this area correspond to junction temperatures lower than that established by the Iso-T J Curve with margins I – I’

B.3.4.2 Constructing the Iso- T J curve using thermal resistance

If Equation B.1 is modified as follows:

T =−θ × + Equation B.2 then a plot of power dissipation vs ambient temperature yields a straight line with slope –θ JA

The Iso-T J curve is defined as the line that intersects the point (T A-Max, P Max), where P Max represents the maximum power dissipation specified by the manufacturer at the maximum operating ambient temperature, T A-Max.

To accurately plot the Iso-T J curve, it is essential to know the junction to ambient thermal resistance, often provided as θ JA in thermal characteristic sections of data sheets This thermal resistance is influenced by various factors, including the thermal conductivity of the printed circuit board, the proximity and power dissipation of nearby devices, airflow speed and pattern, coolant properties, die size, and the thermal radiation characteristics of surrounding surfaces These factors primarily affect the thermal impedance from case to ambient, and not all can be precisely modeled during the early design phase Therefore, collaboration between electrical and mechanical designers is crucial to determine the thermal resistance data using the best available information It is advisable to obtain the test or simulation conditions from the device manufacturer before utilizing the thermal resistance data, and any data used in the initial design stage should be validated through testing in the application environment later in the development process.

A m bi ent t em perat ure

Figure B.1 – Iso- T J curve: the relationship between ambient temperature and dissipated power

B.3.4.3 Constructing the Iso- T J curve using thermal analysis

Thermal simulation software is essential for assessing device performance in specific applications, ensuring that the software's temperature and power range aligns with the requirements The initial step in stress balancing involves creating a thermal model of the device.

The device under test must undergo thermal simulation within the manufacturer's thermal test setup to ensure the model's validity by comparing the results with the manufacturer's thermal data Additionally, the application environment should be modeled using various power dissipation values within the device Finally, it is essential to establish the power-temperature relationship derived from the application thermal model.

The value of thermal resistance is influenced by power dissipation and temperature, which is considered in thermal simulations but not in the simplified single parameter approach outlined in B.3.4.2, where θ JA is assumed to be constant Consequently, the Iso-T J curve derived from thermal analysis or simulation may exhibit non-linear behavior, making the thermal characterization in this subclause more precise.

B.3.5 Assess applicability of the method

If the required maximum system temperature is lower than T Up-Max , then the device can be uprated by stress balancing If the required maximum system temperature is greater than

T Up-Max , then the required power dissipation is lower than P Min , and other options should be considered

A horizontal line representing the desired ambient temperature intersects the Iso-T J curve at a specific power dissipation value, which corresponds to the maximum allowable power (P App) that the device can dissipate at the application temperature.

All selected electrical parameters are then modified to maintain the device power dissipation below P App

B.3.6 Determine the new parameter values

Figure B.2 illustrates a generalized relationship between an electrical parameter and dissipated power within the allowable range of power dissipation, defined by P Min and P Max The vertical line represents the application power dissipation, P App, of the upgraded device The intersection of this line with the vertical axis indicates the modified value of the electrical parameter due to stress balancing.

Application example

This example is presented courtesy of the University of Maryland, CALCE from a 1999 paper:

"Stress Balancing: A Method for Use of Electronic Parts Outside the Manufacturer Specified

The Fairchild MM74HC244 serves as an example to demonstrate the stress balancing process in a hypothetical system, with a maximum ambient temperature of 85 °C and a minimum of –40 °C This component is an octal buffer/driver, highlighting its significance in temperature-sensitive applications.

A 3-state buffer is commonly utilized to buffer a bus prior to connecting to input or output devices This CMOS logic device operates within an ambient temperature range of –40 °C to 85 °C and is offered in a plastic dual-in-line package The data sheet suggests a supply voltage (V CC) of 2 V, 4.5 V, or 6 V for optimal performance, with an absolute maximum rating for power dissipation specified.

At 65 °C, the device has a power dissipation of 600 mW, but with a derating factor of –12 mW/°C for temperatures above this threshold Consequently, the maximum power dissipation at the maximum operating ambient temperature of 85 °C is limited to 360 mW Additionally, the system operates with digital logic levels of 4.4 V.

(minimum) for high and 0,1 V (maximum) for low From the data sheet, this requires V CC = 6 V

B.4.2 Determine the ambient temperature extremes

The system ambient temperature range is –40 °C to +85 °C, however it is assumed that a

A temperature increase of 15 °C is observed from the external ambient temperature to the area surrounding the equipment, caused by internal heating during operation Consequently, the device operates within an ambient temperature range of –40 °C to +100 °C.

B.4.3 Select the parameters that can be derated

CMOS devices typically have negligible quiescent power consumption compared to the power dissipation during switching The power dissipation (P) of a CMOS device is given by the equation:

C PD is the power dissipation capacitance,

C L is the load capacitance, f is the switching frequency,

I CC is the quiescent supply current, and

V CC is the supply voltage

From the Fairchild data sheet:

I CC (maximum specified, for V CC = 6 V) = 160 àA

The load capacitance (C L ) is assumed 50 pF, which is the value used for test conditions in the data sheet

To minimize power dissipation, it is essential to either lower the supply voltage or decrease the operating frequency However, altering the supply voltage (V CC) would directly impact the system's logic levels, making it an unsuitable option Consequently, in this scenario, the focus is on reducing the operating frequency from its maximum capability to achieve lower power dissipation.

The device operates at a frequency of 13 MHz with a power consumption of 360 mW at a supply voltage of 6 V It is essential for the system to function at a minimum frequency of 3.5 MHz, which corresponds to a minimum power dissipation of 100 mW.

In CMOS circuits, the exact value of load capacitance, denoted as \$C_L\$, is often uncertain The main factor contributing to load capacitance in digital logic circuits is the additional capacitance introduced by the printed wiring board (PWB) during device installation.

For accurate power dissipation calculations, it is essential to initially use a conservative maximum estimate of $ C_L $ Once the printed wiring board (PWB) design is completed, the actual value of $ C_L $ can be substituted back into Equation B.3 for a more precise estimate.

B.4.4 Construct an Iso-Tj Plot

An Iso-T J curve for the MM74HC244 plotted from the Fairchild data sheet, is shown in

Figure B.3 The specified maximum power is 360 mW at a maximum ambient temperature of

85 °C, so the curve passes through the (360 mW, 85 °C) point

B.4.5 Determine whether or Not the device can be uprated

The minimum power at which the buffer can operate in the system is calculated as 100 mW

(see B.5.2) The slope of the Iso-T J curve is –83,3 °C/W (corresponding to –12 mW/°C power derating), which is higher than the thermal resistance value quoted by Fairchild for this device

The device can be operated at a maximum ambient temperature of 100 °C, provided that the power dissipation remains below 170 mW The maximum uprated ambient temperature (T Up-Max) for the device is 106 °C, with a thermal resistance of 61 °C/W At the maximum application temperature of 100 °C, the new power value is recorded at 170 mW.

5 It should be noted that there are eight buffers in the MM74HC244

A m bi en t t em perat ure ° C Target application

Ma x P D (us ing 83, 3 °C /W ) Min P D al lo wabl e f or s yst em operat ion = 100 m W

Target appl ic at ion power di ss ipat ion = 170 m W

Figure B.3 – Iso-TJ curve for the Fairchild MM74HC244

B.4.6 Determine the new parameter values

As specified in B.4.3, the minimum power at which the buffer can operate in the system is

The output power ranges from 100 mW to a maximum of 360 mW The relationship between operating frequency and power dissipation is illustrated in Figure B.4, derived from Equation B.3 According to the data sheet values and application conditions, the curve can be expressed by the equation: \$P = 2.88 \times 10^{-8} f + 9.6 \times 10^{-4}\$ (Equation B.4).

Figure B.4 is plotted with the data sheet specified values for the various parameters, with

V CC = 6 V and C L = 50 pF It is also noted that there are eight buffers in the MM74HC244N

The device operates at a frequency of 6 MHz with a power output of 170 mW and can function at an ambient temperature of 100 °C, provided the frequency remains at or below 6 MHz Additionally, this frequency may be further reduced based on the design practices of the equipment manufacturer.

In this example, a basic logic device is utilized; however, for more intricate devices, the assumption that the maximum operating junction temperature equals the average die surface temperature may not hold true due to the presence of hot spots It is advisable to obtain this information from the device manufacturer when feasible, or alternatively, to apply a higher margin above the maximum junction temperatures indicated in Figure B.1.

B.4.7 Conduct parametric and functional tests

Normally, the device is tested at T A-Max to ensure that the device will operate satisfactorily in the required environment

Rated power dissipation at rated maximum T A ,

Target application maximum power dissipation

New value of frequency = 6 MHz

Value of frequency at rated maximum

Assumed minimum frequency of the buffer required (or can be tolerated) by the system = 3,5 MHz

Figure B.4 – Power versus frequency curve for the Fairchild MM74HC244

Other notes

When calculating stress balancing, it is essential to consider the manufacturer-specified ambient temperature and dissipated power, which include intended margins To enhance accuracy, it may be wise to incorporate an additional margin (P M) when determining the application power (P App) to account for potential inaccuracies in thermal modeling, the value of θ JA, and the actual ambient temperature near the device Furthermore, the electrical parameters derived from the derated power dissipation should also undergo derating to address inherent errors in the process, while ensuring that adequate precautions are taken during this procedure.

To ensure reliable performance, parametric pass/fail tests should be conducted on devices after stress balancing analysis at the target application temperature, plus an appropriate thermal margin This approach guarantees that sufficient margins are maintained at the intended operating temperature.

Although stress balancing appears to be straightforward, there are certain hidden difficulties

When calculating stress balancing, it is crucial to exercise caution with data sheet junction temperature limits, as they do not represent the maximum operational junction temperature of the device Additionally, the thermal resistance values of a device vary based on its specific application.

If they are used to construct the Iso-T J curve, thermal validation by test or analysis should also be done

The power dissipated by a device refers to the energy lost as heat, which differs from the output power In CMOS devices, all the power consumed is dissipated since the output current is minimal However, some devices may only specify output power in their data sheets, necessitating the calculation of dissipated power using additional information from the manufacturer.

Stress balancing is applicable solely for upgrading devices beyond their rated maximum temperature limit The equation provided in Equation B.2 is not relevant for assessing devices intended for operation below the manufacturer's minimum specified temperature limit For devices operating within a lower temperature range, alternative methods such as parameter re-characterization should be considered.

B.4.3 Select parameters that can be derated

B.4.5 Can uprating be done through stress balancing?

Explore other uprating method or choose another component

B.4.6 Determine the value of derated parameters

B.4.7 Confirm component functionality using derated parameters at

B.4.2 Determine T A-max in the application

Figure B.5 – Flow chart for stress balancing

1 Rated temperature range (case or ambient):

2 Usage temperature range (case or ambient):

5 Iso-T J plot (please attach separate sheet):

6 Amount of margin (or derating) below junction temperature maximum to be used:

7 Power dissipation in device with derated parameter:

8 Derated value of parameter, chosen from (4):

9 Verification test results agree with calculated value (document results and attach):

Figure B.6 – Report form for documenting stress balancing

General

Device parameter conformance assessment consists of evaluating electrical parameters at target thermal test points that are higher or lower than the manufacturers’ specified ratings

For this uprating method, the specifications, conditions, and test limits used are the manufacturers published data sheet parameters

The following references to statistical methods are suggested as tools in determining the statistical confidence in the options listed in this annex

Montgomery D C., and Runger G C., "Applied Statistics and Probability for Engineers," John

Wiley and Sons, Inc., New York, 1994

Pfaffenberger R., and Patterson J., "Statistical Methods," Irwin Publishers, 1987

Test plan

A test plan should be defined that documents the required tests, process steps, test methods, and number of samples Figure C.3 shows a flow diagram of parameter conformance assessment testing

Identifying all critical application parameters and obtaining values from the manufacturer's data sheet is essential While testing all electrical parameters is ideal, it may not always be feasible due to the unavailability of a complete functional test program from the manufacturer Therefore, the specified test parameters, excluding test temperature limits, should serve as the performance specifications for assessing parameter conformance.

When developing a test plan for parameter conformance assessment, it is essential to consider the specified and target temperature ranges of the device, as well as prior experience with it The plan must detail the temperatures for testing and the sample size at each temperature Target temperature data can be derived from the initial assessment of the equipment's environmental requirements and thermal analysis results Additionally, fluid dynamic conditions, such as air speed and direction, should be factored in if indicated by the initial assessment Utilizing all available thermal environment data is crucial for calibrating the test equipment to accurately reflect the application environment Furthermore, the type of temperature specification—whether ambient, case, or junction—should guide the selection of the test temperature range, along with the accuracy of the thermal assessment method and the test itself.

These test margins provide additional confidence in the applicability of the test results over the target temperature range Figure C.1 shows the relationship of the various temperatures

Test temperature range Target temperature range

T test-min T req-min T rated-min T room T rated-max T req-max T test-max

Figure C.1– Relationship of temperature ratings, requirements and margins

The two options for parameter conformance assessment are (a) Test at the minimum allowable margin, C.2.4.2, and (b) Determine the margin by incremental temperature testing,

The chosen option must undergo a rigorous quality assurance process to ensure that future devices meet the required parameter values at specified temperatures It is crucial to understand the absolute temperature limits of the device during testing, as exceeding these limits should only be done to gain further insights into device behavior and margins Devices intended for real-world applications must not surpass their absolute maximum ratings While a device may meet its parameter specifications, it could still fail in practical applications; thus, functional testing at both the application and circuit levels is essential.

C.2.4.2 Test at the minimum allowable margin

This option involves testing parameters at target temperatures that exceed the specified maximum or fall below the minimum temperature limits It is essential to incorporate sufficient margins by adding to the target temperature limits when they are above the maximum, or subtracting when they are below the minimum.

NOTE Typical margins are 2 °C to 5 °C

To ensure the success of the parameter conformance process, sample sizes must be sufficiently large to achieve the desired statistical confidence It is essential to document the required confidence level, the calculation method, and the results for each device evaluated through parameter conformance assessment.

Parametric pass-fail tests should be conducted for all critical electrical parameters being assessed The parameter test limits should be those of the device manufacturer’s data sheet

If this process is successful, then the quality assurance process of C.2.5 should be followed

If the process is unsuccessful, then either the device should not be considered uprateable, or another uprating process should be considered

C.2.4.3 Determine the margin by incremental temperature testing

Devices are evaluated at or near their maximum or minimum specified temperature limits, followed by testing at progressively higher or lower temperature increments, typically recommended at 5 °C, until the parameters of interest fail to meet the data sheet specifications The temperatures at which non-conformance occurs are recorded, allowing for the establishment of new temperature specification limits based on this distribution.

CI X is the confidence interval of the mean,

A is the number of standard deviations for the margin, and

CL σ is the confidence level for standard deviation

TE is the margin to account for test equipment error, then the device is capable of being used in the application in question

C.2.4.3.2 Example: Determine the margin by incremental temperature testing

The following conditions are assumed for this example:

– device temperature rating of 0 to 70 °C,

– application ambient temperature range near device is –40 °C to +85 °C,

– initial test sample is 10 devices,

– the mean fallout temperature of the 10 devices is 130 °C,

– the standard deviation of the mean fallout temperature for the 10 devices is 6 °C,

– the confidence level chosen for this application is 95 %,

– the margin or "A" term in Equation 10 is chosen to be 4σ

(Test equipment error is not accounted for in this example)

Sample sizes for this option are established based on the observed statistical distribution of non-conformance temperatures, including the distribution parameters, the desired temperature margin, and the required statistical confidence in the results It is essential to document all this information for each instance of the activity.

C.2 illustrates the relationship between mean fallout temperature, T req-max and a typical curve representing the device fallout probability at each temperature

P erc ent f al lo ut

Figure C.2 – Typical Fallout Distribution versus T req-max

C.2.4.3.2.2 Confidence interval for mean fallout temperature

The equation for calculating the one-sided confidence interval for the mean when the variance is unknown, is

CI X = γ , n − 1× Equation C.2 where t γ ,n–1 is the percentile of the t distribution, γ is the confidence limit, n–1 is the degrees of freedom n is the sample size, and

For this example, the following confidence interval would result:

C.2.4.3.2.3 Confidence level for standard deviation of mean failure temperature

The equation for calculating the confidence level for the standard deviation is;

− σ n n S χ , Equation C.3 where σ u 2 is the variance, and Χ 2 1– γ ,n–1 is the percentile of the Χ 2 distribution

This equation actually computes the upper confidence interval of the variance (σu 2)

The confidence limit for the standard deviation is the square root of the variance of the confidence limit For this example, the following confidence limit would result:

At a 95% confidence level, it is expected that no more than 60 parts per million (ppm) will fail to meet the manufacturer's specifications at or below 87 °C If this reliability is adequate for the intended application, the device can be utilized, provided that ongoing monitoring of the device or product is implemented as referenced in section C.2.5.

Quality assurance may be established by device level test per 5.3.2 prior to assembly or higher level assembly testing per 5.3.3

See option B example for guidance

Determine confidence interval/level for mean and SD

Test critical parameters over margins selected

C.2.4.3 Determine temperature increments for test and initial number of samples

C.2.4.3.2 Find initial mean and SD for failure temperature

Choose other method or component

Test additional components and re-calculate mean and SD

Have enough samples been tested?

Apply confidence interval/level for mean and SD to determine margin

Figure C.3– Parameter conformance assessment flow

(to be completed for each lot)

4 #Functional and parametric test passed:

(Record all test temperatures, and list results for each temperature separately.)

5 #Functional or parametric test failed:

(Record all test temperatures, and list results for each temperature separately.)

Figure C.4– Report form for documenting parameter conformance testing

Tiêu đề	Process Management for Avionics – Use of Semiconductor Devices Outside Manufacturers' Specified Temperature Range
Trường học	International Electrotechnical Commission
Chuyên ngành	Process Management for Avionics
Thể loại	Technical report
Năm xuất bản	2005
Thành phố	Geneva

Định dạng
Số trang	58
Dung lượng	0,91 MB