Background on Online Self-Tests
When looking on all potential failures of a safety system caused by a random hardware fault, we identify general types of failures:
- λ_s = Failures which don’t influence the safety function (safe failure)
- λ_d = Failures which influence the safety function (dangerous failure)
- λ_tot = Total number of possible failures = λ_s + λ_d
For each hardware component, which is part of the safety function, the rate for failures is given with a unit, called failure in time (FIT) and is equal to 10e-9/h.
The FIT rates are available from the hardware vendors or from standardized databases (e.g. SN29500). Furthermore, chip vendors addressing the safety market provide the FIT rates for their internal microcontroller blocks like RAM, FLASH, Clocks and CPU. Due to the fact, that it is impossible to decide in general which failure is dangerous and which failure is safe, we split the FIT rates to 50% dangerous and 50% safe.
For a system with a given safety integrity level, the rate of dangerous failures is limited. This limited rate is given as safe failure fraction (SFF) and depends on system architecture and complexity of the used hardware components. This SFF is the relation of safe failures to all possible failures. The relevant range for a system starts from >60% and rises up to >99%.
To achieve this limit for a system using a complex hardware unit like a microcontroller, we need to enhance the system with online diagnostic (also called "online self-test"). Depending on the algorithm and measure of the diagnostic module, we achieve a defined diagnostic coverage (DC). The definition is part of the safety standard and classified in three classes to simplify the use:
- low : diagnostic coverage is 60%
- medium : diagnostic coverage is 90%
- high : diagnostic coverage is 99%
With this diagnostic coverage, we are able to guarantee that we handle all detected dangerous failures in a system specific way. So that no influence on the safety function is guaranteed. Finally we can calculate with:
- λ_dd = Detected dangerous failures = DC * λ_d
- λ_du = Undetected dangerous failures
the safe failure fraction to:
- SFF = (λ_s + λ_dd) / λ_tot
One Functional-Safety-Engineer in every team
We continuously educate engineers in software development for Functional Safety. After at least 5 years of activity in this functional safety, the training leads to an examination by an external company. The recognition of the knowledge takes place with the certificate "Functional Safety Engineer".
We are able to contribute a lot of experience, knowledge and technologies from more than 270 projects and research projects to your project.
Mentoring and Training
We support your project from concept inspection to certification by an independent approval body. At all stages we can contribute with experience, tips and tricks to ensure the success of the project.
You get source code of the self-test algorithms, including all artifacts you will need for approval by the certification authorities. We are happy to integrate the documentation artefacts into your existing process model.
Guarantee and Trust
We are committed from the very beginning to your project and provide you with full support for the realization of your vision. Driven by the enthusiasm of the team work and the fun on the joint success of your project.
Individual selection of algorithms and measures for your project guarantee optimal runtime behavior. Prepared are Online RAM-Tests, ROM/FLASH-Tests, Clock-Tests, CPU-Tests, Program Execution Checks and more.
You are Interested in Functional Safety ...
Arrange an expert meeting
We offer a free and non-committal interview with one of our experts. We can get to know each other, answer your open questions and also discuss the first requirements of your project.
Related Products and Services
The so-called Flexible Safety RTOS is the real-time kernel with memory protection, pre-certified for applications in the field of functional safety according to IEC 61508, ISO 26262 and IEC 62304
For a first impression of the nature and extent of completed projects, we have put together a small selection for you. For more details read: Reference Projects.