This video had me turning my Rigol DS1102E scope on just to play with the averaging! Now I know how that works, thank you! As usual, great video. Please keep them coming.
If you would be my english and electronics professor every people in my university would go to your classes. You teach electronics son fun and usefully, you are an awesome man.
I think Dave left out an important point (or I had a senior moment). All the fir filters, whether boxcar or more complicated, reduce bandwidth, which may be important in certain situations e.g. high speed timing measurements.
High resolution can show you smallest details. And I can see that. It is important for communication. High resolution shows you true signal from your probe.
I think this effectively is boxcar averaging although it doesn't look like it. I remember many years ago using boxcar averaging and it meant replacing every N samples with N values equal to the average of the N samples. So the result looked like a string of boxcars in a train. A ramp would therefore look like a staircase. The digital scope doesn't quite do that it replaces every N samples with a single value equal to the average of the N samples and then draws vectors between them producing a smoother looking curve and a ramp looks like a ramp.
In NMR this is called digital oversampling, reduces physical audiofilter interference and also digitalization noise. And for a real amazing application, Sonys DSD for audio recordings, a one bit ADC with sampling rates in the 20 MHz range.
Averaging mode can tell you where an over-modulated FM signal could be picked up on other transmitter frequencies, which is a bit of a problem in 'Murica. Might be useful for transmitters who want to play nice with their neighbors.
Great explanation of what the two modes do but is "boxcar" the right name for high res oversampling. Wikipedia describes a Boxcar averager as a gated averager and that seems to be something different. It actually seems to correspond to the normal scope averaging mode for averaging periodic signals. Dare I suggest Dave has it the wrong way round?
Yes they seem to use it to mean simply averaging over n successive samples but the rest of the internet seems to think it is something different. See sunnytek.net/admin/xiazaifiles/2010114103920522.pdf. That describes a gated integrator before the ADC and then averaging the samples digitally over successive cycles of a periodic waveform. The gate pulse is then swept over time to build up the waveform. Lots of other references to the same thing.
Wow! That built-in waveform generator does have some pretty advanced features (for a built-in gen) like modulation and noise. But the question is, can it do sweep?
Heh. I had to discover this stuff myself in Electronics Lab at uni. I figured out averaging mode fairly quickly, but it took a while to understand High Res mode. Now I default to High Res unless I know the signal is periodic with *no changes* (like data packets that might have different values). Averaging tends to make the cleanest signals, but you can't always use it.
Can the average/high resolution mode be used to determine if a "noise" is random or not? What looks like noise might not be random but some form of HF coupled into your signal. Based on the amount of "noise" reduction between high resolution and average you should be able to determine if it's random or some form of repetitive HF.
Although, if you adjust the trigger holdoff you could probably make averaging work the same way as high res. However, this is more of a kluge than a viable measurement solution
I would think that high resolution mode and Peak detect mode could possibly be used at the same time? when your sweep is slow you might want Peaks to stick out even though you're even higher frequency noise is being averaged out on the Fly. or maybe you don't want that? haven't really thought about it all that long
Just wondering if anyone else has EVER heard or seen it called Boxcar averaging, besides from Dave. Never heard it elsewhere, and I've used a LOT of scopes.
+Aurelius R : The term "boxcar" is standard, at least for me. It is used in DSP as a type of windowing function. Any scope that offers FFT functionality should offer boxcar along with hamming, hanning, and other windowing functions. Using the phrase "boxcar" just means that you look at the last N samples. Another way is to use constant averaging, where each new sample will make up N% of the total, and (100-N)% of the previous average. This means that one stray signal will eventually decay (as if the average had an RC time constant). I hope this helps.
Very nice explanation, people will tell me RTFM but I'm more in this kind of example explanations (and I think a lot more people 😎) so keep them coming BRILLIANT!!
+Joost B It's a different type of averaging. 'Average' requires multiple samples, it then averages all samples together. Eg. data point 1 is averaged across all samples, data point 2 is averaged across all samples, etc, etc. 'Hi-Res' increases the sample rate. It then performs a moving average on the sample. Eg. data point 1, 2, 3, and 4 are all averaged if it was a 4 bit average. Next data point 2, 3, 4, and 5.
Come a gutsa: english.stackexchange.com/questions/130057/what-do-australians-mean-when-they-say-he-came-a-gutsa So basically, "It can really go wrong", if i understood that correctly?
+Daniel Segel Really "come a gutser'" 1. (Australian informal) to fall while you are walking or running I was in a rush, tripped and came a gutser on the step. 2. (Australian informal) to fail at something Paul's too confident for his own good. I hope he doesn't come a gutser.
3:10 - $12000+ scope yet look at the delta Y jumping between mV and V. Hard to follow when it is changing ranges so fast and hard to believe a scope in this price range wouldn't fix that!
Wow, I'm very disappointed hi-res & averaging can't be combined. Figured that was just a Tek thing. Really, scopes have no need for a "hi-res" mode. They should simply store the highest res possible at any given sample rate.
+qwaqwa1960 It's a way to trade bandwidth for bit depth. We don't live in a perfect world, so it's hard to manufacture cheap high rate ADCs with high bit depth (and low enough power etc..).
peteabc1 You perhaps misunderstood. Why have a switchable hi-res mode? Just max out the res automatically, so that resolution goes from e.g. 8b @ full rate, to 16b or more at slow sampling.
"Didn't anyone ever tell you to make sure your optics are clean?" -Kent, Real Genius (1984) Clean your screen Dave! Those big smudges are driving me crazy!
+EEVblog Dave, got that all along... Bob's my uncle, but he stops by once a week and cleans all my touchscreens. Maybe it's just a "Yournita States of Uhmerica" thing... Love your videos!
This video had me turning my Rigol DS1102E scope on just to play with the averaging! Now I know how that works, thank you! As usual, great video. Please keep them coming.
If you would be my english and electronics professor every people in my university would go to your classes. You teach electronics son fun and usefully, you are an awesome man.
I think Dave left out an important point (or I had a senior moment). All the fir filters, whether boxcar or more complicated, reduce bandwidth, which may be important in certain situations e.g. high speed timing measurements.
High resolution can show you smallest details. And I can see that. It is important for communication. High resolution shows you true signal from your probe.
I think this effectively is boxcar averaging although it doesn't look like it. I remember many years ago using boxcar averaging and it meant replacing every N samples with N values equal to the average of the N samples. So the result looked like a string of boxcars in a train. A ramp would therefore look like a staircase. The digital scope doesn't quite do that it replaces every N samples with a single value equal to the average of the N samples and then draws vectors between them producing a smoother looking curve and a ramp looks like a ramp.
In NMR this is called digital oversampling, reduces physical audiofilter interference and also digitalization noise. And for a real amazing application, Sonys DSD for audio recordings, a one bit ADC with sampling rates in the 20 MHz range.
Averaging mode can tell you where an over-modulated FM signal could be picked up on other transmitter frequencies, which is a bit of a problem in 'Murica. Might be useful for transmitters who want to play nice with their neighbors.
Great explanation of what the two modes do but is "boxcar" the right name for high res oversampling. Wikipedia describes a Boxcar averager as a gated averager and that seems to be something different. It actually seems to correspond to the normal scope averaging mode for averaging periodic signals. Dare I suggest Dave has it the wrong way round?
+nophead Both Keysight and Lecroy at least use the term boxcar averaging. It's in their app notes on the subject of high res modes.
Yes they seem to use it to mean simply averaging over n successive samples but the rest of the internet seems to think it is something different. See sunnytek.net/admin/xiazaifiles/2010114103920522.pdf. That describes a gated integrator before the ADC and then averaging the samples digitally over successive cycles of a periodic waveform. The gate pulse is then swept over time to build up the waveform. Lots of other references to the same thing.
Wow! That built-in waveform generator does have some pretty advanced features (for a built-in gen) like modulation and noise. But the question is, can it do sweep?
Heh. I had to discover this stuff myself in Electronics Lab at uni. I figured out averaging mode fairly quickly, but it took a while to understand High Res mode.
Now I default to High Res unless I know the signal is periodic with *no changes* (like data packets that might have different values).
Averaging tends to make the cleanest signals, but you can't always use it.
Can the average/high resolution mode be used to determine if a "noise" is random or not? What looks like noise might not be random but some form of HF coupled into your signal. Based on the amount of "noise" reduction between high resolution and average you should be able to determine if it's random or some form of repetitive HF.
Have you noticed that while pressing the Single Shot Button the Acq Mode switches to High Res...(at min 4:10)?
Although, if you adjust the trigger holdoff you could probably make averaging work the same way as high res. However, this is more of a kluge than a viable measurement solution
Congrats on 1000 videos m8.
Simply speaking - high resolution mode is just an additional low-pass digital filter (FIR in this case).
Nice to see my old Agilent slides on high-res and averaging still in use. :-)
I would think that high resolution mode and Peak detect mode could possibly be used at the same time? when your sweep is slow you might want Peaks to stick out even though you're even higher frequency noise is being averaged out on the Fly. or maybe you don't want that? haven't really thought about it all that long
Just wondering if anyone else has EVER heard or seen it called Boxcar averaging, besides from Dave. Never heard it elsewhere, and I've used a LOT of scopes.
+Aurelius R I mean, I know that's what it actually is, I've just never seen a scope label it that way.
+Aurelius R Both Keysight and Lecroy use the term, but not in the scope, just in app notes and things.
+Aurelius R : The term "boxcar" is standard, at least for me. It is used in DSP as a type of windowing function. Any scope that offers FFT functionality should offer boxcar along with hamming, hanning, and other windowing functions.
Using the phrase "boxcar" just means that you look at the last N samples. Another way is to use constant averaging, where each new sample will make up N% of the total, and (100-N)% of the previous average. This means that one stray signal will eventually decay (as if the average had an RC time constant).
I hope this helps.
Very nice explanation, people will tell me RTFM but I'm more in this kind of example explanations (and I think a lot more people 😎) so keep them coming BRILLIANT!!
So high res mode is like moving average with extra samples. Is that correct?
+Joost B That's how I understood it as well.
+Joost B It's a different type of averaging. 'Average' requires multiple samples, it then averages all samples together. Eg. data point 1 is averaged across all samples, data point 2 is averaged across all samples, etc, etc.
'Hi-Res' increases the sample rate. It then performs a moving average on the sample. Eg. data point 1, 2, 3, and 4 are all averaged if it was a 4 bit average. Next data point 2, 3, 4, and 5.
I like to think of high res mode as point-to-point averaging, whereas averaging mode is waveform-to-waveform averaging.
I like to think of high res mode as point-to-point averaging, whereas averaging mode is waveform-to-waveform averaging.
KEYSIGHT OSCILLOSCOPES That's a concise way of putting it. I like it. Are you the/an official UA-cam channel for Keysight?
What's Dave saying at 11:51? "Really xXxX". I've heard the phrase before but don't understand it. Australian slang?
Come a gutsa: english.stackexchange.com/questions/130057/what-do-australians-mean-when-they-say-he-came-a-gutsa
So basically, "It can really go wrong", if i understood that correctly?
+Daniel Segel Really "come a gutser'"
1. (Australian informal) to fall while you are walking or running I was in a rush, tripped and came a gutser on the step. 2. (Australian informal) to fail at something Paul's too confident for his own good. I hope he doesn't come a gutser.
+Daniel Segel
he said "really come a gutter", it's auzzy land slang
3:10 - $12000+ scope yet look at the delta Y jumping between mV and V. Hard to follow when it is changing ranges so fast and hard to believe a scope in this price range wouldn't fix that!
I haven’t got the idea: why, in HiRes, greater the resolution less is the bandwidth?
Very informative video, thanks! A quick look at the merits of peak detect would be appreciated too! Best regards
+Koffi Banan Yeah, decided to leave that one out and do separately, video was already too long. Although I have mentioned it several times before.
Very handy. Thanks!
thank you good video ... oh what ever happen to the Microsoft InstaLoad Battery Technology haven't heard of any its been a few years
Time for a Mail Bag
Why is it so? Just ask Julius Sumner Miller. Physics is his business.
Wow, I'm very disappointed hi-res & averaging can't be combined. Figured that was just a Tek thing.
Really, scopes have no need for a "hi-res" mode. They should simply store the highest res possible at any given sample rate.
+qwaqwa1960 It's a way to trade bandwidth for bit depth. We don't live in a perfect world, so it's hard to manufacture cheap high rate ADCs with high bit depth (and low enough power etc..).
peteabc1
You perhaps misunderstood. Why have a switchable hi-res mode? Just max out the res automatically, so that resolution goes from e.g. 8b @ full rate, to 16b or more at slow sampling.
+qwaqwa1960 Aha now I understand. That would be nice feature.
"Didn't anyone ever tell you to make sure your optics are clean?" -Kent, Real Genius (1984)
Clean your screen Dave! Those big smudges are driving me crazy!
+Stephen Furr It's a Touchscreen
+Michael Schneider Yes, and I got giant smudges all over my iPhone so I'm a bit of a hypocrite... ;)
+Stephen Furr It's a touch screen.
+EEVblog Dave, got that all along... Bob's my uncle, but he stops by once a week and cleans all my touchscreens. Maybe it's just a "Yournita States of Uhmerica" thing...
Love your videos!
I look at this and dream of my (non-existent) oscilloscope. 6 year old $10 multimeter is the best!!!
+Trish Mapow
Same, I have one of those that dave keeps throwing away in his mailbag, like the ones people in the usa get for free
yay, sinc waveforms
hi
that shows: modern oscilloscopes are all just confusers :P
1st
+Akarsh Agarwal Damn it! 70 views in under 3 mins!
+Paul Bearer And it's not even peak time to release the video. Mornings my time usually produce the fastest view results.
+EEVblog dave it just does not matter. We will watch the videos any time of the day you publish it. :)
EEVblog us Brisbane folk like to have a little after dinner tutorial with our after dinner beer Dave. Cheers :D
+EEVblog some of us work night shift and you release videos while we're awake. it's a nice treat!