Timeline for UNI-T UT181A question: what is confidence interval used in the datasheet?
Current License: CC BY-SA 4.0
31 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 14 at 23:38 | history | edited | Валерий Заподовников | CC BY-SA 4.0 |
added 81 characters in body
|
| Apr 14 at 23:05 | history | edited | Валерий Заподовников | CC BY-SA 4.0 |
added 4 characters in body
|
| Apr 14 at 22:30 | answer | added | Brethlosze | timeline score: 0 | |
| Apr 14 at 22:24 | answer | added | Prashanth C | timeline score: 1 | |
| Apr 14 at 22:23 | vote | accept | Валерий Заподовников | ||
| Apr 14 at 22:13 | answer | added | Marcus Müller | timeline score: 1 | |
| Apr 14 at 21:56 | review | Close votes | |||
| Apr 15 at 21:26 | |||||
| Apr 14 at 21:56 | comment | added | Marcus Müller | @ВалерийЗаподовников thanks for the edit ! That really helps a bit. I'll try to answer. | |
| Apr 14 at 21:55 | comment | added | Marcus Müller | @Brethlosze please, then, go ahead and answer the question. I've really really really tried to help and explain, work with OP to improve their question. The result was that OP added "what does it mean for military equipment to have more than one sigma?", which frankly, makes no sense on multiple levels. I hope you can appreciate my frustration here, being accused of being unhelpful after spending half an hour trying to help. | |
| Apr 14 at 21:55 | comment | added | Валерий Заподовников | @Brethlosze I edited the question. It should be "nicer" to answer now. | |
| Apr 14 at 21:54 | history | edited | Валерий Заподовников | CC BY-SA 4.0 |
deleted 376 characters in body; edited title
|
| Apr 14 at 21:53 | comment | added | Brethlosze | @MarcusMüller Please, stop acting this way. You are not helping the OP, and you are being absolutely biased. This is not new from you. | |
| Apr 14 at 21:47 | comment | added | Marcus Müller | I think you meant to really fundamentally modify your question – make it about "why do 68% of measurements fall into \$1\sigma\$ of the mean value; but instead, you just added more confusing text to your question that was already based on wrong assumptions. You'll have to remove all the things that aren't relevant to the question you want to ask from the text, please. | |
| Apr 14 at 21:45 | comment | added | Marcus Müller | no, this is not about your multimeter. This is just about the fact that the assumptions about distribution do not hold. And they don't – you're the one making the surprising claim that they do, and the absolutely wrong claim that 68% is the right measure for every case. I don't have to prove anything here, your question is really just based on wrong assumptions, and your newest edit really just added total nonsense. | |
| Apr 14 at 21:43 | comment | added | Валерий Заподовников | @MarcusMüller Prove that for this multimeter before voting close | |
| Apr 14 at 21:41 | history | edited | Валерий Заподовников | CC BY-SA 4.0 |
added 176 characters in body
|
| Apr 14 at 21:41 | comment | added | Marcus Müller | I’m voting to close this question because it's based on the wrong assumption that the measurement of frequency had normally distributed error, and that the accuracy given was the standard deviation of that. Neither is the case, as discussed in the comments. | |
| Apr 14 at 21:39 | comment | added | Marcus Müller | No. That makes no sense. Again, this is really basics of normal variables, and I can't explain that better than a textbook on measurement theory or the wikipedia article on normally distributed variables could. don't speculate – read! | |
| Apr 14 at 21:23 | comment | added | Валерий Заподовников | @MarcusMüller Is it possible it actually does some form of continuous measurement and the datasheet specifies it for 1 measurement? Of course if you have 68%, you can measure 50 times and see what is most accurate (more than 50% value) | |
| Apr 14 at 21:20 | comment | added | Marcus Müller | "military equipment will have more than one sigma accuracy": Uh, I think you're really confused here. I've mentioned that several time now, so maybe you'll really need to read up on normally distributed random variables, and what sigma means in context of these. | |
| Apr 14 at 21:14 | history | edited | Валерий Заподовников | CC BY-SA 4.0 |
added 137 characters in body
|
| Apr 14 at 21:12 | comment | added | Валерий Заподовников | @MarcusMüller Okay, but Military equipment will have more than one sigma accuracy. So why is that not the case with my multimeter? Intuitively, I would say it is accurate enough or at least 95%. | |
| Apr 14 at 21:09 | comment | added | Marcus Müller | You already mentioned the key word here, \$\sigma\$! Because 68,something% of realizations of a normal random variable fall within one standard deviation (\$\sigma\$) of its mean. | |
| Apr 14 at 21:08 | comment | added | Валерий Заподовников | @MarcusMüller Okay, but with Voltage I also am very sceptical it is 68.27%. | |
| Apr 14 at 21:07 | comment | added | Marcus Müller | you're misinterpreting your search results. Voltage fluctuation under Johnson-Nyquist noise is a normally distributed random variable. Frequency estimation isn't; nothing in the datasheet claims that, that's just you! | |
| Apr 14 at 21:05 | comment | added | Валерий Заподовников | @MarcusMüller Yes, they do. google.com/… Well, the internals of modern multimeters is complex. Indeed, it would make sense that typical statistics doesn't work for them, even if the datasheets say that it does. | |
| Apr 14 at 21:01 | comment | added | Marcus Müller | No, they don't. That makes sense for amplitudes of normal variables, and I don't see an indication for that here. | |
| Apr 14 at 20:59 | comment | added | Валерий Заподовников | @MarcusMüller Everyone assumes Confidence interwall of one sigma (that is 68.27%) unless specified otherwise. | |
| Apr 14 at 20:58 | comment | added | Marcus Müller | "everyone assumes": Please give a reference for that. I'm not assuming that. I'm also not assuming 100% of time or 99.73% of time, you'll really need to explain where these numbers came from, Valerij! | |
| S Apr 14 at 20:53 | review | First questions | |||
| Apr 14 at 22:28 | |||||
| S Apr 14 at 20:53 | history | asked | Валерий Заподовников | CC BY-SA 4.0 |