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 Show all replies About this forum

Original Message:

Posted by: Chris Wright (christopher.d.wright@baesystems.com )
Organization:BAE SYSTEMS
Date posted: Wed Mar 6 15:32:13 US/Eastern 2002
Subject: RACRates IC Model Questions
Message:
I have been evaluating the PRISM IC RACRates model for my company and have a few questions. First, how is it that the pi-factors tend to go down for non-hermetic devices versus hermetic? This seems counter intuitive, especially for pi-RHT. The overall device failure rate ends up being lower for a non-hermetic part. Is this true? Second, it appears that pi-TO, pi-RHT, pi-DT and Lambda-SJ all vary with device temperature, but there is nothing in the manual which describes these relationships. I am also puzzled as to why pi-RHT would change when I add a junction temperature rise, since this is supposed to be a non-operating factor. The junction temperature rise should only apply to operating factors. When non-operating the device should be at its non-operating ambient temperature with no junction temperature rise since it wouldn't be dissipating power at that point.


Reply:

Subject: RACRates IC Model Questions
Reply Posted by: David Dylis (ddylis@alionscience.com )
Organization: Reliability Analysis Center (RAC)
Date Posted: Thu Mar 7 13:36:47 US/Eastern 2002
Message:
1. First, how is it that the pi-factors tend to go down for non-hermetic devices versus hermetic? This seems counter intuitive, especially for pi-RHT. The overall device failure rate ends up being lower for a non-hermetic part. Is this true?

In some cases this is true, however, this will not be true for all factors in all environment selections. For instance, the Operating Duty Cycle (PI-DCO), the operating temperature acceleration factor (PI-TO) and Temperature acceleration rate factor (PI-DT) are higher for non-hermetic devices in a ground stationary environment.

The failure rate of a non-hermetic device can be lower than the failure rate of a hermetic device. The RAC has been collecting data on parts and systems for over 30 years and has an extensive database with over 10^12 hours of field experience data. This data was used in the development of the PRISM models. Based on the data, plastic encapsulated microcircuits (PEMS) can exhibit failure rate comparable or less than hermetic counterparts. The OEM using PEMS must have a goods parts control program, assure that components are used within specification and in an operating environment. We believe that the reliability growth of PEMS can be attributed to the fact that they are produced using more robust processes than that of their hermetic counterparts. In addition, die/devices and plastics being used in the development of PEMS today exhibit less contaminants than devices manufactured in the past reducing failures that result from contamination/moisture.

When using PRISM to calculate failure rates of any system it is recommended that PRISM users define the processes that are being used in the design, development, manufacturing, maintenance, etc. of their systems. These processes define what steps are (or are not) being taken to assure reliability in a design and some questions specifically address the use of PEMS.

2. Second, it appears that pi-TO, pi-RHT, pi-DT and Lambda-SJ all vary with device temperature, but there is nothing in the manual which describes these relationships.

Each of these factors is based on a form of Pecks Model. A more complete description of the basics behind the IC model used in PRISM has been provided in the RAC publication “Reliable Application of Plastic Encapsulated Microcircuits” (PEM2). The model was updated to reflect more recent data and to account for hermetic integrated circuits.

3. I am also puzzled as to why pi-RHT would change when I add a junction temperature rise, since this is supposed to be a non-operating factor. The junction temperature rise should only apply to operating factors. When non-operating the device should be at its non-operating ambient temperature with no junction temperature rise since it wouldn't be dissipating power at that point.

The PRISM Help section and User Manual contains a description on how to perform a dormant reliability prediction. What you are saying is true, however, you should not be doing this when performing a dormant reliability prediction as the procedure identifies that you must reduce the part (internal) temperature rise to zero for each and every diode, IC, resistor, thyristor and transistor in the system. There are also several additional steps required to perform a dormant reliability prediction. I recommend that you review this section of the manual to assure that you are performing your calculations correctly.

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


Reply to this message