Congrex whole of life assurance

Document Sample
Congrex whole of life assurance Powered By Docstoc
					         TRISMAC 2008
ESA-ESTEC Noordwijk, NL 14-16 April 2008

       The Evolution of Product
    Assurance Approaches for EEE
         Space Components

     Ralf de Marino, Components
         Division ESA ESTEC
The basic PA Objectives :
… ensure that EEE components used in space projects will allow the project as a
whole to meet its requirements in terms of functionality, quality, reliability, schedule
and cost ….
Translates into a complex conglomerate of engineering and management
Below a level of basic principles formulated in the top level requirements for Space
Product Assurance – Policy and Principles (e.g. ECSS-Q-00A) and Quality
Assurance (e.g. ECSS-Q-20B) PA is not uniform but expands soon into an
extended set of rules and ‘recipes’ depending on technologies, component types
and other conditions.
Optimisation through tailoring of requirements has been and continues to be
one of the fundamental principles of PA.
While we tried to achieve a balance of all the objectives in the past it appears that
the trend has turned to a prioritisation of economic factors : cost and schedule.
But we must not go too far in sacrificing technical requirements.

                  TRISMAC 2008, PA Approaches for EEE Space Components
Another top level requirement – risk assessment and management - should
provide the balance.
Objective risk assessment is a difficult matter because we have to quantify risks.
Quantification means the availability of relevant data.
For EEE Components this means extensive testing and large amounts of data,
which are driving efforts and total cost.
Standardisation is one of the obvious solutions to achieve a degree of economics
which is applied for a long time.
It provides an infrastructure in the form of requirements documents detailing
evaluation, qualification and procurement criteria and conditions down to the level
of Test Methods, Generic and Detail Specifications (e.g. US MIL, ESCC, JAXA).
These requirements are deemed to be representative and sufficient to ensure the
suitability of components for space applications.
The ultimate goal is to produce catalogs of readily available (=procurable)
components that users can choose from without further assessment.

                 TRISMAC 2008, PA Approaches for EEE Space Components
Qualification Concepts :

Qualified Parts List (QPL)
         Individual part type qualification and qualification by similarity (family)
Capability Approval (ESCC – CA)*
         at present mainly for Hybrids and Microcircuits (other products possible)
Qualified Manufacturer List (MIL-QML, EQML)
         product qualification based on technology qualification

These concepts coexist although the promise of higher efficiency pushes a
preference for the QML concept.

However, the extended level of responsibility placed on the QML manufacturers
and the reliance on representative test structures and test regimes (periodic testing
in place of lot based acceptance) also has its price and risks.

* European Space Component Coordination

                 TRISMAC 2008, PA Approaches for EEE Space Components
How do we populate those Qualified Parts Lists ?

In the general perception Space Qualified Products still enjoy the reputation of
being among the best there is : highly reliable = robust, stable, long lasting

For some manufacturers Space Qualification is therefore a marketable recognition.

‘Bread and Butter’ type components :
          Passives : e.g. resistors, capacitors, connectors, cables, wires, etc.
          Actives : e.g. diodes, transistors, standard logic, linears
Many of those have a secure market, been qualified a long time ago and
qualification maintenance (cost) is covered by continuous production.

For new parts to be added :
A market needs to be established based on an objective usage forecast.

In Europe we try to establish an up-front user consensus for new developments
and the candidate component types and manufacturers in the frame of the ESCC
Component Technology Board (CTB) which defines an Annual Qualification
Programme (AQP).

                 TRISMAC 2008, PA Approaches for EEE Space Components
How do we populate those Qualified Parts Lists ? (2)

It was and is mandatory for the manufacturer to share the investment in the
qualification effort but few are willing to bear the full cost and require a financial

Until the early 90’s ESA had a moderate Annual Qualification Budget to contribute
to the evaluation and qualification cost.

When possible we then resorted to the inclusion of qualification efforts in the frame
of dedicated space product development activities (limitations apply).

In 2004 ESA initiated the European Component Initiative (ECI phase 1) for the
development and qualification of critical space components.

ECI will continue into phase 2 in 2008.

The ESCC-QPL is growing again and so is the ESCC Preferred Parts List.

                  TRISMAC 2008, PA Approaches for EEE Space Components
The function of the ESCC Preferred Parts List

Originally a component policy tool for expressing preference among different
suppliers for the same part type created at a time when second sourcing was a
wider option.

It remains the top-level space component selection reference in the ECSS system.

In the mid 90’s we introduced a second part listing components which had been
evaluated to ESCC requirements, deemed to have the potential for a full ESCC
qualification (if someone was prepared to invest) and

the flight heritage criterium was added for the listing in EPPL Part 1 (with the
anticipated effect of reducing the incentive of full qualification).

The intention of EPPL part 2 is to achieve a better exploitation of evaluation
activities performed in the frame of ESA (or other) projects by giving a higher level
of recognition to the achievement of a complete evaluation programme on the
bases of an independent data assessment.

                 TRISMAC 2008, PA Approaches for EEE Space Components
Further Extending the Scope – ECSS-Q-60B

All of the above is unfortunately not able to meet the rather diverse range of project
demands for Space components.

The often unique and ambitious performance characteristics of scientific payloads
require rather sophisticated components based on very advanced technologies.

This applies similarly also for commercial telecom payloads which need
distinguishing performance parameters and features to be competitive.

ECSS-Q-60B Space Product Assurance – EEE Parts
Defines the European requirements for the selection, control, procurement and
usage of EEE components

It distinguishes between 3 component classes (1 = highest) which represent
different quality and risk levels with detailed annexes defining the applicable
standards per component family. Class 3 includes the use of commercial

                 TRISMAC 2008, PA Approaches for EEE Space Components
                                 CLASS 1                          CLASS 2                              CLASS 3

- Compliance to ECSS-M-00        required                         not required                         not required
- EEE parts control plan         required                         compliance matrix                    compliance matrix
- PCB                            required                         required                             not required
- “as built” DCLs                required                         required                             not required

- Type red. & pref. process      required                         not required                         not required
- Commercial parts               -                                allowed (some families)              allowed (some families)
-- Mfr assessment (evaluation)   required                         not required                         not required
- Approval process               DCL (qualified)                  DCL (qualified & EPPL/NSPL)          DCL (qualif & not qualif)
                                 PAD (not qualified)              PAD (others)

- Procurement spec               normative or project             normative -> datasheet               normative -> datasheet (for review)
- Quality levels
  + integrated circuits          ESCC or QML/V                    ESCC or QML/Q-M + PIND               ESCC or 883B screening
  + discrete active              ESCC or JANS                     ESCC or JANTXV + PIND                ESCC or JANTXV
  + standard passive             ESCC/C, EFR-R                    ESCC/C, EFR-R, CECC qual + BI        ESCC/C, EFR-R, CECC qual + BI
  + relays                       ESCC/B                           ESCC/B or MIL/R + ESCC screen        ESCC/B, MIL/R + ESCC screen
  + hybrids                      ECSS-Q60-05 level 1 or QML/K     ECSS-Q60-05 level 2 or QML/K         ECSS-Q60-05 level 2 or QML/H + PIND

- Customer precap                required (non qual & few qual)   required (some non qual types)       not required
- Lot acceptance test            required (data < 2 years)        required (content to be discussed)    required (content to be discussed)
- Customer buy-off               required (non qualified)         required (some non qual types)       not required
- DPA                            required (non qualified)         required (some non qual types)       required (non qual. relays)

- Alerts                         required                         only handle alerts received          only handle alerts received
- Lot homogeneity                required                         not required (except for rad)        not required (except for rad)
Dealing with non-space parts – Up-Screening

Another method that has been suggested to rapidly obtain project approval for
existing industrial components is the so-called up-screening.

This is based on the assumption that the component design and process
technology margins of existing industrial products are sufficiently large to
demonstrate the required project specific level of reliability and quality.

Inherently this is a rather risky approach as the overall reliability and quality
margins are unknown and because it usually happens without the support and
involvement of the manufacturer.

This particularly creates a confidence problem with respect to the test coverage
that can be obtained without detailed design knowledge (black box).

This is definitely not a recommended approach and particularly space qualified
manufacturers take exception to it for their similar commercial products.

                 TRISMAC 2008, PA Approaches for EEE Space Components
Dealing with non-space parts – example DRAMS

There are of course some note worthy exceptions such as DRAM (Dynamic
Random Access Memories) which are the primary ingredient of on-board mass
memory systems for the past 10+ years.

Only available as commercial products for primarily computing applications in a
fiercely competitive market and based on the most advanced processes the
average life cycle is less than two years and includes die shrinks and process
optimisation along the way.

Such life cycles do not lend themselves to formal qualification and the actual
product selection/approval for a space application must be performed on the
prospective flight lot to ensure the validity of in particular radiation test data.

These type of very high volume components are supported by high level quality
requirements and
The mass memory application allows the use of effective reliability (redundancy)
and radiation effects mitigation.

                  TRISMAC 2008, PA Approaches for EEE Space Components
Dealing with non-space parts – Assembly and Test House

The number of fabless device manufacturers are increasing and among the ranks
of medium to large size semiconductor industry there is a trend toward the use of
pure-play foundries.

To enable design houses using smaller or bigger foundries and to exploit radiation
hardened cell libraries developed for pure play foundries, etc.
an Assembly and Test House Qualification/Certification concept is being developed
under the responsibility of the ESCC-PSWG (Policy and Standards Working

This is concept applicable for discrete semiconductors and microcircuits will be
embedded in the existing ESCC requirement framework and places a clear end
product responsibility on the design house or the ATH.

It is based on the assumption that the necessary insight into the front end process
and access to wafer lot acceptance test data from the foundry can be achieved
and therefore requires a minimum level of foundry cooperation.

                 TRISMAC 2008, PA Approaches for EEE Space Components
Dealing with the speed of semiconductor technology innovation is
no easy matter.

Radiation Hardness Assurance is getting increasingly complex when testing for
Single Event Effects in high speed, high density integrated circuits.
The demands on test set-up and software are high and it is increasingly difficult to
determine if events observed are caused by radiation or test induced.
Problems that are expected to increase with deep submicron CMOS technologies.

These may be significantly life time limited (<< 10years) due to Time Dependent
Breakdown failure modes unless mitigation approaches can work on the design or
application level.

As performance limits are being pushed out it is no longer clear if prevailing
assumptions on worst case conditions are still applicable as design and process
generations have evolved.

Advanced packaging concepts for high complexity microcircuits with pin counts
above 350 I/Os rely on Column Grid Array packages that introduce a new
connection between component and assembly process qualification.

                 TRISMAC 2008, PA Approaches for EEE Space Components
Other Important Circumstances

RoHS – the legislation for the Reduction of Hazardous Substances has banned the
use of lead (Pb) in electronics (Space and several others are exempt)

The industry is moving toward compliance using pure tin finishes and Pb-free

Pure tin is forbidden in space applications due to its propensity to grow whiskers
posing a short circuit threat.

Reliable test methods to detect non-conforming material and rework procedures
are to be put in place.

Counterfit components represent a definite risk for space projects.
Their avoidance requires a strict procurement discipline.

                 TRISMAC 2008, PA Approaches for EEE Space Components
Human Resources

Product Assurance for EEE components is an engineering job and by no means
automated or dispositioned administratively.

Industrial engineering and procurement teams in established space suppliers have
been downsized.

Newly entering space system providers have to build experience.

Component PA groups in Space Agencies (e.g. ESA and CNES) carry high and still
increasing workloads.

Suitable young PA engineers are difficult to find.

And we all need training

                 TRISMAC 2008, PA Approaches for EEE Space Components

Shared By: