EEVblog
ฝัง
- เผยแพร่เมื่อ 22 พ.ค. 2024
- What's all this AC RMS and Standard Deviation measurement stuff on your oscilloscope anyhow?
And how does it differ from "normal" RMS measurement?
Another trap for young players, how to make sure you measure noise correctly.
Dave made an oopsie in the 1GHz Siglent scope review, did you spot it?
Forum: www.eevblog.com/forum/blog/ee...
#Oscilloscope #Tutorial #RMS
Bitcoin Donations: 38y7DE8HEHNj8fGDtUr4PkCn9nWxiorvvy
Litecoin: ML7oQokTwB38bgzzjLDbRV97HKAHuwRfHA
Ethereum: 0x11AceA38DCA9DbFfB4F35f3F746af65F9dED28ce
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
AliExpress Affiliate: s.click.aliexpress.com/e/c2LRpe8g
Buy anything through that link and Dave gets a commission at no cost to you.
Stuff I recommend:
kit.com/EEVblog/
Donate With Bitcoin & Other Crypto Currencies!
www.eevblog.com/crypto-currency/
T-Shirts: teespring.com/stores/eevblog
Likecoin - Coins for Likes: likecoin.pro/@eevblog/dil9/hcq3 - วิทยาศาสตร์และเทคโนโลยี
Apparently it’s not just a trap for young players.
Dave considers him as a young person.. so he is technically still a young player
I love it when channels notice a slight mistake from one video, admit they made a mistake, and release another entire video covering it. Awesome.
and you actually learn something from it
Learn from your mistakes, best way. You'll kick yourself so hard you won't forget. Second best way is to learn from other people's mistakes.
Yes responsibility and honesty are great character traits.
Dave, could you make a video about RMS, True RMS, DC +AC and all the uses and pitfalls of all of them?
The rhode'n'schwartz oscilloscope has such a nice GUI that it is just on another level compared to the other ones. It even has an explanation for the parameters included.
Nice Bob Pease reference, Dave. I remember his column in Electronics Design, one of the few magazines that I was bombarded with at work that was actually worth reading
If you don't have that feature just manually do Sqrt( RMS^2 - MEAN^2 )
Good for you with stepping up an owning your mistake. That's the professional way to handle it and set a good example.
Watching this video on full screen and seeing the reflection on the scopes, made me look over my shoulder for Dave ^^
04:15 am... And it took him almost two hours to film everything. Just shows his devotion...
Very instructive ! I did not know to use STDDEV for AC rms. Thanks !
OMG, still learning. Thank you !
This is how an honest channel should be.
Errors happens all the time, to everyone. Acknowledge you made an error and correct yourself; it is the only right thing to do, there is no shame in that. On the contrary.
Other youtubers, instead, choose to persist in their errors even when they are given proper feedback, and in doing so, they do a disservice to the whole community, by spreading ignorance.
Kudos to you, Dave.
Errors are opportunities to learn or re-enforce what you have learned.
Thanks. Good opportunity to show a concept in this case.
I just had an exam involving this exact thing a couple weeks ago, it's nice seeing it actually applied in real life.
Wouldn't it make more sense to properly terminate the inputs or at least fully shield them when measuring the input noise? If you leave them open, 1) the one with the lower resistance to ground has lower voltage noise and wins, 2) the inputs are very susceptible to surrounding e fields (and thus of the equipments position in your lab and even your position in your lab) as is visible in your traces as a periodic pattern.
I also thought it was strange to leave the inputs flapping in the breeze like that
Good stuff! I didn't know: 1) that AC coupling could give me an offset anyway, 2) that RMS != AC RMS, 3) that AC+DC in a multimeter didn't mean the display was dual. I kind of knew (2) because I wasn't getting the expected (AC RMS) results from RMS, but now I know how to get them!
Hey. Good on you for all of your frank, honest content.
Thank you
It's worth mentioning that the reason that Volts RMS *doesn't* remove the offset is because it was created to calculate the DC equivalent heating value ( or total energy value) of the source into a load, & for that purpose, obviously any DC offset *should* be included.
By coincidence measuring tiny AC noise signals with a SIGLENT scope is exactly what I am doing this week.
So, thanks for the timely reminder to use stdev rather than RMS, and to run the scope self-calibration!
(SIGLENT advises that the scope be left running for 30 minutes prior to self-calibration to let it thermally stabilise first).
Yep, always let it thermal stabilise first.
I started watching a video about AC/DC RMS. Ended up on the Highway to Hell.
EEVBlog did an oopsie. A lil' oopsie-doopsie!
… a trap for »young« players, is it?
And the wooden spoon award goes to Dave at EE vblog😭
Bell curve standard deviation is the sample standard deviation it is supposed to predict. Same same.
I think you shoud short your input, since you're measuring voltage noise. Open input when measuring current noise.
Dave, why don't you short-circuit the inputs to get a true 0V input signal?
Proof that maths IS important!!! While you don't need to do the calculations. You need to know enough maths (including summation, algebra and some calculus topics) to understand why 1/n vs 1/(n-1) actually matters.
Great video Dave and I think this is showcases while maths might not be people favorites, it does still come into play.
6:16 okay, hit me, why is it 1/(n-1)? If xdash is the mean and zero, then RMS and stdDev should give the same results, but they won't because of the (n-1)?
@@YensR sample mean vs population mean; long winded maths proofs are available on the internet ....
@@robiniddon7582 Hm, googling that turns up nothing that would explain (n-1).
keydifferences.com/difference-between-sample-mean-and-population-mean.html
Are you sure you are using the right terms?
Would appreciate any useful pointer that explains (n-1)
@@YensR en.wikipedia.org/wiki/Bessel%27s_correction
It's still essentially bandwidth. Normally, let's say audio. We use 20-20khz rms noise. Another truth is as the update rate of the scope, it's not possible to measure low frequency accurate at all. Whether using higher time/ division will help idk. Probably.
I've done a video related to update rate and noise th-cam.com/video/Znwp0pK8Tzk/w-d-xo.html
Hi Dave, I'm looking at buying some new multimeters one handheld and one benchtop. So what do you thing about TANMA meters?
good vid. like the old ones :)
Bonus info: you can use the Vmean measure function to get the DC component only, and sub this from Vrms to get "AC RMS" if your scope does not have the standard deviation function
Simply subtracting Vmean from Vrms is incorrect. The correct way should be "AC RMS" = Sqrt(Vrms^2 - Vmean^2).
Went Bob Pease style with the title, nice :)
Thanks for the nice explanation - nice focus on detail. Good advice for the young players starting out. As an old lab guy there's a German saying: Wer viel misst misst Mist. I don't think this translates well to English. Basically it means to switch on the brain machine before during and after measuring. Measurements can easily confuse you and no Murphy is not on vacation: even the fanciest measuring setup can and will fool you 😜. PS: I am working in a physics lab for more than 18 years now. If I make no oopsies for a longer period of time I am getting nervous. Mostly because I am not realizing the oopsies I make already.... It happens. There's no greater opportunity to learn than from your own oopsies.
EEVBlog did an oopsie. Nicely amended.
So this is essentially an apology video?
Time for pew ... I mean Ham News!
Hi Dave
Just checked my DS1054 and no Std Dev measure function. Doing a self cal as I write. While I've got your attention. I'm getting back into electronics as a hobby and purchased the DS1054 as a good entry level scope. Have you considered doing a series of videos on the why and how of using a DSO? Cheers
He has covered a lot of DSO info in the past; probably all you are looking for and more. Additionally, Tektronix has a very nice Oscilloscope guide pdf you should read. Just google "Tektronix oscilloscope fundamentals .pdf".
@@SuperSSSSooonniicccc Hi Thanks. Yes, I know Dave has done a lot of stuff on using a DSO and I've looked at that. Thanks for the link however that is a bit to basic as I'm familiar with the basic workings of an analog scope. I was hoping to encourage Dave to do some more videos on the why and how of some of the more advanced DSO functions. Any function that's not available on a basic analog scope is advanced to me. Cheers
I've done quite a few videos on various aspects, but they are scattered over the place.
@@EEVblog Hi Dave and thanks for getting back to me. Yes, I've look at quite a few of those videos and they are all, as usually, very helpful. I especially liked the one on using the DSO as a bode plotter as I was wondering if that was possible. As I said getting back into electronic, I'm comfortable with the basic of the analog CRO however the why and how of the feature of a DSO not so. It's having someone with the knowledge to confirm yes that is the way to use that feature and that is what to expect and here are the traps. I've had a bit of a look around and there does not appear to be a modern reference to using the DSO. I'd be surprised if I'm alone, so anything you could offer in this area would be appreciated. Thank again. Bill P
Love the Bob Pease title.
What's all this TH-cam comment stuff anyhow?
Will those scopes report the BW they are in based on the voltage selection? I don't know any that do, but I wish they did. It's all great to say you have a 200MHz or 300MHz scope, but what voltage range do you have to be in to get the full BW? If BW reduces with smaller voltage ranges, I want to know (without trying to find it in specifications) what the scopes current BW is. That is a trap for new players and us old farts alike.
BTW, after self Calibrating my precious (lol) TDS3032, at full BW (???) on the 1mv range, 1us/div (as in your video), I am getting about 37uv RMS (no stats function other than my eyeball on the old 3032). Before the recalibration, it was around 350uv RMS. For what I use it for, noise really doesn't matter, but very interesting. The BW at Vertical Division selection has me trying to look the specs up (it should be reported on-screen, dang it!).
Reflective screens are terrible except here where we can see you gesture while you're talking. Is that why they did it?
Puting N-1 instead of N (i.e Bessel correction) makes the estimation of the standard deviation asymptotically unbiased.
:)!
For modern scopes having from at least 1024 samples and up to many megasamples memory depth I would not care about the bias at all.
Hi Dave, I was curious and did a google of your multimeter. Would you say this is common for any high-end meter? In our labs at my uni, we sometimes would like to measure current, but most cheapy current measurement boards are for well above the
Is the input terminated with 50 Ohms?
👍 nice video
I uploaded it two minutes ago, you haven't watched it all yet, it might get progressively more sucky toward the end...
I know it is always a good video
Could you do a video on lock-in amplifiers?
What would be the N in this case? Would it be the number of points it acquired? If so, N would be sufficiently large such that the N - 1 doesn't matter and the sample std would approach the population std. Right?
I want one of those 121GW multimeters.
Looks like my Rigol DS2072A doesn't have ac rms, unless its hiding somewhere weird.
in the formula of the standard deviation, the "1/(N-1)....."??? why -1?
Actually, there are two! formulas for standard deviation, one uses 1/N to calculate stddev for the "population" (population includes all possible measurements) i.e. it would be the case if you have an oscilloscope with unlimited memory and you're measuring for an infinite period of time. And another stddev formula uses 1/(N-1) when you have a fraction of all possible values which is called a "sample". So, if you know all possible deviations from the mean - you have a population and then you can use 1/N multiplier. But when you only know a fraction of deviations from the mean - you have a sample. In case of sample - 1/(N-1) is used as a multiplier because when you have a sample, you have limited knowledge about all possible deviations, so you should be pessimistic and think that actual deviation is slightly bigger than you see, so, 1/(N-1) does this "correction" for you. It simply increases stddev value slightly to incorporate measurements which are not included into the sample.
Check out this for more clues - www.mathsisfun.com/data/standard-deviation.html
Dave, please do that again, make me feel good 😜 welcome to the human deviation!
I make a mistake in every video, guaranteed!
Dave, you just make me feel good 🤪 (but thx)
As a yank, guess I'm not going to sleep...thanks Dave.
I usually release in the morning my time which is evening Yankee time, but haven't done a video for a while so couldn't help being a tad premature this time.
@EEVblog
Could you make a video about the official Raspberry Pi power supplies and take them apart? Especially the ones for Raspberry Pi 3 and 4. How save are they to use compared to laboratory power supplies?
That would be a nice teardown video topic for bigclive!
Greetings: Sorry, but that R&S offset after the AC coupling is not acceptable to my way of things- fail.
Not to be naughty, but an offset of approx. 350µv makes you say it isn't good enough, then you just have to spend more $$$
Is the “fault” there, yes, is it important, not for me, 350µv has no meaning at all, special when you can auto-calibrate, but I guess it depends on what you are using it for. I use oscilloscope for audio, power electronics and incoming repair.
I would swap my 15 yer old trusty leCroy 9450 in a heartbeat with the R&S
That seems to happen on quite a lot of scopes. That's why they have the self calibration feature.
I would like to own a multi meter like the one you have. Where can I get one, Dave?
It's my custom meter, available on my website.
@@EEVblogdon't delay, buy today
👍👍👍
Hey Dave, you know I kinda wish you did a math lesson! Might learn alot!
Trust me, you wouldn't learn much from me about maths!
@@EEVblog want to bet? 😀
why RS has worse DC input offset than siglent? are they to cheap :D
Hey Dave, something different: have you ever heard about a memristor? I think speaking of which would make a great video!
Have to compare the power spectral density to be fair.
Shouldn't you have adjusted the DC offset before starting?
TRMS (is there false RMS ? no, it is just marketing ) should be AC+DC but almost all TRMS DMMs measure AC coupled. If you choose the AC+DC mode you get the thing that is the one and only true RMS (just called RMS ). You can measure AC, also AC-coupled (so the TRMS of most meters, then measure DC and then calculate the real RMS from that)
What! Did you not get given the latest raspberry pi 4 B?
Now I'm more confused than ever! Which doesn't mean much, could be the solvents, could be the chemicals or alcohol or maybe the repeated head trauma... and mom smoked while she was pregnant...
Bruh, who needs those fancy and weird measurements when your oscilloscope has a CATHODE RAY TUBE!
Dave! Who hates math? Math is your friend.
I love ohms law
I m with Dave on this one. Math is just a tool which I have to use to do engineering. You don't have to love your wrenches to be a mechanic.
Mr Dave doesn't it become exponential
I know this video is for Std. Deviation Demo but it would be better to terminate the inputs to see the real figures.
I love oopsees..
Mr Carlson's Lab oscilloscopes be like "yeah, but our operating voltage is bigger than yours"
I can't tell you how many times I tried to make my colleagues aware of the difference between STD and RMS. In my field, people tend to calculate the SNR of a signal as P2P/RMS, because that's just "what everybody" does. Nobody ever talks about the fact that RMS includes the mean and for clarity it would make much more sense to use STD. Why does nobody actually use STD? Well, because then people would tread onto the field of statistics and make themselves vulnerable. As long as you use RMS people will assume an engineering/electronics context and not ask too many maths questions. *facepalms*
too bad the 1000z series is too low end for this advanced feature
tldr either know your maths and your scoe for that matter or simply calibrate this darn thing again
Clickbait ;-)
Lower noise is artificial. I rather see everything that is going on, lol