EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: dirtyfly on August 07, 2020, 09:46:41 am
-
Im replacing my failing 40MHz analog scope.
I never needed anything more than the dual 40MHz.
Now I cannot devide if I want 200MHz 2 channel or 50MHz 4 channels
price is the same.
Can anyone give me some insights to help me decide ?
thanks
-
What do you use it for?
200Mhz isn't as great as it sounds unless you're using BNC cables. Probing 200Mhz with a standard probe is difficult.
-
thanks
I usally use it to debug retro computer and some ham radio stuff, never needed more than 35MHz
I'll be using the i2c and rs232 decoders that up to now ive used in my pc
Never used more than 2 channels.
If the 1054 was 100mhz there qwouldf be no doubt...
I'm afraid ill need more than 2 channls (cant imagine why :D
)
-
I'ld always have the four channels rather than higher bandwidth. Even the DS1054Z (hacked), with only one or two channels active has a considerable bandwidth margin over 100MHz. So, if the Rigol scope is otherwise okay for you, I'ld recommend to get the DS1054Z. There are other brands possibly offering better "alltogether" packages, but that's a different story, to be read in many other threads ;).
-
I write here on another post, my DS1054Z is +250MHz after upgrade with key (maybe I have luck, it is 4 year old). I think that any will be +140MHz and work very very good :)
-
I write here on another post, my DS1054Z is +250MHz after upgrade with key (maybe I have luck, it is 4 year old).
FAKE NEWS. Please don't spread this BS without proving.
-
I write here on another post, my DS1054Z is +250MHz after upgrade with key (maybe I have luck, it is 4 year old).
FAKE NEWS. Please don't spread this BS without proving.
@tv84 is correct. 250MHz 3dB bandwidth is BS. You will see "something" @ 250MHz if only a single channel is active (1GSa/s) but for a "real" measurement, forget it. Here's some information (https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg2592723/#msg2592723) that I could verify on my DS1000Z.
For general use, assume the 3dB point of the hacked DS1000Z to be at 135MHz (no more than two channels active, i.e. 500MSa/s). With all channels active, it's very advisavle to limit the highest frequency component of the input signal to 100MHz, otherwise you may end up with a lot of aliasing products.
Still, the four channel mode is pretty useful. Consider analyzing SPI protocols or even communicating with an I2C A/D or D/A converter and observing the serieal input contemporary with the analog output...
-
Yes I measure with 1ch mode (99% use), and my instrument have rise time of 1.35 to 1.4ns.
Also read this (it is not my rigol)
https://hackaday.com/2016/10/05/choosing-a-scope-examining-bandwidth/
-
I'd go for four channels. You'll find uses for them if you have them.
-
Yes I measure with 1ch mode (99% use), and my instrument have rise time of 1.35 to 1.4ns.
Also read this (it is not my rigol)
https://hackaday.com/2016/10/05/choosing-a-scope-examining-bandwidth/
That magic 0.35 formula is valid only for analog oscilloscopes. The formula does not stand for digital oscilloscopes with much sharper DSP filters. Also, what is shown on a digital oscilloscope is not the real signal, but a reconstruction of it after applying lots and lots of DSP. Those 1.4ns are not the real raise time of the oscilloscope's amplifier.
DS1054Z is NOT a 250MHz analog bandwidth oscilloscope.
To answer the OP question, I'll buy the 4 channels one.
-
Yes I measure with 1ch mode (99% use), and my instrument have rise time of 1.35 to 1.4ns.
Also read this (it is not my rigol)
https://hackaday.com/2016/10/05/choosing-a-scope-examining-bandwidth/
That is incorrect. You cannot estimate bandwidth like that, without knowing input characteristics.
Proper way is to make frequency sweep with leveled signal generator, which was done and cca 135 MHz is right figure.
-
- Analog bandwidth is the frequency at which the measured amplitude is 3 dB below the actual amplitude of the signal
- Rise time is the time required for a pulse to rise from 10 per cent to 90 per cent of its steady value
I remember that I measure only rise time, and it is about 1.4ns. Later I will check BW with signal generator, to be 100% like in definition :D
-
That magic 0.35 formula is valid only for analog oscilloscopes. The formula does not stand for digital oscilloscopes with much sharper DSP filters. Also, what is shown on a digital oscilloscope is not the real signal, but a reconstruction of it after applying lots and lots of DSP. Those 1.4ns are not the real raise time of the oscilloscope's amplifier.
by the way for much sharper DSP filter with same rise time it will be more not less BW
https://www.tek.com/support/faqs/how-bandwidth-related-rise-time-oscilloscopes (https://www.tek.com/support/faqs/how-bandwidth-related-rise-time-oscilloscopes)
-
A picture is worth a thousand words :D
-
Trigger is not very good after 150MHz, you will see phase noise. Also the counter is not working properly.
-
A picture is worth a thousand words :D
:palm: |O
The bandwidth of an oscilloscope is determined by several factors, mainly 3, if it does not meet the three parameters well, the result is that you are not seeing the correct signal.
Factors:
1 - Signal amplitude, when the signal falls 3db is the limit.
2 - Samples, you need at least 10 samples per cycle to be able to have a "real" signal. Really more, but that would be the minimum.
3 - Rise up and rise down time, you need it to be fast enough or it will distort the signal.
If only one of those factors fails, you will not see the real signal and it will not work correctly
other important factors are:
1 - The amount of WFMS / S that you are able to see.
2 - The "treatment" (falsified) that the oscilloscope does when interpreting the captured points and converting them into vectors to display a reconstructed signal. Rigol DS1054z in this falsifies more than other oscilloscopes
If you don't believe me, do an experiment:
First...
Put a 100Mhz SINE signal and capture it on the oscilloscope.
Now...
Put a TRIANGULAR signal of 100Mhz and you capture it, tell me if you can differentiate them?
Let's not say 200MHZ and much less 300MHZ
-
A picture is worth a thousand words :D
:palm: |O
The bandwidth of an oscilloscope is determined by several factors, mainly 3, if it does not meet the three parameters well, the result is that you are not seeing the correct signal.
Factors:
1 - Signal amplitude, when the signal falls 3db is the limit.
2 - Samples, you need at least 10 samples per cycle to be able to have a "real" signal. Really more, but that would be the minimum.
3 - Rise up and rise down time, you need it to be fast enough or it will distort the signal.
If only one of those factors fails, you will not see the real signal and it will not work correctly
other important factors are:
1 - The amount of WFMS / S that you are able to see.
2 - The "treatment" (falsified) that the oscilloscope does when interpreting the captured points and converting them into vectors to display a reconstructed signal. Rigol DS1054z in this falsifies more than other oscilloscopes
If you don't believe me, do an experiment:
First...
Put a 100Mhz SINE signal and capture it on the oscilloscope.
Now...
Put a TRIANGULAR signal of 100Mhz and you capture it, tell me if you can differentiate them?
Let's not say 200MHZ and much less 300MHZ
Er, no...
It is not that simple, yet it is simpler than that.
1. Yes, -3db point is considered bandwidth. That is correct.
2. You need more than 2.5 samples per period of highest frequency sinewave component your input filtering will pass into A/D converter.
3. Rise/fall times and bandwidth are function of frequency/phase characteristics of combined input attenuators/amplifiers/antialiasing filters all the way up to A/D converter(including A/D converter inputs).
4. Waveforms/seconds (triggers per second really) are important because of short retrigger time, so we can reliably trigger on quick short bursts (like in bursts of data packets on CAN bus for instance) without loosing data in a sequence. For people working on slowly moving or simple repetitive signals not that important. But generally speaking shorter retrigger time is always better.
But, make note that is one of the specifications ALL manufacturers (even the A brand) are very "creative" with, and numbers will be applicable to very specific situations and timebases . Most of the time you will NOT get specced wfms/s. I really wish that ALL manufacturers would simply publish table off retrigger times per time/div because that is all you need, really.
5. The "treatment" is called signal reconstruction. It is important part of A/D conversion and D/A conversion of signal back to analog domain to display it on screen.
Realy, and pay attention here, everybody thinks scope does A/D conversion and than creates data dots to display on screen.
It doesn't. It does A/D conversion, and then does D/A conversion of data back to analog domain. Except it doesn't convert it back to electric signal domain, but it converts it back to graphical, visual domain.
But to recreate signal as was seen by A/D converter (on it's input), you have to low pass filter output from D/A converter too. On a scope D/A converter/output filter is it's graphical plotting engine....
Scope is a band limited system. If you put a 100MHz squarewave or triangle wave into perfect 100 MHz oscilloscope you have to see something resembling sinewave on a screen in both cases.
Because scope have to filter out anything above 100MHz. 100 MHz scope is a 100MHz lowpass filter with screen...
And since any signal that has a repeating period of 100MHz that has any shape different than perfect sinewave, is a 100MHz sinewave with added harmonics at 2,3,4, 5, 6... times it's fundamental frequency (100MHz in this case), scope will filter those out and only show 100 MHz sinewave. Depending of spectral content of signal, you need scope with 5, 6, ...10.. times more bandwidth than signals repetition rate (fundamental frequency) to accurately show signal shape... 100 MHz triangle wave has significant harmonic content at 300, 500 and 700 MHz.... So you would need a 500Mhz scope to nicely show a 100 MHz triangle wave.
And that applies to analog scopes too. They are also band limited systems...
-
A picture is worth a thousand words :D
(https://www.eevblog.com/forum/testgear/rigol-ds1054-or-ds1202/?action=dlattach;attach=1042780;image)
(https://www.eevblog.com/forum/testgear/rigol-ds1054-or-ds1202/?action=dlattach;attach=1042784;image)
(https://www.eevblog.com/forum/testgear/rigol-ds1054-or-ds1202/?action=dlattach;attach=1042788;image)
The amplitude response looks pretty good. That property alone can serve well for some particular types of measurements, for example to look at an AM carrier, to measure modulation index, etc. However, the amplitude response, alone, is not enough for an oscilloscope.
An oscilloscope is supposed to correctly show the shape of an arbitrary signal (composed of many spectral components, each with its own phase shift and own delays). Preserving the phase of each spectral component is as important as the amplitude, in order to display the correct shape of an arbitrary waveform. If the amplitude is correct for each frequency, but the phase is not, then the displayed shape will be very different from the original. Same with the group delay vs frequency.
Would be interesting to measure phase shift vs frequency, too, and also the group delay vs frequency.
-
Definitely go 4 channels DS1054Z. It’s a massive advantage having more channels, particularly when working on serial protocols and more complex analogue systems and you’re already getting a bandwidth boost from your analogue scope.
With respect to bandwidth, if you’re doing HF ham stuff, then it’s fine for signal tracing and output monitoring. But really I’d probably reach for a NanoVNA more for radio stuff.
I would, as most of these infernal threads, ignore the semi-academic flame war that inevitably interlaces with the meat of the content.
If you want to save yourself some cash just buy another second hand analogue scope or fix yours :)
-
If you want to buy perfect scope, sell Lamborghini, and buy maybe 20 high/middle/lowend (no one can have all with ideal specification). But for <1000$ scope all was good. I don't work for Rigol, and also buy this only to have one more toy in home :D. Also Analog is best buy todays, maybe to have one analog (100/200MHz old tek,HP is very cheap) with some digital.
Also before many years oscilloscopes have 100MHz with 100Msps, and 40GHZ with 200ksps, and no one says that is impossible. I know, we are want real time, 100Tsps for 100$, but maybe in our experiment 1Msps is max that we need :D
-
2 - Samples, you need at least 10 samples per cycle to be able to have a "real" signal. Really more, but that would be the minimum.
Not true.
-
Er, yes but no.
1 - with only 2.5 samples per wave period you will see anything but not the actual wave.
It is a "convention" of the manufacturers but it is still false.
look here:
http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html (http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html)
(http://support.ircam.fr/docs/AudioSculpt/3.0/res/aliasing.png)
2 - Wfms / s are very important to see glitches or other non-periodic wave failures
3 - The signal treatment carried out by the oscilloscope is important and normally falsifies the reconstruction looking for a sinusoidal finish (Rigol does it among many others)
https://circuitglobe.com/digital-storage-oscilloscope.html (https://circuitglobe.com/digital-storage-oscilloscope.html)
(https://circuitglobe.com/wp-content/uploads/2017/06/interpolation-of-oscilloscopes.jpg)
4 - Digital oscilloscopes in general are rubbish because it is true that you need an oscilloscope with a frequency almost 10 times higher to see the real wave, this is so because the manufacturers considered that valid and as a standard, but it is rubbish that we have accepted.
Normally, 1GS ADC oscilloscopes are not able to distinguish a square signal from a sine wave beyond 25 or 30 Mhz, you can do the test yourself.
5 - It is false that analog oscilloscopes work the same, I have an analog one and taking it to the maximum bandwidth it is perfectly capable of differentiating between square, triangular or sinusoidal waves, and even seeing glitches.
Therefore I know what there is, I accept it reluctantly, but I know that a digital oscilloscope with a 1GS ADC does not see the real signal beyond about 30Mhz (using a single channel), that if they contribute many other things.
Therefore putting a photo of a Rigol ds1054Z with a signal of 100MHZ, 250MHZ and 300MHZ makes me laugh, you are seeing anything, but not reality.
I have a Siglent SDS1104X-E with 2 1GS ADCs and the same thing happens to it, beyond about 30MHz you do not distinguish square, sinusoidal or triangular signal, I see the fantasy that the oscilloscope wants to show, but not reality
-
I have a Siglent SDS1104X-E with 2 1GS ADCs and the same thing happens to it, beyond about 30MHz you do not distinguish square, sinusoidal or triangular signal, I see the fantasy that the oscilloscope wants to show, but not reality
If I were you I'd sell it and buy a Rigol MSO5000.
-
I have a Siglent SDS1104X-E with 2 1GS ADCs and the same thing happens to it, beyond about 30MHz you do not distinguish square, sinusoidal or triangular signal, I see the fantasy that the oscilloscope wants to show, but not reality
If I were you I'd sell it and buy a Rigol MSO5000.
I can't buy a Rigol MSO5000, it is too expensive for me, but I have no stupid illusions putting a picture of a DS1054Z at 300MHZ saying "measure it"
-
A picture is worth a thousand words :D
Unless your scope is fundamentally different from all the other DS1000Z known to me, these photos are fake. What is your vertical multiplier? My DS1000Z switches ranges between 330 and 335 mV/div (at 1x multiplier). At 335mV/div and above, I can duplicate your findings, below, the 3dB bandwidth is about 135MHz. See the link to the DS1000Z main thread (https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg2592723/#msg2592723) from my previous contribution.
-
I have a Siglent SDS1104X-E with 2 1GS ADCs and the same thing happens to it, beyond about 30MHz you do not distinguish square, sinusoidal or triangular signal, I see the fantasy that the oscilloscope wants to show, but not reality
If I were you I'd sell it and buy a Rigol MSO5000.
I can't buy a Rigol MSO5000, it is too expensive for me, but I have no stupid illusions putting a picture of a DS1054Z at 300MHZ saying "measure it"
What you can do is the same as what hacked Rigol owners ought to be doing: When a measurement is life or death critical you should reduce the number of channels to get maximum possible sample rate on the channel of interest.
I suspect it won't make much difference in reality though, getting a 200+MHz signal all the way to the ADC without distortion is difficult.
-
Er, yes but no.
1 - with only 2.5 samples per wave period you will see anything but not the actual wave.
It is a "convention" of the manufacturers but it is still false.
look here:
http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html (http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html)
(http://support.ircam.fr/docs/AudioSculpt/3.0/res/aliasing.png)
2 - Wfms / s are very important to see glitches or other non-periodic wave failures
3 - The signal treatment carried out by the oscilloscope is important and normally falsifies the reconstruction looking for a sinusoidal finish (Rigol does it among many others)
https://circuitglobe.com/digital-storage-oscilloscope.html (https://circuitglobe.com/digital-storage-oscilloscope.html)
(https://circuitglobe.com/wp-content/uploads/2017/06/interpolation-of-oscilloscopes.jpg)
4 - Digital oscilloscopes in general are rubbish because it is true that you need an oscilloscope with a frequency almost 10 times higher to see the real wave, this is so because the manufacturers considered that valid and as a standard, but it is rubbish that we have accepted.
Normally, 1GS ADC oscilloscopes are not able to distinguish a square signal from a sine wave beyond 25 or 30 Mhz, you can do the test yourself.
5 - It is false that analog oscilloscopes work the same, I have an analog one and taking it to the maximum bandwidth it is perfectly capable of differentiating between square, triangular or sinusoidal waves, and even seeing glitches.
Therefore I know what there is, I accept it reluctantly, but I know that a digital oscilloscope with a 1GS ADC does not see the real signal beyond about 30Mhz (using a single channel), that if they contribute many other things.
Therefore putting a photo of a Rigol ds1054Z with a signal of 100MHZ, 250MHZ and 300MHZ makes me laugh, you are seeing anything, but not reality.
I have a Siglent SDS1104X-E with 2 1GS ADCs and the same thing happens to it, beyond about 30MHz you do not distinguish square, sinusoidal or triangular signal, I see the fantasy that the oscilloscope wants to show, but not reality
You have no clue do you?
-
Unless your scope is fundamentally different from all the other DS1000Z known to me, these photos are fake.
I will add: "and known to Rigol".
Rigol's meeting room:
"- Well we have here are good old DS1054Z capable of 300Mhz that we have been selling up to 100. Shall we rebrand the sticker or shall we devise new ASICs and create a new line of +300MHz scopes?
- I don't know, that's a hard one... :-\"
-
Er, yes but no.
1 - with only 2.5 samples per wave period you will see anything but not the actual wave.
It is a "convention" of the manufacturers but it is still false.
look here:
http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html (http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html)
(http://support.ircam.fr/docs/AudioSculpt/3.0/res/aliasing.png)
2 - Wfms / s are very important to see glitches or other non-periodic wave failures
3 - The signal treatment carried out by the oscilloscope is important and normally falsifies the reconstruction looking for a sinusoidal finish (Rigol does it among many others)
https://circuitglobe.com/digital-storage-oscilloscope.html (https://circuitglobe.com/digital-storage-oscilloscope.html)
(https://circuitglobe.com/wp-content/uploads/2017/06/interpolation-of-oscilloscopes.jpg)
4 - Digital oscilloscopes in general are rubbish because it is true that you need an oscilloscope with a frequency almost 10 times higher to see the real wave, this is so because the manufacturers considered that valid and as a standard, but it is rubbish that we have accepted.
Normally, 1GS ADC oscilloscopes are not able to distinguish a square signal from a sine wave beyond 25 or 30 Mhz, you can do the test yourself.
5 - It is false that analog oscilloscopes work the same, I have an analog one and taking it to the maximum bandwidth it is perfectly capable of differentiating between square, triangular or sinusoidal waves, and even seeing glitches.
Therefore I know what there is, I accept it reluctantly, but I know that a digital oscilloscope with a 1GS ADC does not see the real signal beyond about 30Mhz (using a single channel), that if they contribute many other things.
Therefore putting a photo of a Rigol ds1054Z with a signal of 100MHZ, 250MHZ and 300MHZ makes me laugh, you are seeing anything, but not reality.
I have a Siglent SDS1104X-E with 2 1GS ADCs and the same thing happens to it, beyond about 30MHz you do not distinguish square, sinusoidal or triangular signal, I see the fantasy that the oscilloscope wants to show, but not reality
You have no clue do you?
oh you hunted me
It is true that an oscilloscope with a 1GS, 100Mhz and 4 channel ADC can perfectly see a glitch of a square signal at 300MHz, or will it be sinusoidal? maybe triangular? could it perhaps be modulated?
What does it matter, surely the wave signal generator is broken, whatever I put, I see it sinusoidal and perfect.
-
oh you hunted me
It is true that an oscilloscope with a 1GS, 100Mhz and 4 channel ADC can perfectly see a glitch of a square signal at 300MHz, or will it be sinusoidal? maybe triangular? could it perhaps be modulated?
What does it matter, surely the wave signal generator is broken, whatever I put, I see it sinusoidal and perfect.
Look up Dunning–Kruger effect.. It applies to all waveforms ...
-
Er, yes but no.
1 - with only 2.5 samples per wave period you will see anything but not the actual wave.
look here:
http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html (http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html)
(http://support.ircam.fr/docs/AudioSculpt/3.0/res/aliasing.png)
Ummm... yeah. That's less than 1 sample per wave.
-
I really need to work out how the unsubscribe works on this forum.
Every time these threads descend into people slapping their dicks against each other about signal integrity. The only truism seems to be that no one understands it.
-
Er, yes but no.
1 - with only 2.5 samples per wave period you will see anything but not the actual wave.
look here:
http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html (http://support.ircam.fr/docs/AudioSculpt/3.0/co/Sampling_1.html)
(http://support.ircam.fr/docs/AudioSculpt/3.0/res/aliasing.png)
Ummm... yeah. That's less than 1 sample per wave.
It has been an exaggerated example to understand what we are talking about, I have taken it from the internet, I have not done it with the exact measurements.
I hope you understand
-
I really need to work out how the unsubscribe works on this forum.
Every time these threads descend into people slapping their dicks against each other about signal integrity. The only truism seems to be that no one understands it.
jajajaja
ok finish for me
-
I really need to work out how the unsubscribe works on this forum.
Every time these threads descend into people slapping their dicks against each other about signal integrity. The only truism seems to be that no one understands it.
If you find out how, please let us know..
-
It has been an exaggerated example to understand what we are talking about, I have taken it from the internet, I have not done it with the exact measurements.
I hope you understand
[/quote]
I understand that it agrees with the math exactly, ie. The math works.
The math says: 2.5 samples per wave is enough to avoid the AM effects that start to appear as you approach Nyquist.
-
It has been an exaggerated example to understand what we are talking about, I have taken it from the internet, I have not done it with the exact measurements.
I hope you understand
I understand that it agrees with the math exactly, ie. The math works.
The math says: 2.5 samples per wave is enough to avoid the AM effects that start to appear as you approach Nyquist.
[/quote]
(https://www.picotech.com/images/uploads/library/_med/oscilloscope_tutorial_3.png)
Do you want to explain to me how you can know what wave it is, for example "a ladder pattern" with 2.5 samples per wave, like the one in the photo? It is impossible.
But for me I end the issue.
You want to be right because yours is
-
That is a video signal. :)
-
(https://www.picotech.com/images/uploads/library/_med/oscilloscope_tutorial_3.png)
Do you want to explain to me how you can know what wave it is, for example "a ladder pattern" with 2.5 samples per wave, like the one in the photo? It is impossible.
The square wave and triangle wave must have gone to the same place as the sine wave.
-
Every time these threads descend into people slapping their dicks against each other about signal integrity. The only truism seems to be that no one understands it.
jajajaja
ok finish for me
It's clear that some people understand it a lot more than others. :horse:
-
Do you want to explain to me how you can know what wave it is, for example "a ladder pattern" with 2.5 samples per wave, like the one in the photo? It is impossible.
IMO "2.5 samples per wave" should better be phrased "2.5 samples per period of the highest frequency component contained in the signal" (-> according to a fourier analysis of the signal, which decomposes the signal into a sum of sine waves). Theory tells that the sampling rate must be more than 2x the highest frequency component; with 2.5 there is already some headroom beyond the theoretical minimum, in order to deal with practical limitations of the reconstruction.
Regarding the screen shot:
Standard-compliant PAL composite video signals (as broadcasted) have a video bandwidth of about 6 MHz, so 6 MHz is the highest possible frequency component. (For a synthetic, non-standard video signal, this may not apply, though.)
Ideal triangle or square wave signals have an infinite bandwidht (due to an infinite number of harmonics), i.e. they cannot be reconstructed exactly at all from samples captured at any finite sampling rate. So the practical goal can only be a "good enough" reconstuction up to the first N harmonics. And how good is "good enough" can be quite subjective, depending on the particular use case.
-
One formulation of Nyquist's theorem is that given an infinitely long sequence of samples made at uniform time intervals, there is one and only one curve that can be drawn which passes through every sample point and has a bandwidth no more than half the reciprocal of the sample interval. A good DSO will display this curve, or a close approximation to it (*). Whether this curve is or is not the signal from which the DSO originally measured the samples is a matter for the skill and judgement of the user.
(A finite-length set of samples, such as a DSO record, is considered to repeat indefinitely, for the purposes of the theorem.)
(*) unless told not to
-
Definitely go 4 channels DS1054Z. It’s a massive advantage having more channels, particularly when working on serial protocols and more complex analogue systems and you’re already getting a bandwidth boost from your analogue scope.
With respect to bandwidth, if you’re doing HF ham stuff, then it’s fine for signal tracing and output monitoring. But really I’d probably reach for a NanoVNA more for radio stuff.
I would, as most of these infernal threads, ignore the semi-academic flame war that inevitably interlaces with the meat of the content.
If you want to save yourself some cash just buy another second hand analogue scope or fix yours :)
Interrupting the usual bantering about signal integrity and jumping back to the original question for a moment
Personally, for digital protocols, I find a protocol analyzer (even a low cost 8 channels Saleae clone) a much better tool than an oscilloscope with more channels. Saleae clones cost <$10 and DSlogic clones cost ~$60. And they support more than a hundred of protocols, including older 8 bit processors that the OP seems to work on, vs the half a dozen protocols supported by oscilloscopes. And usually you need to analyze a long data stream, for which you need a PC anyway
I work mostly with digital systems and microprocessors. Given the choice of a 4Ch schope only or a 2Ch scope+logic analyzer, I'd choose the latter every time
Complex analog systems are clearly a different matter
-
Luckily for us the Oscilloscope manufacturers make input circuits that naturally low-pass filter the signal and help it stay within the Nyquist limit.
Many of them even measure the -3dB point of their input circuits and print it on the front of the device. :-+
-
I really need to work out how the unsubscribe works on this forum.
Every time these threads descend into people slapping their dicks against each other about signal integrity. The only truism seems to be that no one understands it.
(https://media1.tenor.com/images/4b763468b97504a3d00aae06b5aabab4/tenor.gif?itemid=14271597)
Finish
-
snip
Why don't you tell us the frequency of the signal in this image?
(https://www.picotech.com/images/uploads/library/_med/oscilloscope_tutorial_3.png)
-
Definitely go 4 channels DS1054Z. It’s a massive advantage having more channels, particularly when working on serial protocols and more complex analogue systems and you’re already getting a bandwidth boost from your analogue scope.
With respect to bandwidth, if you’re doing HF ham stuff, then it’s fine for signal tracing and output monitoring. But really I’d probably reach for a NanoVNA more for radio stuff.
I would, as most of these infernal threads, ignore the semi-academic flame war that inevitably interlaces with the meat of the content.
If you want to save yourself some cash just buy another second hand analogue scope or fix yours :)
Interrupting the usual bantering about signal integrity and jumping back to the original question for a moment
Personally, for digital protocols, I find a protocol analyzer (even a low cost 8 channels Saleae clone) a much better tool than an oscilloscope with more channels. Saleae clones cost <$10 and DSlogic clones cost ~$60. And they support more than a hundred of protocols, including older 8 bit processors that the OP seems to work on, vs the half a dozen protocols supported by oscilloscopes. And usually you need to analyze a long data stream, for which you need a PC anyway
I work mostly with digital systems and microprocessors. Given the choice of a 4Ch schope only or a 2Ch scope+logic analyzer, I'd choose the latter every time
Complex analog systems are clearly a different matter
Yeah that's exactly what I use. Got an aliexpress clone. Does the job.
-
Thanks all, got myself a 1054z
-
Can someone explain how the Bandwidth hack works on DS1054Z? I recently bought 1054Z from distributor; it has all features enabled so really no need to hack to decode I2C.
I looked at rise and fall times - the scope reports ~5ns; however, the rise of 70+% of the transient is ~2ns, and then there is an a round corner of the transition which lets the scope think that the rise time is longer. So the approximate bandwidth is ~70MHz for 0.5V signal which meets the specs. And the minimum timebase setting is 5ns.
Now I am wondering what the "bandwidth hack" is doing - looking at the schematics of the frontend found on Internet (attached), there are no software configurable/programmable components on it. Is there any digital filtering to artificially reduce the bandwidth?
Does the hack increase the resolution of the timebase?
The front end is very similar to the legacy Tektronix FET front-ends - with the high speed FET AC amplifier and the OpAmp-based DC restoration channel in parallel. So it may achieve something around 100-150MHz, but probably not more. It is not the fully configurable integrated front end found on DS2xxx series of Rigol scopes. And, because the scope uses "logic" trigger, the trigger is in fact the condition in the sequence of the samples, and there is no separate trigger channel as in other scopes...
Is there any sense in increasing the bandwidth - the scope samples at 250Msa/s in 4-channel mode, and reasonable bandwidth to avoid aliasing is around 70MHz... Any comments?
-
I think you are overthinking it. Even if it were no difference, with the next firmware update might be, who knows.
Just unlock everything, that's what I did from the day one and it all worked great, including the 500uV/div.
-
Is it really bandwidth? First impressions on different settings of Rigol DS1054Z - apparently trigger/acquisition system is made in attempt to randomize the sampling time and save the location of waveform relative to virtual trigger. This means that the signal you see is not the signal where your probes are attached.
First: Vector display is applying SinX/X interpolation to samples. But it is valid only when the ADC is preceded by a true rectangular filter in front - which is not true. Look at dots display and vector display of the same step fig 1 [attach=1] and 2 [attach=2]. Now, when multiple channels are enabled, apparently there is a bug or anomaly on this filter - it messes with the samples (fig 3) [attach=3] - you see ringing and jitter which is not present in the source.
Second: Ok, let us disable interpolator and see the samples as dots. One can get very interesting effects - like time is going backwards between ADC samples (!) [attach=6]. And the rise/fall depends on the position of the trigger relative to the signal step (fig 4-6)[attach=4] [attach=5]. I am very surprised...
Anyway, this is the extreme measurement case. The scope is OK to measure signals below 10-20MHz which is its primary purpose of low-speed, student oscilloscope. Comments welcome.
-
Is it really bandwidth? First impressions on different settings of Rigol DS1054Z - apparently trigger/acquisition system is made in attempt to randomize the sampling time and save the location of waveform relative to virtual trigger. This means that the signal you see is not the signal where your probes are attached.
First: Vector display is applying SinX/X interpolation to samples. But it is valid only when the ADC is preceded by a true rectangular filter in front - which is not true. Look at dots display and vector display of the same step fig 1 (Attachment Link) and 2 (Attachment Link) . Now, when multiple channels are enabled, apparently there is a bug or anomaly on this filter - it messes with the samples (fig 3) (Attachment Link) - you see ringing and jitter which is not present in the source.
Second: Ok, let us disable interpolator and see the samples as dots. One can get very interesting effects - like time is going backwards between ADC samples (!) (Attachment Link) . And the rise/fall depends on the position of the trigger relative to the signal step (fig 4-6) (Attachment Link) (Attachment Link) . I am very surprised...
Anyway, this is the extreme measurement case. The scope is OK to measure signals below 10-20MHz which is its primary purpose of low-speed, student oscilloscope. Comments welcome.
This all was explained and proven to death, both theory, math and by measurement numerous times ago. Google "site:eevblog.com search phrase here"
In short: it has digitally controlled filter. Bandwidth was measured at some 120-130 MHz (-3dB). Which makes is barely Nyquist for 4 ch. Search for Gibbs phenomenon. Higher bandwidth is useful, even if you need to use it with only 2 ch. The scope is OK to use for any purpose, including professional, as long as operator knows what he/she is doing. It is good to measure up to it's full bandwidth, that is why bandwidth is there. It is simple, entry level scope. If you want more, pay more.. Simple as that.
This is all I am going to contribute here. I'm really fed up with :horse: stuff that keep resurfacing because people can't be bothered to read.. Just search for it, it's all there... If you want to learn, that should be your effort, not ours..
-
Hello,
I got my DS1054 several weeks ago direkt from RIGOL.
The measurement options were already released except 100MHz.
With the well known way, the bandwidth is now 100MHz.
About the 500 uV I read a lot of positiv and negatives in the net.
Because nothing can be destroyed and it can be also uninstalled if it is necessary,
I try it.....and recalibrate it.
[attach=1]
And it works. Attached you find a screen picture of a 145MHz signal with 500uV vertical timebase.
---
www.dg8fbv.de (http://www.dg8fbv.de)
-
Hello,
I got my DS1054 several weeks ago direkt from RIGOL.
The measurement options were already released except 100MHz.
With the well known way, the bandwidth is now 100MHz.
About the 500 uV I read a lot of positiv and negatives in the net.
Because nothing can be destroyed and it can be also uninstalled if it is necessary,
I try it.....and recalibrate it.
(Attachment Link)
And it works. Attached you find a screen picture of a 145MHz signal with 500uV vertical timebase.
---
www.dg8fbv.de (http://www.dg8fbv.de)
Btw...that is a picture of a 435 MHz Signal 500 uV
[attach=1]
-
Is it really bandwidth?
With 1 or 2 channels on? Yes.
With 4 channels on? Not so much - sample rate is on the very limits.
If this isn't good enough for you then feel free to pay for more than bottom-end equipment.
This all was explained and proven to death, both theory, math and by measurement numerous times ago.
Yep.
(nb. All this applies to low-end Siglents, too)
-
Hello,
[…..]
About the 500 uV I read a lot of positiv and negatives in the net.
Because nothing can be destroyed and it can be also uninstalled if it is necessary,
I try it.....and recalibrate it.
(Attachment Link)
And it works. Attached you find a screen picture of a 145MHz signal with 500uV vertical timebase.
---
www.dg8fbv.de (http://www.dg8fbv.de)
Sorry I forgot to mention my system info….