How useful is conditional probability in electrical engineering field ?

[EEstyleZ90]

Specially in digital/analogue electronics troubleshooting ?

 

[rjenkinsgb]

It applies in such as when you discover a failure due to overload in one component, there is a high probability other parts of the same circuit or subsystem have also failed or been overloaded and are suspect.

Also, the chances of any particular part failing are often inversely proportional to ease of replacing it and the availability of a replacement - that's just Murphy's Law..

 

[EEstyleZ90]

that's just Murphy's Law..

Non sense

 

[EEstyleZ90]

Next opinion please...

 

[Pommie]

Next opinion please...

You don't seem to understand a joke when you see one so I doubt you'll have many takers.

Mike.

 

[rjenkinsgb]

And zero real-world experience.

 

[Tony Stewart]

Your question has no correlation, unless you can define the FITs or MTBF of every component with its, application, stress factors, supplier and environmental conditions. A design validation test (DVT) might highlight a failure mode from stress tests, which later biases fault isolation methods. Perhaps statistics might be used create parameters from a large database of defective field returns and root causes that results in design or process improvements. Perhaps it's a fuzzy guess. https://tinyurl.com/2o9yuof7 Imagine the analysis of warranty field returns of 1 million rotating hard disk drives.

 

[EEstyleZ90]

You don't seem to understand a joke when you see one so I doubt you'll have many takers.

Mike.

And zero real-world experience.

Sorry if I am rude or something I am a bit psychotic sometimes.

But you are right . I am just a beginner in this EE field but I have a solid ground in electronics though (I hope so).

My intuition helps sometimes in analog . But digital designs are very hard to troubleshoot without some assistance from probability theory . There is so many combinatory logic and it gets hard when they are tied to FPGAs.

That's why I need to explore new approaches .

 

[Tony Stewart]

If you need a new approach, use diode law and diode capacitance characteristics using a sig gen current sensing on inactive CMOS to look for ESD diode-pf loops that indicate silicon port stress aka Huntron Tracker approach. It works. If you have a solid ground in electronics it will be immediately obvious the relationship between diode bulk resistance and junction capacitance.

 

[rjenkinsgb]

Faultfinding, whether a design that's not working as expected or a failure in previously working equipment, is along the lines of a "Differential diagnosis", to steal a medical term.

What parts of the circuit/system/machine could be related to the fault?
What parts could not be related?

Of the parts that could be related, which could be causative and which are side effects?

Pin it down as far as possible, then compare what it it doing to the components/parts involved, to try and figure out what type of fault or failure could cause the effects you are seeing.

Use or add test points; with an FPGA, you can use spare outputs or temporarily swap outputs from an area that is unaffected, so you can see what is actually happening on a scope or logic analyser.

If you are working with other people, talk it over with them and explain the system and fault in as much detail as you can. If on your own, still explain it out loud as if you were trying to explain to someone else, down to the finest detail.
It forces your brain to re-process what you know and can often lead to new answers to a problem.

Rubber duck debugging - Wikipedia

en.wikipedia.org

And don't rely purely on simulations - they are useful but not always perfect. As another user on here has in their signature:
“In theory, theory and practice are the same. In practice, they are not.”