Great question. I cover this in my other video on accelerated life testing, which you can find on my channel. Thanks!

I deeply appreciate your constructive feedback regarding my prior commentary in the DFMEA video. I wholeheartedly acknowledge your perspective on the matter of risk prioritization through the utilization of Risk Priority Numbers (RPNs). Allow me to address the salient points you have raised: RPN Calculation Validity: It is astutely observed that conventional Risk Priority Number (RPN) calculations tend to yield significantly elevated RPN values when both the severity (Sev) and occurrence (Occur) parameters are assigned the maximum value of 10. This, in turn, often renders the assignment of low detection values (Det=1 or Det=2) improbable, given the inherent inverse relationship between these variables. I must underscore that the extreme example presented was employed deliberately to underscore the limitations associated with an exclusive reliance on RPN as a means of risk assessment. While such occurrences are rare, it is plausible for a "1" to be allocated within a DFMEA, bearing in mind that the determination of detection may occasionally be influenced by the expertise level of the engineer engaged in this manual assessment. The Role of Detection: Your emphasis on the pivotal role of detection within the RPN calculation is well-founded. The detection factor assumes a critical function within the purview of risk assessment, serving to pinpoint areas necessitating heightened attention and control measures to avert the actualization of the identified risks. Addressing All Severities: Your proposition that all instances of severity attaining a rating of 10 warrant immediate consideration is a valid assertion. In numerous risk assessment frameworks, the sheer gravity of a high-severity scenario in isolation is sufficient to warrant prioritization, regardless of other influencing factors. It is imperative to recognize that not all situations neatly align with the confines of the RPN calculation model, and instances exist wherein the mere existence of a high-severity issue mandates expeditious action. Context Matters: It is paramount to acknowledge that risk assessment is a bespoke undertaking, devoid of universal applicability. The suitability of employing RPNs is intrinsically contingent upon the specific context, industry, and organizational nuances. While RPNs prove advantageous in many scenarios, some contexts may indeed incline towards an emphasis on severity, as you have aptly surmised. In summary, I concur with your appraisal of the example provided and your accentuation of the primacy of addressing high-severity concerns. It is imperative to comprehend that risk assessment is a multifaceted discipline, wherein RPNs, while undoubtedly valuable, may not emerge as the panacea for all situations. The choice of approach should be judiciously determined by the contextual factors, the industry domain, and the unique intricacies of the risk under consideration. I extend my sincere appreciation for your discerning insights and the elucidation provided on this matter. Thanks for watching!

It depends on whether or not the intent is an operational or non operational test for your test case.

@@tomreshtesting tell me in the perspective of operational condition

There is IEC spec for tumble/random drop testing. The total number of tumbles that you will perform will be specific to what your company has set for design reliability targets, or it could be you want to tumble your product to failure. The intent and number of cycles will be product, use case, and industry specific, and there isn't a one size fits all # of cycles.

Great question. Thermal Shock is Great for any electronics regardless of environments. One example is even in a temperature stable environment, an electronic device will vary in temperature depending on the load or task it's performing and the corresponding heat that is generated from the work performed. Thermal Shock can be leveraged to accelerate years of expansion and contraction events in a short period of time to find failure modes/ weakest link in the design or manufacturing process.

Not too familiar with tremors in Alzheimer's patients, but I do know there are white papers on using wireless accelerometers with Parkinson's patients. Only concern I would have is how to reliably create a "connection " of the accelerometer to the patient with minimal decoupling. Most likely a tightly fitted ban with an attached wireless accelerometer would be the best route to go.