Role within my team: We are a small R&D team with many hats. My role is at any time one or more of the following.
designer for electronics and embedded control, software and hardware tester, develop the functional specs and user interface for a new product, investigate product improvements, verify that the product meets all specs and does not introduce unintended features, technical writing for product manuals, user guides, installation and troubleshooting instructions.
We are also the product support experts for the service department and customer trouble calls.
Management, Project Engineering, Architecture and Specification, System Integration, Transition to Production ...
Would like to see a slide or two in presentation one on approach going from 1) Specification to 2) Test Plan to 3) Modeling approach. Then two or three slides at the end of the session (class 5) amplifying and summarizing the entire approach - not really outside the scope of the presentation, but just to tie the entire process flow together ...
It was about 9 yrs. ago so they were not as sophisticated but it performed it the calculations as prescribed so results were valid for the ASIC but the resulting values did not produce the desired effect, so the bit shifter was born.
VLSI chip verification is the most stringent form of verification because of the cost of getting it wrong. Some poeple do limited amounts of PCB verifciation and system verification is required by many mil/aero project. What kind of design are you working on Mohammad_Zohaib?
Controllability/observeability - partly dealt with by doing a lot of block level verification where the problem is not so bad. Have to restrict the system level tests carefully because of the time they take. Also some people use emulationm or prototyping to speed up the process. We will talk about those tomorrow.
Question: In the need for verification you talked about controllability and observability due to complex design challenges. How we engineers trade off when trying to check the various blocks for functinality vesus verification time?
I will mention formal in several places including in the SystemVerilog session and tomorrow in models and their execution. Will not however have time to go in depth. That could be a course all on its own!
Brian,The last test I worked we used what we called requirements verification matrix based on design and worked through to develop tests from this. I remember one time the design guys had a case where they dropped what they thought were the least significant bits but where actually the MOST significant. It was a little endian big endian thing. It verified Ok in simulator of course but not in operation.
When we very a blkock in the conext of a system, we are in effect making sure that the sepcification was understood and the specification is consisent. In other words we are beginning to do some validation up front rather than at the end where it is done today.
Brian: Yes I have an idea about the need for system verification. We were maintaining Motorola MSF-5000 UHF base stations, and occassionally we would get a failure with an obscure error code. I called Motorola engineering and after sending them the exact werror code, asked why they didn't publish an error code library. He said even the engineers didn't know what the error cosdes meant, or what exactly caused each one; they had to run down the hal to the softheads and ask them. The codes were altered with every production version. Talk about no one steering the ship.
@farhad00110 - At 1:00 PM (assuming you're on Cental time like me), a sound bar should appear at the top of the page. If it doesn't appear, refresh your browser page. It doesn't start automatically. You'll have to click the play button. There's no video. You'll need to download the slide deck above and follow along. If you don't have Power Point, you should be able to download a Power Point Viewer for free.
I saved the presentation as a PDF. I have uploaded it to a file hosting site called 2shared.com. Some users on this forum have had problems with this site in the past. If you don't have AdBlock running on your browser, there might be a few advertisement URLs they want you to click on instead. Just look for the big blue download button for the actual file.
Are they robots or androids? We're not exactly sure. Each talking, gesturing Geminoid looks exactly like a real individual, starting with their creator, professor Hiroshi Ishiguro of Osaka University in Japan.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.