Oscilloscope recommendation DS1054Z / SDS1104X-E / GDS-1104B / etc?

[Performa01]

As for FRA, Keysight unveiled it in december of 2017 of their "low cost" 1000X series and to 3000T. Previous models did not have any.
Siglent unveiled their BODE analysis 2 months later. Meaning it was already fully developed and tested to production and was awaiting FW release scheduling... It was not reaction to Keysight...

Exactly. The idea of a Bode Plot application has been accepted by Siglent management back in September 2016, when there was only Vellemann with some cheap USB oscilloscope (and I don't know about the quality of their implementation) and Keysight had indeed one as part of their expensive PA-package, of course not available in any entry level scope.

The specification for Siglents first version of the FRA was written back in February 2017 and already specified a frequency selective detector, but this was something not yet implemented in the first version.

[nctnico]

We've used the Siglent Bode function often, and have also reported it's slow, real slow (we are not a fanboy)!! With this in mind it does give superb results as mentioned by 2N3055, slow yes, but accurate with a very wide dynamic range if utilized properly by someone that understands how the Bode function operates within the DSO.

If getting the correct, accurate answers is important, one can often overlook the speed!!

Now you make it sound like slow = accurate but that simply isn't true. The results from R&S's implementation (the one I have experience with but I don't see why others should be different) are just as accurate. So they deliver the same answer in less time. I have repeated several of the measurements you have shown using my R&S oscilloscope and got to the same results.

[2N3055]

We've used the Siglent Bode function often, and have also reported it's slow, real slow (we are not a fanboy)!! With this in mind it does give superb results as mentioned by 2N3055, slow yes, but accurate with a very wide dynamic range if utilized properly by someone that understands how the Bode function operates within the DSO.

If getting the correct, accurate answers is important, one can often overlook the speed!!

Now you make it sound like slow = accurate but that simply isn't true. The results from R&S's implementation (the one I have experience with but I don't see why others should be different) are just as accurate. So they deliver the same answer in less time. I have repeated several of the measurements you have shown using my R&S oscilloscope and got to the same results.

He never said slow was accurate.. He said (same as I did) it is dead accurate but somewhat slower then R&S and Keysight... And it is not seconds to hours difference...

And it is debatable if you got exactly the same result. You don't have a Siglent scope to really set it up the same... It looks OK though, and we said that.

Again, against what sub 1000 USD scope you are comparing Siglent SDS1104X-E bode plot feature/speed/accuracy set?
How many scopes (at any price range) can perform bode plot analysis of a 3-way active xover (low, mid, hi) at the same time on the same screen.
How long does it take to do that on Keysight or R&S..

I did check against the Keysight 3000T and got better dynamic range with Siglents I have..
Is that because I'm comparing a 500uV/div 12 bit scope against 8 bit scope with 4mV/div with more noise ??
That could be the case.... But the both SDS6000 and SDS2000X HD had better dynamic range than MSOX3104T...

For general use both will be good enough to gather -3dB point frequency, general shape of curves, etc..

There is also one more thing that were discussing. It is much easier to measure filters (passive and active) in quiet environment, than loop response of switching PSU with aggressor signals being orders or magnitude larger.
With switching PSU, ability of BODE plotter (whatever it is) to extract phase/gain of stimulus is very much a question of extracting very small signal from switcher natural output ripple.
In which case frequency selective detector and averagning helps a lot.. In which case slower is better....

So yeah, it is MUCH more complicated than: SLOW!!!! BAD!!!!!

[nctnico]

R&S is also using frequency selective detector. I did those tests as well. Anyway, Siglent could speed up their FRA a whole lot without compromising much so it would make sense for Siglent to implement fast (low dynamic range) and slow (high dynamic range) modes. That would be the proverbial cherry on top.

[2N3055]

R&S is also using frequency selective detector. I did those tests as well.

I defer to you for this. I have no way to verify it and no reason to doubt you.

Anyway, Siglent could speed up their FRA a whole lot without compromising much so it would make sense for Siglent to implement fast (low dynamic range) and slow (high dynamic range) modes. That would be the proverbial cherry on top.

That is what I said. If they can, that would be another win for Siglent. But if not, still OK.. I personally find it mildly annoying, not a dealbreaker..
If other scopes could do 2 sweeps per second than I would be also annoyed. But they cannot. I need to get a 5000 USD FRA to get order of magnitude faster sweeps.
This way, 45 seconds or 90 seconds, for both cases I take that as as moment to get up, stretch my back and go get some water or something.. By the time I'm back, result is there...

If I would do FRA every day, all day long, I would then go for BODE100 device, because not a single scope (even 20000USD+ ones) are as fast, accurate or have as good dynamic range for that purpose.

But, for occasional use, speed is not primary..  And again, Siglent has that capability, and better features than other manufacturers, for a very low entry price, available to hobby users and small business that cannot afford to pay 15X times more for 30 seconds less...

[mawyatt]

We've used the Siglent Bode function often, and have also reported it's slow, real slow (we are not a fanboy)!! With this in mind it does give superb results as mentioned by 2N3055, slow yes, but accurate with a very wide dynamic range if utilized properly by someone that understands how the Bode function operates within the DSO.

If getting the correct, accurate answers is important, one can often overlook the speed!!

Now you make it sound like slow = accurate but that simply isn't true. The results from R&S's implementation (the one I have experience with but I don't see why others should be different) are just as accurate. So they deliver the same answer in less time. I have repeated several of the measurements you have shown using my R&S oscilloscope and got to the same results.

No, no not saying that slow always equals accurate. Just saying that within a given measurement/equipment scenario that if one needs specific accurate results, often slow is acceptable. Suspect the R&S is faster and also accurate, we don't have one so can't say for sure, we don't comment on equipment we don't have and put very little faith in comments by others that don't have said equipment in hand.

Would like to see how well the R&S behaves with the Peltz Oscillator Injection Locking experiment mentioned (wish we had one to compare), would expect good results tho as from what we've gathered from various reviews, and the informative videos and responses by R&S Engineer pdenisowski, R&S MSO features seem well done.

Best,

[mawyatt]

There is also one more thing that were discussing. It is much easier to measure filters (passive and active) in quiet environment, than loop response of switching PSU with aggressor signals being orders or magnitude larger.
With switching PSU, ability of BODE plotter (whatever it is) to extract phase/gain of stimulus is very much a question of extracting very small signal from switcher natural output ripple.
In which case frequency selective detector and averagning helps a lot.. In which case slower is better....

So yeah, it is MUCH more complicated than: SLOW!!!! BAD!!!!!

Agree, this is a condition that really stresses the measurement with a high interfering signal hiding the wanted low level signal, kinda like what a radio receiver must deal with!! The Injection Locking experiment placed the interfering signal (oscillator output) quite close to the tiny wanted Bode signal, and thus our comments

Best,

[Fungus]

It is actually rather interesting that none of the Siglent fan crowd has ever commented on the slowness of Siglent's FRA. Either they didn't want to tell or just don't know any better. Rudi's comparison where it comes to the time needed to do a sweep came as a total -shocking- surprise to me!

Oh, now you've done it ... you said something bad about a Siglent! 

[mawyatt]

It is actually rather interesting that none of the Siglent fan crowd has ever commented on the slowness of Siglent's FRA. Either they didn't want to tell or just don't know any better. Rudi's comparison where it comes to the time needed to do a sweep came as a total -shocking- surprise to me!

Oh, now you've done it ... you said something bad about a Siglent! 

Everyone that's actually used the Siglent Bode or FRA function, whatever you want to call it, has said it's slow; ourselves, 2N3055, Rudi, and so on. This is likely a consequence of the massive amount of computational energy required to do such, and puts a serious strain on the available computational resources (likely why everything else slows down or stops completely when Bode is active). We must remember that all the Siglent Bode effects are created purely in the digital domain (read software) without any hardware participation nor acceleration (to the best of our knowledge). The R&S may have higher computational resources available, may have hardware assist, or maybe just a better method to implement the function, will let the folks that actually have and know this fine instrument comment tho.

The real discussion as pointed to by 2N3055 is "at this price point" nobody has a competing instrument period!!

We challenge you or anyone to show this isn't true, and if someone finds a better solution then we all benefit

We'll repeat again what we've said all along, the Siglent Bode implementation is technically a superb effort, and at this price point a great value added feature. Recall some actually stating that the Siglent Bode function was just a toy feature, obviously they were/are uninformed!!

All this discussion begs the evaluation of the higher end Siglent instruments for Bode speed. We call all these instruments (DSO & MSO) because they are really highly sophisticated, high speed/resolution, data acquisitions systems with excellent low noise and high dynamic range analog front ends with wonderful digital display renderings


Anyway, maybe someone with various instruments (1000, 2000, 5000 & 6000 series) at hand could do a comparison of Bode Function speed? Would be interesting to see if the newer Siglent instruments are significantly faster than the entry and mid level devices. This might hint at where the bottleneck is regarding speed.

Best,

[mawyatt]

We find it somewhat amusing that the very mention of a "scope" of any form, brings up the usual heated discussions and arguments when the reality is that the "scope" offerings by the major sources discussed are actually all superb instruments, and far more that just a visual rendering of a time domain waveform.

They all are so darn good, we should be thankful and admiring the engineering genius behind them rather than trying to nitpick them apart 

Best,

[nctnico]

Don't rule out the Analog discovery:

They all are so darn good, we should be thankful and admiring the engineering genius behind them rather than trying to nitpick them apart 

Only through nitpicking you'll learn about the flaws. I'm not interested in how well something works on a sunny day, I want to know where the limits of a piece of equipment are.

[mawyatt]

Only through nitpicking you'll learn about the flaws. I'm not interested in how well something works on a sunny day, I want to know where the limits of a piece of equipment are.

Would argue that nitpicking doesn't verifiably revel anything, hands on test and verification verifiably does

Best,

[nctnico]

That is just semantics. Doing detailed tests to get to the finest details is the definition of nitpicking 

[mawyatt]

From the original definition:
"Nitpicking is a term, first attested in 1956, that describes the action of giving too much attention to unimportant detail." 

Edit, from Wiki

https://en.wikipedia.org/wiki/Nitpicking

Best,

[nctnico]

Define unimportant detail  At some point every detail becomes important. For some reason I tend to get bitten by details that other people deem unimportant. 

[2N3055]

Don't rule out the Analog discovery:

They all are so darn good, we should be thankful and admiring the engineering genius behind them rather than trying to nitpick them apart 

Only through nitpicking you'll learn about the flaws. I'm not interested in how well something works on a sunny day, I want to know where the limits of a piece of equipment are.

A device without BNC inputs (and outputs) and proper case is not a real scope in my opinion.  Try using one on a daily basis...
OTOH, it is really useful little thing and I wish I had something like that when I was a kid... But it's a gadget.

[nctnico]

There is a 'pro' version that has BNC connectors. But nothing stops you from adding connectors or a dedicated test fixture yourself. AFAIK the analog discovery has differential inputs. If so, you could create much better test fixtures then you can ever achieve on a common-ground oscilloscope.

[Martin72]

Anyway, maybe someone with various instruments (1000, 2000, 5000 & 6000 series) at hand could do a comparison of Bode Function speed? Would be interesting to see if the newer Siglent instruments are significantly faster than the entry and mid level devices. This might hint at where the bottleneck is regarding speed.

Hehehehe....
I got a SDS1104X-E and my SDS2540X-HD at home, and access to a SDS2104X+ at work.
As I´ve tested the batronix demoboard, bodeplot with the 1104X-E was the fastest.
Could re-test it again, making sure the settings were all the same.

[tautech]

Anyway, maybe someone with various instruments (1000, 2000, 5000 & 6000 series) at hand could do a comparison of Bode Function speed? Would be interesting to see if the newer Siglent instruments are significantly faster than the entry and mid level devices. This might hint at where the bottleneck is regarding speed.

Hehehehe....
I got a SDS1104X-E and my SDS2540X-HD at home, and access to a SDS2104X+ at work.
As I´ve tested the batronix demoboard, bodeplot with the 1104X-E was the fastest.
Could re-test it again, making sure the settings were all the same.

Yes you would need to, things like points # and autoscaling engaged could dramatically slow performance.

However all this my Bode plot is better than yours is nothing but hot air  without a standardized test ! 

[Martin72]

As a German cabaret artist once said, comparing times only makes sense if everyone starts at the same time.
This can be reinterpreted according to the situation, always fits...

nothing but hot air

In the meantime, I ignore such things as far as possible.
Do it yourself, experience it yourself, then you can say something meaningful about it.

[Fungus]

A device without BNC inputs (and outputs) and proper case is not a real scope in my opinion.  Try using one on a daily basis...

The BNC adapter board is a cheap accessory. You can 3D print a case:

https://www.eevblog.com/forum/testgear/3d-printable-case-for-analog-discovery-2/

[mawyatt]

Do it yourself, experience it yourself, then you can say something meaningful about it.

Exactly, as we mentioned hands on experience defines ones level of detail, what's important, and not, and we don't pay much attention to those without hands on, which quickly whittles things down to verifiable information and use.

Even spec sheet racing can lead to misinformed decisions, with "hands on" one gets an actual "feel" for how the instrument behaves in ones technical arena! This is why value added distributers are important, they can quickly get the instruments into ones hands for evaluations in local regions.

Way back when we were employed, Tektronix and HP reps were always bringing new instruments by for our evaluation, and if we liked something (seldom looked a spec sheets) we could place a PO in the next capital procurement cycle, of course having a total new equipment budget over $10M/Yr helps

Best,

[mawyatt]

A device without BNC inputs (and outputs) and proper case is not a real scope in my opinion.  Try using one on a daily basis...
OTOH, it is really useful little thing and I wish I had something like that when I was a kid... But it's a gadget.

We've had an Analog Discovery for some time now, and have never even turned it on. It's too much of a hassle to use for our needs, and more of a learning tool IMO, evidently a pretty good tool in that respect!! Agree this would have been a blast as a kid to have 

Heck, we have a PicoScope 4262 that hardly ever gets used, also too much of a hassle to get up and running with the laptop. For some reason we just feel that a scope should be stand alone and not require a laptop, or any other "fixture" to make it usable, but we are old school (very old) and other folks may have a different view , which is fine

Anyway, our Bode plot is not better than anyones, but it does serve our needs quite well indeed, altho could be faster

Best,

[2N3055]

A device without BNC inputs (and outputs) and proper case is not a real scope in my opinion.  Try using one on a daily basis...

The BNC adapter board is a cheap accessory. You can 3D print a case:

https://www.eevblog.com/forum/testgear/3d-printable-case-for-analog-discovery-2/

 A person comes and asks for  replacement for his old and trusty analog scope.  He never mentioned he wants scope in a kit.

There is a 'pro' version that has BNC connectors. But nothing stops you from adding connectors or a dedicated test fixture yourself. AFAIK the analog discovery has differential inputs. If so, you could create much better test fixtures then you can ever achieve on a common-ground oscilloscope.

Again, no wishes of DIY scope. Or "I want to create text fixtures so what would be best instrument for that?"..
And AD Pro is a joke.. Same mediocre device, just this time really expensive... scope with 32k buffer for 1300 USD +VAT...
Again OP asked for specific class of scope in specific price range...


People read the title:  "Oscilloscope recommendation DS1054Z / SDS1104X-E / GDS-1104B"

that excludes:

- anything that is more expensive than some 500-600 USD
- USB scopes

[Fungus]

A person comes and asks for  replacement for his old and trusty analog scope.  He never mentioned he wants scope in a kit.

But then the discussion turned to Bode plots, and the AD2 has one of the best for half the price of a 'scope setup.

Try to keep up.