Does anybody know if the DS1104Z supports ETS (Equivalent Time Sampling)?
No, none of the Rigol UltraVision DSOs do.
I disagree with the simplicity of this answer.
They have something functionally equivalent.
On a DPO style oscilloscope using only digital triggering...
...the reconstruction happens before triggering which yields a comparable trigger to clock delay measurement. If this did not happen, then timing measurements would be limited to the base sample rate which even at 1 GS/s would only be 1 nanosecond which is pretty poor.
On a DPO style oscilloscope using only digital triggering...But who said the Rigol UltraVision scopes are only using digital triggering? It's not specified in any Rigol datasheets.
Quote...the reconstruction happens before triggering which yields a comparable trigger to clock delay measurement. If this did not happen, then timing measurements would be limited to the base sample rate which even at 1 GS/s would only be 1 nanosecond which is pretty poor.Why don't you look at a DPO that is actually advertised as using a digital trigger? Such as the Siglent SDS2000 - a 2GSa/s DSO (just like the Rigol DS2000) - that, according to it's datasheet, has a trigger timing and resolution of precisely 1ns.
Normally, when people talk about ETS, they're talking about the ability to sample at higher than RT rates. Even if what you're describing is functionally equivalent inside the DSO, it's operationally completely different - and thus irrelevant - to the end user of the DSO.
Mark_O: can't one just turn off the other channels to check how the signal looks in one channel mode if you get a feeling something might be off? Then if you have confirmed it, you can safely turn on the rest of the channels knowing that what you see is the accual signal. Of course, this assmums that one is triggering off the channel in question.
Would there be any accual problems going about it this way?
P.S. I know that that would be a hassel and I'm not saying it's convenient. Nor am I saying I understand this stuff all that well, it's just an idea I got while reading the previous comments.
...the reconstruction happens before triggering which yields a comparable trigger to clock delay measurement. If this did not happen, then timing measurements would be limited to the base sample rate which even at 1 GS/s would only be 1 nanosecond which is pretty poor.
And then I will look at the early LeCroy DSOs which were advertised as having digital triggers and find that they had timing resolution significantly higher than their real time sample rate would suggest.
It is relevant to an end user who expects ETS like performance out of his DSO which says nothing about ETS in its documentation or marketing.
Are you suggesting these Rigol oscilloscopes are crippled compared to the obsolete ETS oscilloscopes they replaced?
Capture a pair of synchronous high frequency sine waves in single shot mode and measure the resolution of the delay between them. Or do the same with one or more fast transition edges which displays pattern sensitive jitter or controlled delay. Can these oscilloscopes make these measurements?
...the reconstruction happens before triggering which yields a comparable trigger to clock delay measurement. If this did not happen, then timing measurements would be limited to the base sample rate which even at 1 GS/s would only be 1 nanosecond which is pretty poor.
In depends on context whether a 1 ns timing measurement is poor or not.
QuoteAnd then I will look at the early LeCroy DSOs which were advertised as having digital triggers and find that they had timing resolution significantly higher than their real time sample rate would suggest.
That's true, but so what? I have a 9300-series LeCroy, and two 9400-series. And you are correct about their timing resolution/capabilities. But they all had ETS (well, RIS), so they got that for "free", because they had a clock (or facsimile thereof) that ran 40x-50x faster. I see interpolation capabilities in the ps range. Back about 50 years ago, when I was using LeCroy scopes in the Physics labs at the Uni, picosecond events were extremely important. But the current "affordable" scopes we're talking about were never intended for that purpose.
QuoteIt is relevant to an end user who expects ETS like performance out of his DSO which says nothing about ETS in its documentation or marketing.
Why would anyone expect that? If it says nothing about some aspect of its performance, I have no expectations.
QuoteAre you suggesting these Rigol oscilloscopes are crippled compared to the obsolete ETS oscilloscopes they replaced?
Marmad may not, but I would. Though I wouldn't use the word crippled. They're simply more limited, in some ways. And I also don't understand why it would surprise you that scopes designed to sell for 50x less, would be less capable in some regards?
QuoteCapture a pair of synchronous high frequency sine waves in single shot mode and measure the resolution of the delay between them. Or do the same with one or more fast transition edges which displays pattern sensitive jitter or controlled delay. Can these oscilloscopes make these measurements?
No. Not really. Not effectively. The 1000z series doesn't specify, but the higher-performance 2000-series can have up to 2 ns of skew (nominally 1 ns) between the two channels. That's up to 4 clock periods at it's max sample rate. And there's enough jitter in their trigger systems (somewhere between 4-8 ns, IIRC) to make that look small by comparison.
The smallest resolution on any trigger-related setting in the 1000z specs is 8 ns, and for the 2000 series is 2 ns. Step increments are 4 ns* and 1 ns, respectively. What does that tell you?
*AFAIK. I hope Marmad (or anyone with those Rigols) will correct me, if that is incorrect about the step-size. The setting may be less, even if it can't really honor it. But the 1000z may have a 1 GHz clock, even when it's not using it to drive sampling directly.
Nowadays, SPI busses aren't just 2 MHz, or 5, or 8 MHz. I've got one here that's running at 60 MHz, and I wouldn't be surprised to hear about faster. So let's say he's looking at that signal on his embedded system. He's perplexed, because he can't get the darn thing to decode properly. It should work (a la, pascal), because it's "only" 60 MHz. But there are numerous problems that prevent it.
The Rigol user manual is explicit about supporting a timebase scale of 5 ns/div.
Without interpolation or reconstruction at 250 MS/s, that would produce a pretty awful looking display of a 3.5 nanosecond transition time signal (single-shot or not)
I wish someone would do the basic simple tests which would reveal how exactly these DSOs perform and then compile a wiki with the results. I would do it myself but lack the necessary hardware, namely the DSO.
The Rigol user manual is explicit about supporting a timebase scale of 5 ns/div.
Yes, given it's maximum sample rate of 1GSa/s.
QuoteWithout interpolation or reconstruction at 250 MS/s, that would produce a pretty awful looking display of a 3.5 nanosecond transition time signal (single-shot or not)
Who would think that the DSO could produce a decent looking 3.5ns rise time without interpolation when it's only sampling every 4ns? When looking at 5ns/div @ 250MSa/s (unless you're just examining the paltry 15 acquired sample points), the entire displayed waveform is nothing but interpolation. Expecting 'detail' at that time base and sample rate is fairly silly.
QuoteI wish someone would do the basic simple tests which would reveal how exactly these DSOs perform and then compile a wiki with the results. I would do it myself but lack the necessary hardware, namely the DSO.
Virtually ANYTHING can be bought and returned within 30 days if you're willing to absorb the shipping costs. Instead of posting new speculation every few days (a couple of weeks ago it was that the Rigol secretly turned off sin(x)/x interpolation at higher sample rates to hide interleaving errors), perhaps you should get one of the scopes, run the tests you want and post your results - returning the DSO afterwards. I would be curious to see your results, and no offense, but judging by your posting frequency here, you have the requisite free time.
Nowadays, SPI busses aren't just 2 MHz, or 5, or 8 MHz. I've got one here that's running at 60 MHz, and I wouldn't be surprised to hear about faster. So let's say he's looking at that signal on his embedded system. He's perplexed, because he can't get the darn thing to decode properly. It should work (a la, pascal), because it's "only" 60 MHz. But there are numerous problems that prevent it.
Yep. This is why discussing Nyquist limits is a mistake in relation to DSOs. The "ten to one" rule for sample rate vs. bandwidth isn't predicted by theory, it comes from experience working with real signals.
nb. Theory can justify the rule (hindsight vision is 20:20...)
As far as buying and then returning an item I never intended to keep, I consider that rather dishonest.
Great! So where in the manual does it say that the fastest timebase scale is slower when more channels are used? I must have missed that or maybe Rigol left it out.
I obviously do not expect any detail faster than the analog bandwidth or between sample points but I do expect sin(x)/x reconstruction to produce something closely resembling a 3.5 nanosecond transition. All of the necessary information baring aliasing is there.
I still suspect they took steps to hide interleaving and aliasing errors. Run the above test and find out.
As far as buying and then returning an item I never intended to keep, I consider that rather dishonest.
...a sample rate to-bandwidth ratio of 4:1 is sufficient to produce reliable digital measurements.
So while we would all love to have a 10:1 ratio or even better, I think it is going a bit far to say we must have 10:1. No doubt that the sample rate when using 4 channels on the 1000Z scopes are there downside, but the price is incredible.
One could argue that if you planned to return it from the get go that that might not be so ethical.
I'd certainly argue that. The seller will have a hard time selling it as "new" if it's had 30 days use (especially since the DS1054Z has trial features that tick away as you use it).
So while we would all love to have a 10:1 ratio or even better, I think it is going a bit far to say we must have 10:1. No doubt that the sample rate when using 4 channels on the 1000Z scopes are there downside, but the price is incredible.
Does anybody know if the DS1104Z supports ETS (Equivalent Time Sampling)?
No, none of the Rigol UltraVision DSOs do.
Thanks marmad. That's interesting and, to a certain extent, somewhat disappointing.
Yep. People are arguing over this as if it were a $4000 oscilloscope, not $400.
For $400 the only question you should be asking is, "Where can I get one???"
Yep. People are arguing over this as if it were a $4000 oscilloscope, not $400.
For $400 the only question you should be asking is, "Where can I get one???"
Agreed. That's why it is such a shame that the main thread dedicated to what is probably the best oscilloscope buy in quite a few years has been polluted with irrelevant and unhelpful noise. Perhaps the thread could be tidied up so it can add value long into the future ?
Yep. People are arguing over this as if it were a $4000 oscilloscope, not $400.
For $400 the only question you should be asking is, "Where can I get one???"
Agreed. That's why it is such a shame that the main thread dedicated to what is probably the best oscilloscope buy in quite a few years has been polluted with irrelevant and unhelpful noise. Perhaps the thread could be tidied up so it can add value long into the future ?
+1
Yeah, but where is my logic analyzer?