EEVblog
ฝัง
- เผยแพร่เมื่อ 9 เม.ย. 2014
- Why do digital oscilloscopes appear noisier than traditional analog oscilloscopes?
Dave busts the myth that digital scopes are noiser than analog scopes, and demonstrates what inherent advantages digital scopes can have over analog scopes in terms of true waveform capture. And also why your analog scope may be hiding important signal detail from you.
Demonstrations of how memory depth, analog bandwidth, averaging, and intensity graded displays can all effect the signal detail you see on your digital oscilloscope.
And how long exposure camera shots on analog oscilloscopes can reveal detail you can't see with your eyes.
Demonstrations are done on the new Tektronix MDO3000, the Rigol DS1052E, the Tektronix TDS220, and Tektronix 2225 analog oscilloscope.
Previous video on common mode noise measurement: • EEVblog #442 - Analog ...
Discuss on the Forum here: www.eevblog.com/forum/blog/eev...
EEVblog Main Web Site:
www.eevblog.com
EEVblog Amazon Store:
astore.amazon.com/eevblogstore-20
Donations:
www.eevblog.com/donations/
Projects:
www.eevblog.com/projects/
Electronics Info Wiki:
www.eevblog.com/wiki/ - วิทยาศาสตร์และเทคโนโลยี
![[TH]FFWS SEA Spring 2024 Knockout Stage - Day 16](http://i.ytimg.com/vi/jpvPyp7LESA/mqdefault.jpg)
This tutorial may be 9yrs old but it still holds up today. Brilliant, just BRILLIANT and informative.
Oh, stay in darkness for half an hour or so, like you would for stargazing. Then look at an analog scope with the brightness juuust high enough to see. It's just amazing how much information you can see. I found it interesting with music on the scope. You could see all sorts of details that are invisible under normal light.
i'm learning so much with you, things we never do at the university, so thx
+Yacine Wlid Ahlem Yeah, totally. 99% of my knowledge came from this channel and on-the-job trial and error!
@@stuartkerr1012 what job do you do/have?
My Rigol 1054Z shows the same consistent noise as the one Dave got a few years after he made this video. I've managed to determine that this "noise" is a 250MHz sawtooth wave about 10mV p-p; it isn't random. I'm not sure if it's sampling noise introduced by the ADC or switching noise from the power supply riding one of the voltage rails; either way, it seems to me this is something that could have been snubbed in hardware or filtered out in software. So high above its specified 50MHz bandwidth; you'd think the designers wouldn't want it showing up on a trace.
All of us poor buggers that wish we had a scope look at Dave's bench and think "He has more scopes than a big game hunter!"
You can always get a cheap analog scope.
I finally got a scope (Siglent 1102CML). Dave should do a review on Siglent.
Digger D
I bought today a Kenwood 20MHz scope, I was going to get a Tektronix 60MHz one but the guy sold it and I only was able to get the Kenwood one xD.
There's something about analog scopes that I love.
Digger D I got myself SDS1052DL when I had no money, now I'd get 1104 or somethin. It's not that good but enough for me.
I think most inexpensive 50 and 100 MHz scopes would be fine for most hobbyists. I'd still like to see a Siglent tear down :)
Dave, you are reading my mind. THANK YOU for this video explanation. I have only used analog scopes and recently bought a new digital one and thought it had a defect because it looked so noisy.
This is awesome.
Thanks for educating me!
Thanks for explaining this. Well done!
Excellent work Dave. Really appreciated.
Great explanation Dave. Thanks for sharing....
Nice, thanks for sharing Dave
Thank you for these great videos!
Absolutely Brilliant Dave ! You're the Best !
Thanks for the explanation, with a good demonstration.
Excellent explanation of digital vs analog. You hit a home run with this one !! Thank you
You saved my day. Something with a very expensive new scope, high expectations and too less experience (but much motivation to learn)
Nice, thanks for sharing Dave,
There are lots of ways to explain something, and after all those ways, there is the Dave-Way to explain things which of course, is the best of all... Bravo!
Thanks for the great video, learning all sorts of things.
8:38 very nice way of showing your point
Thanks Dave, point taken. Cheers
Thank you. I did not know this. been using an analog scope my whole life and never wanted to switch to digital before this.
This has driven me bananas on my Rigol for years. I had a basic idea of what was going on since turning on BW limit and shortening the memory depth cleaned up the trace, but I never knew why noise was ALWAYS present. Thanks.
I've been searching for these answers to why digital scopes do this.
This is awesome.
One of the simple things I want to use a scope for is just to find clipping of an analog audio frequency waveform.
Great video. I am very impressed. Best regards Erik.
Good stuff, Dave!
Super useful information, thanks a bunch!
8:07 "you gonna be able to see it! watch!" .... yeeeeeep... I've seen it very very clear!
Great video, makes the point well.
Good demo and excellent explanation. Nice one Sir :-)
thanks Dave, I learned something, great video!!!
I have a Tektronix TDS 210 that I bought brand new way back then, weirdly enough it's about at the time I stopped doing electronic. Now that I'm back in the Hobby I'm very glad to have it. It still works great.
I was thrown by the 'noise' on a new Rigol . That was overshadowed by two probes that created large glitches when the leads were moved. Telonic (UK) listened and replaced the probes under warranty. I'm now very happy with the scope and have accepted this facet of digital scope use. Nice vid explaining this non issue,
very nice work dave
I've always thought about it in this way: The more "noise", the more sensitive the equipment is at picking up data... (Camera, Audio recorders etc...)
Actual AC RMS input noises I've just measured @ 5, 10, and 20 mV/div:
- Tektronix 454
(150 MHz, on Ch1 output): 60, 80, and 110 uV
- Keysight DSOX1102G (2 GS/s, 70 MHz) : 120, 180, and 400 uV
- Siglent SDS1204X-E (1 GS/s, 200 MHz) : 120, 50, and 400 uV
The strange (and strongly varying) result on Siglent at 10 mV/div is the result of quantization error. The thermal noise rarely reached the threshold of 1 LSB, therefore it was truncated to 0 too often to get a correct measurement. And this is exactly the reason why noise of a digital oscilloscope must be higher than one of an analog equivalent.
Dave you saved my day. I thought my new MSO1104 was faulty. I know now it is only more sensitive and just found High Res Acquire mode that shows graphs I expected. Thanks!
I just stumbled upon this universe of oscilloscopes, and as a musician and sound design hobbyist, holy shit if this isn't interesting.
I love that there's a universe of people doing the inverse of what I'm doing, and that in this universe and that this universe had their own arguments about analog vs digital, just like in the music universe I reside in.
I am going to buy an analog oscilloscope, and I may quite possibly get tangled up in this universe I stumbled upon as well. If so, thank you for being part of my future universe.
Fantastic video!
Nice video Dave.
As someone who has been brought up on Analogue scopes, Gould, Hameg, Tektronix, etc....having to use a Tektronix 720 Portable Digital was quite worrying.
I have been fooled into thinking my circuit I was fault finding on was worse than it was. I saw many times on your videos, the apparent noisy waveforms you seemed not to be worried about and thought why was this.... This video has helped enormously.
Thanks again Dave.
Very cool video Dave!
Nice explanation of "noise" on scopes
Brilliant video.
Very helpful, thank you. I was interested in this as I am a 'young player' and recently purchased a Hantek DSO 2D15 which seems noisy.
Could have been a great video but those no signal tests should have been done with terminated inputs on the scopes aka grounding caps. Is the noise we see being picked up by the little BNC stub in the female socket on the scope? It's high enough frequency noise that the connector probably makes a good antenna. Or is the noise from the A/D converters or numerous processors in the digital scope. Without grounding the inputs all the data is invalid.
Very informative, Dave. Thank you.
that temperature waveform palette mode is incredibly useful seeming
instead of losing the noise information to gain the average information, you can see both
I purchased a Hantek 100MHz scope about a year ago and I honestly have to say that it's the best piece of kit I've ever spent money on. While it isn't super top of the range, it's certainly more than adequate for my needs at the moment. I would really love to get a network anayliser next but sadly, I couldn't even afford the dust cover for one lol.
Brilliant! As a recent buyer of a digital 'scope I was ready to send it back to the supplier because of the noisy traces. Now I understand what is really going on, the analogue 'scope will go into the skip - well eventually ;-)
great vid, just back from national electronics week here in uk and while I was there over heard some old guys raving over how analogue scopes are much better, man just cant see it myself and this is just more proof why, sometime I think they get stuck in the past
Hmm ... I've got both an old analogue and a modern digital scope (Siglent SDS1202X) and have noticed this same effect. But all the scopes in the video are set to 1 volt / division and on the digital scopes we're seeing 'noise' of about 1/20 of a division. That equates to 50 mv. Full screen deflection is 8 divisions or 8 volts. So the noise is 160th of the full screen, give or take. You might expect something like this with an 8 bit A/D converter. So maybe the problem is insufficient resolution. Certainly, when I use my Siglent in the single trace mode to capture a one-off event, you can see the digitalisation and it's about that magnitude.
Also, there is nothing connected to any of the scopes. I would like to see them with the inputs shorted to ground, just in case they're picking up external noise.
With the money they charge for modern scopes, why cant they put a 2560x1600 display on them and get closer to around 260-300ppi?
I high pixel density display on a smartphone of around 5 inches, is often around $50
With oscilloscope companies getting number happy and charging nation state bankrupting prices for their scopes, they should at least be including a high resolution display.
The asic doing all the signal processing from memory is in essence also producing the waveform display. The asic is the limiting factor for display resolution. Accommodating a larger resolution would mean the entire asic would need to be an order of magnitude more powerful. Besides, if the ADC is doing 8bit conversion, you'll only get 256 lines of vertical position. Increasing the resolution does not increase the adc bitrate
But it will allow for less graphical aliasing of the lines. Think of it like when you move from a 1920x1080 display to a display that may be only 1-2 inches larger but the resolution is 2560x1600, and you suddenly don't need enable antialiasing, (or at most, only need around 2X antialiasing in order to make things smooth. The points of data making up the waveform or anything else being displayed may not increase, but things will look smoother. It will also make the text and other GUI elements look much better, or allow for more efficient use of the screen space, by making smaller GPU elements that are still easily readable. For the prices these companies charge for high end scopes, they should throw in a R9 290x while they are at it, just so it can drive the display a little better, and also mine bitcoin while you are while you are using the scope.
***** Is there a good reason why they can't make that a 12 or 16bit ADC, or is it just not needed?
Razor2048 It's not aliasing, the asic displays only absolute data points (data events that it registered and stored). Each pixel on the display is an actual data point (or combination thereof). Interpolating your data is the same as data manipulation.. You use an oscilloscope to get answers, not to get pretty interpolated lines. You can compare this to upscaling photos from a 1.2mpixel webcam to 5mpixel. With some smart algorithms and aliasing, you can probably get the image to look nicer, but it's not giving you more information
Kevo F I think Tektronix has a couple of 12 bit oscilloscopes, maybe also 16 bit but I'm not sure. They're not really necessary, thus not very popular. With increased bit depth, you lower your noise floor, but decrease sampling speed (and probably increase costs). With increased bit depth, you also make quantification a bit better. It's not very common that the noise floor interfered with what you use the scope for. And an oscilloscope is by nature not a quantification platform, even if it is possible to do that
wow this one is super helpful for engineering students like me :D now the errors are more explainable to instructors :D
Your camera set up reminded me of something.
>Be at fleamarket.
>Pick up CCD camera on table full of junk.
>See it's a 55mm f1.2 Nikkor O (oscilloscope lens)
>Price 2 dollars.
>Know more than enough to know it's a fast lens.
>Before I can pull out wallet seller says you can have that for a dollar if you can use it.
>Maintaining poker face as hard as possible 😐
>Hand him a dollar and say, "I'm just buying it for the lens, you got anymore of these?" 😐
>"No that's the only one." 😐
>Get home and look up lens.
500-600 dollars. 🤑
This was a brilliant video Dave. Thanks!
"There's a myth about oscilloscopes that will not go AWAAAAIIIIII!!!"
This is excellent @_@ i learned so much today
Thanks from 2020!
I'm a software guy these days but I never knew about the camera hoods for scopes, very interesting video mate.
this is so cool!!!
Nice you got nearly the same analog Tektronix oscilloscope like mine, but mine's 2205
Makes me kinda sad to hear the DS1052E being called a "cheap old" scope. It still does alright!
Yeah, but it's around 7 years old or so, that's more than a lifetime in this market.
Analogue grey beard here. Just bought a DSO because the old 20MHz unit lacks bandwidth. So, now I know why it appears to be noisy, thanks!
Interesting Video!
Its interesting that digital osc's have done exactly what digital cameras have done. As time passes they get higher and higher resolution and now have extremely high resolution. I come from the old school of film and large format. Digital is the only way to go now.! The difference however, is that film started with higher resolution and digital cameras had to catch up. Additionally, osc' s try to emulate there analog counterparts. The same thing happened to digital cameras emulating film.
Would be nice to see a comparison with an analog MCP (microchannel plate) CRT scope like the Tek 7104 or 2467 series.
Awesome video! I didn't know that. Although I prefer and use digital scope all over my lab and chop :) Cheers and thanks for sharing!
Thanks very much, that was absolutely brilliant. My old Tek 'scope just blew up and I'm trying to get my head around the new entry level digitals
So if a manufacturer offers 3 oscilloscopes in the same series, let's say Hantek 5000 series, the only difference is the memory depth, the DSO 5102P is 40K, the 5102M is 1 M, and the 5102BM is 2 M but they should otherwise be identical. If they were all lined up in a row, the 5102BM would appear "noisier" than the M, and the M would appear "noisier" than the P?
Thanks bro. I am an old timer who has used nothing but analog scopes, so I had no idea that square noise ever existed. LOL
For my first digital scope I bought a Tek 2232 which had analog and digital modes. For some measurements with noisy signals, analog was better. One analog technology not mentioned is CRT storage. I have a Tek 7834 which can capture single shot waveforms at high sweep speeds, but its best use is as a long persistence display for a spectrum analyzer plug-in.
@9:11 The wider the scope bandwidth, the more noise you'll see. Analog scopes don't have as wide a bandwidth so higher frequency components (such as noise can't be recorded)
Looking at the time on the oscilloscope is kinda weired. In Germany, it's now 13:21 and the oscilloscope shows 18:13. The same day :D
Funny coincidence looking at the stack of equipment, I have a Rigol DS sitting on top of of a Tektronix analog scope sitting on top of a Tektronix TDS. Not the exact same model numbers but close enough to catch my attention - 1052E, 2215 and TDS 784C.
20:35 oh god lol, I was really hoping you weren't going to tell me those ridges were for long exposures
Any application I would have for an oscilloscope would only involve audio frequencies perceptible by chordates. Within this frequency spectrum, what advantages do DSOs offer over analog CROs?
I bought a Hantek digital DSO5202B 100 Mhz scope some years (9) ago when Silicon Chip paid me for an article. I find I still prefer my many Tektronix analogue CRO's to do most servicing work and it's not about the perceived noise. I am in general not looking for a random event but signals that are repetitive so a digital storage scope does not usually help me. I bought the instrument in order to capture pictures of waveforms to include with electronics articles. I believe in some procedures the analogue scope does a better job.
Some advice please dave,if given the choice of the FNIRSI 1014D or Hantek DSO2000,for audio work on valve/tube stuff,and the fact you want to use the display to give a good account of Fhz and Vpk-pk to use as calibration for your Farnell LFM4 sig genny.
The only digital i envisage would be pwm for on either boost converters using an inductor or a push pull transformer type for HT supply,otherwise for flip flop switch control for channel sellection.
I can't get anything on ebay or local with screen display measurements accurate enough(trace) to do any calibration,and Fhz meters on sale used may need calibration anyway,plus the scope has i have has one channel fully working with one channels V/div not working on some settings,i can repair it as the y amp has to be working,but the V divider must be faulty(i assume).
So chasing my tail trying to find accurate pk-pk voltages and Fhz would require more test gearwhen a DSO seems to have self test and Fhz and signal voltages on screen.
Both the DSO units are under £200 on ebay have a guarantee and seem the best candidates for what i want,but which would you go for,as the cash difference is only £50.
Johnamptech in Scotland,
Cheers.
You sould make a shoot out on intensity grading and color corrected temperature grading, as its not soemthing many vendors are that op0en about the levels of intensity grading and even then it can vare a lot from one scope to the next
intensity grading and CCT is features I personally use a lot.
When I spotted this I wondered "Is it because the digital scopes are more sensitive than the old analogue ones?" Seems I was on the right track.
Oh neat. On my Siglent 1202X-E I got a variable result. The color grading worked as expected, but the intensity really did not. It just uniformly dropped the intensity of the waveform on the display... made the whole thing uniformly more or less bright but didn't make the waveform look any thicker or thinner. Averaging mode worked quite well on its own but didn't change how intensity worked.
-Matt
Just as an extra and I belive worth a mention ! My ex students had problems with digital oscilloscopes with digital aliasing ! "The wrong time base setting would sample incorectly and would look like the wrong frequency on the display" where as good old analogue scopes wouldn't do this!
I wish i could have what you have, Teckronix oscilloscope. Might get Hantek or Rigol for me.
Very err, enlightening.
Very intresting
The best thing with the Tektronix 2225 ,50MHz is that is has no fan, -perfect for does late night operations.
Hello there-just a tip-maybe you could do a tutorial on inductance matching and termination and all this high speed design stuff? Just putting that out there as an idea!
when is that Tek 3000 review coming? ;)
12:16 - So, now I understand the noise I’m seeing; but, what about signals that appear as ghosts 17:16 (like I’m seeing double or sometimes even ten(ple)? Noise makes it fuzzy up and down; but, the ghosts make it fuzzy side to side even when triggered.
It is likely (hopefully) the case that a tiny signal averaging 1/2 least significant bit is added to the input to "dither" around the LSB. Without it, you get a huge step increase going from zero to 1 but your signal might have only gone from 0.99 to 1.01
So a non-harmonically related sawtooth, or just thermal noise, (random but at least non-harmonic) noise will, for instance, at one moment in time be 0.95 and now your tiny signal is enough to push it over the threshold from a zero to a 1. But a moment later it doesn't. The *average* of these sometimes-0 and sometimes-1 will converge on the actual value. Let's say three measurements of 0, and one measurement of 1, repeat. The average becomes 0.75 and can be displayed as 0.75 which is accurate.
This is why (probably) choosing less memory depth on the Tek scope reduces noise; under the hood the sampler is going full speed, and with some dithering noise added to average out the quantizing error, then subtracted. So you might take a thousand noisy samples in extremely rapid succession but store the average as a single sample. The effective ADC bits is increased by doing this; a Keysight DSOX1202 can deliver up to 12 bits of effective vertical resolution even though its ADC is only 8 bits. You get this effect more or less automatically on a Keysight (unless you tell it not to); it optimizes memory depth based on sweep speed and probable zoom setting. If you want maximum zoom then you'll have to run the ADC at maximum speed and now you're back to 8 bits per sample.
I just realised that Dave took his shirt off to capture the analogue scope. Haha...
Hi
Can you explain how to meassure siemens with a osciloscope this to measure conductance on like a battery
The thing I don't like on the digital scopes is the slow refresh time & resolution of the display.
Fascinating, especially after I'm now in the digital oscilloscope world after finding a dead one in the skip that just needed a new psu!! (videos about it on my channel). Would it be possible to do a video on triggering some day, I haven't a clue about it :o)
What digital scope model was it?
I've done two videos that involve triggering:
EEVblog #159 - Oscilloscope Trigger Holdoff Tutorial
EEVblog #387 - Oscilloscope Trigger Jitter
Ive never done it at all :(
Its a tek tds 3012 Dave. Ive not come across those vids so ill check them out tonight :)
BIG THUMB UP.
But depending on what you are probing and if you know roughly what you’re expecting, could the analog scope be more easy to read coz it’s kind of filtered out the noise?
That's my feeling as well. It's kind of like seeing too much detail in a photo, so photographers sometimes add a little blur to make the photos more readable.
I like the fact the digital scopes are so much smaller and lighter. As Dave can attest work benches get crowded and having a few extra inches on your bench is a huge help.
That is. I like analog scopes. But have no room for them at my home workbench which even almost have no space for digital scope.
Why does my digital scope shift the trace when I increase the sensitivity of the input channel with no input? My old Tek scope has a DC Balance adjust so the trace is at the same place no matter what the vertical sensitivity is set at.
From 09:00, it is time for a question.
Are digital scopes able to display a short spike in intensity with reduced brightness?
That is, can a vertical trace be 25% luminescent, if only 256 of 1024 samples in that X axis are 1v, while the remaining 768 are at 0.02v, +/- 0.005v?
The intent here is to know when your peaks are either too short to be significant, or too quick for your current timescale.
( 15:00, answered. )
Answered at 15:00 ; Waveform / Graticule Intensity.
Busted
So given all the visible noise on digital Scylla scopes, how does one differentiate between a noisy audio amplifier stage and the noise in the aucilla scope? With my analogue 20 to 65B, I can see if an amplifier stages putting out a pure sine wave or has some noise on it. But these scopes have noise on the sine wave no matter how clean they are. And the other problem is the low resolution of these displays exacerbates the noise I think that's the real reason why these scopes look noisier. These displays in 2020 should be able to show 300 points per inch Just like our smartphones do. It's ridiculous that and over $10000 instrument is stuck in 1990 as far as display resolution.