PRISM Forum - Message Replies


Topic: PRISM Questions and Answers
Topic Posted by: SRC (src_forum@alionscience.com )
Organization: SRC
Date Posted: Wed Jan 12 8:33:33 US/Eastern 2000
Topic Description: Welcome to the PRISM forum! Please feel free to post your questions and comments about the PRISM assessment software here.

Back to message list About this forum

Original Message:

Posted by: Ralph Listl (ralph.listl.t-systems@eads.com )
Organization:EADS
Date posted: Mon Feb 27 5:27:44 US/Eastern 2006
Subject: Prism Reliability Model
Message:
The Prism Reliability Model uses an initial failure rate multiplied by a process grade factor. The calculation of the inital failure rate for all via RACrates model supported items (Capacitors, Resistors...) makes use of a temperature dependent factor in their models. Therefor all base failure rates are temperature dependent. The process grade factor does also include an environment factor that is temperature dependent. If a complete system has been set up and the temperature is changed it influences the base failure rates as well as the process grade factor. From my point of view a temperature dependency should either be applied on the base failure rate or for the process grade factor but not on both, because this is double influence. A temperatue dependence on item level is understandable, but a temperature dependence on a process grade is for my understanding not correct. Could you please explain why the Prism model is set up like this?


Reply:

Subject: Temperature dependency
Reply Posted by: David Dylis (ddylis@alionscience.com )
Organization: System Reliability Center (SRC)
Date Posted: Mon Feb 27 15:19:02 US/Eastern 2006
Message:
As you identified, a component has a certain inherent failure rate that is affected by application and temperature. Processes that are used in the design, development and manufacturing of a system are also affected by temperature. Poor processes will result in more early life failures than good processes. Temperature is one factor that will precipitate these events. When PRISM was developed data was analyzed and the affect of temperature on failure rate was distributed between the component and process grade portions of the model. If you have additional questions, feel free to contact me.


Reply:

Subject: Temperature Dependency
Reply Posted by: Ralph Listl (ralph.listl.t-systems@eads.com )
Organization: EADS
Date Posted: Tue Feb 28 7:40:15 US/Eastern 2006
Message:
Thank you very much for your prompt response. I agree to your statement of having the temperature as a parameter that has influence on the inherent failure rate of an item as well as to the process of manufacturing of the item. Temperature has from my point of view no influence on the desgin and development process. The modell that is used by PRISM makes it possible to change the temperature in the environment area. if you do this and change for example the OP. temperature from 85°C to 70 °C, trickle down through the system, you receive a different failure rate for the system and a different proces grade factor. This is as far as I understand reliability predictions not correct. The OP. temperature is the temperature used during operation of the item and is used for calculation of the inherent failure rate. Within PRISM it is also used for the environmental factor in the process grade factor. This must not be mixed. For the process grade factor other temperature settings must be used. How is the temperature really used within PRISM during calculation of the environmental factor ? I hope this explanation makes it more clear, where I see the difficulties in the PRISM model.


Reply:

Subject: Temperatue Dependancy
Reply Posted by: David Dylis (ddylis@alionscience.com )
Date Posted: Tue Feb 28 9:52:18 US/Eastern 2006
Message:
PRISM differs from other prediction methodologies as it takes into account processes that affect the reliability of a system. Prior to PRISM these factors were typically ignored and only the failure rates of components were considered.

The achievement of system reliability relies on the following elements:
• Obtaining valid system requirements and managing interface dependencies
• Good quality parts must be chosen
• The parts must be designed into a system in a robust manner to ensure that they can meet performance requirements over the system’s design life
• The design must be validated through analysis and test
• The system must be manufactured without inducing damage or adding defects
• A management philosophy that supports the achievement of the above five elements

The PRISM design PI factor considers failures resulting from an inadequate design. Examples are tolerance stack-up, unanticipated logic conditions, a non-robust design for given environmental stresses, etc. Below is a sample of some questions from the PRISM design process grade that address temperature related issues:

Is design of experiments (DOE) used to ensure robustness of the FRU in the product under all operational and environmental variations?
Will mechanical stress analysis be performed on relevant components, materials and structures?
Will the chassis that this FRU is mounted in be thermally measured and analyzed and operating temperatures assured to be at a safe margin below device limits?

In addition, PRISM does not have a PI-Q factor at the component level as in MIL-HDBK-217 and other methodologies. Part quality is assessed at the system level …

I will send a copy of a report that describes the basic process on how the PRISM system level models were derived and the process that addresses the PI-E factor in PRISM.

If you have additional questions, feel free to contact me.


Back to message list About this forum

Reply to this message