REVIEW - Siglent SDS2000 series - A Comparison of Features with Rigol DS2000

[Mark_O]

FFT have 1k sample points what give 0.5k resolution. (pity)

I agree that is unfortunate.  Even the cheap(er) Rigols are 2k-sample FFTs, which is also fairly anemic.  Siglent should consider increasing this, in a future update.

I assume this is being done to maintain a decent refresh rate, but when the acquisition is stopped, it would provide an opportunity to do a lot more with the available sample set.  Siglent should take advantage of that.

[Mark_O]

...in my opinion the way FFT is implemented in modern low-end scopes (not just Siglent, Rigol isn't really much better) is absolutely atrocious.

I agree.  And I suspect the reason it is that way is simply so they (all the lower-end guys, not picking on Siglent) can claim they have FFT, regardless of how poor it may be in practice.  Just a box to check off on their Feature List.

There's a big difference between things done right, and things just done.  Many vendors these days are opting for quantity of features over quality.  And considering that the ones they do have on their Feature Lists often don't even work (or work poorly), it's probably too much to ask that they do it "right".  That will only come as things evolve (and competition forces it), and they can leverage off a core foundation of existing code, in their next-gen product lines.  Rather than starting from scratch and re-inventing the wheel each new hardware/firmware cycle.

[Wuerstchenhund]

Even the cheap(er) Rigols are 2k-sample FFTs, which is also fairly anemic. 

It is. What's even worse, though, is that for the DS4000 and DS6000 (both pretty expensive scopes) it's apparently only 700 points:
http://beyondmeasure.rigoltech.com/acton/attachment/1579/f-00f2/0/-/-/-/-/file.pdf

This is pretty embarrassing. The listed 600pts for the old DS1000 Series looks great in comparison.

I assume this is being done to maintain a decent refresh rate, but when the acquisition is stopped, it would provide an opportunity to do a lot more with the available sample set.  Siglent should take advantage of that.

The reason is very likely the severe limitations in processing capabilities in these low-end scopes (which as a result will of course also affect the refresh rate). FFT is pretty heavy in CPU ressources, and I doubt the simple controllers in these scopes can deal with anything like say 10kpts at a somewhat reasonable speed.

And I suspect the reason it is that way is simply so they (all the lower-end guys, not picking on Siglent) can claim they have FFT, regardless of how poor it may be in practice.  Just a box to check off on their Feature List.

I fully agree, FFT seems to be merely a check-box exercise so that they can list it on the data sheet, no matter that at this stage it's borderline useless.

There's a big difference between things done right, and things just done.  Many vendors these days are opting for quantity of features over quality.  And considering that the ones they do have on their Feature Lists often don't even work (or work poorly), it's probably too much to ask that they do it "right".  That will only come as things evolve (and competition forces it), and they can leverage off a core foundation of existing code, in their next-gen product lines.  Rather than starting from scratch and re-inventing the wheel each new hardware/firmware cycle.

Indeed. That's one reason why I rather buy a not too old second-hand scope than a new entry-level scope. The performance per buck in the low end class, while having increased dramatically in recent years, is for me still much better with a second hand scope.

[Mark_O]

Even the cheap(er) Rigols are 2k-sample FFTs, which is also fairly anemic. 

It is. What's even worse, though, is that for the DS4000 and DS6000 (both pretty expensive scopes) it's apparently only 700 points:
http://beyondmeasure.rigoltech.com/acton/attachment/1579/f-00f2/0/-/-/-/-/file.pdf

This is pretty embarrassing. The listed 600pts for the old DS1000 Series looks great in comparison.

I believe that document was from 2012.  According to this Oct/2013 document,

"The DS2000, DS4000, and DS6000 Scopes use 2048 data points from the waveform to calculate the FFT." 

That's where I obtained the 2k figure I mentioned previously.

[Wuerstchenhund]

I believe that document was from 2012.  According to this Oct/2013 document,

"The DS2000, DS4000, and DS6000 Scopes use 2048 data points from the waveform to calculate the FFT." 

That's where I obtained the 2k figure I mentioned previously.

Ah, ok. That's better of course but still abysmally low. And pretty pathetic for a scope at the price of the DS4k and DS6k.

[nctnico]

There's a big difference between things done right, and things just done.  Many vendors these days are opting for quantity of features over quality.  And considering that the ones they do have on their Feature Lists often don't even work (or work poorly), it's probably too much to ask that they do it "right".  That will only come as things evolve (and competition forces it), and they can leverage off a core foundation of existing code, in their next-gen product lines.  Rather than starting from scratch and re-inventing the wheel each new hardware/firmware cycle.

Indeed. That's one reason why I rather buy a not too old second-hand scope than a new entry-level scope. The performance per buck in the low end class, while having increased dramatically in recent years, is for me still much better with a second hand scope.

IMHO it depends on what you find important. Most second hand scopes don't have serial protocol decoding and digital inputs which is why I bought a new oscilloscope. It just sucks that in the low costs range you appearantly have to trade protocol decoding + digital for proper signal analyses functions.

[Mark_O]

Ah, ok. That's better of course but still abysmally low. And pretty pathetic for a scope at the price of the DS4k and DS6k.

I would agree that the higher-end Rigols should support something closer to the (up to) 64k FFTs on the Agilent MSO-X scopes.  At least as an option, that could be disabled if it was too slow.

[Wuerstchenhund]

IMHO it depends on what you find important. Most second hand scopes don't have serial protocol decoding and digital inputs which is why I bought a new oscilloscope. It just sucks that in the low costs range you appearantly have to trade protocol decoding + digital for proper signal analyses functions.

That's true, serial decode severely limits the choice for 2nd hand scopes to a few newer scopes. The Agilent MSO6k comes to mind but these scopes are still pretty expensive. LeCroy X-Stream scopes (with the exception of the WaveSurfer 400) can do serial decode for a wide range of protocols (serial, CAN, USB, Firewire, Ethernet, ARINC/MIL-1553 and so on) as a software option, but unless the scope in question already comes with the required option it's hardly a financially viable way (the options are still available from LeCroy but aren't cheap).

And if you need MSO capabilities (digital inputs) then it gets even worse.

At the moment I can live with serial decode only (my WavePro has all the decode options) but should I need digital channels then I'll probably just get one of the USB logic analyzers.

[tautech]

IMHO it depends on what you find important. Most second hand scopes don't have serial protocol decoding and digital inputs which is why I bought a new oscilloscope. It just sucks that in the low costs range you appearantly have to trade protocol decoding + digital for proper signal analyses functions.

That's true, serial decode severely limits the choice for 2nd hand scopes to a few newer scopes. The Agilent MSO6k comes to mind but these scopes are still pretty expensive. LeCroy X-Stream scopes (with the exception of the WaveSurfer 400) can do serial decode for a wide range of protocols (serial, CAN, USB, Firewire, Ethernet, ARINC/MIL-1553 and so on) as a software option, but unless the scope in question already comes with the required option it's hardly a financially viable way (the options are still available from LeCroy but aren't cheap).

And if you need MSO capabilities (digital inputs) then it gets even worse.

At the moment I can live with serial decode only (my WavePro has all the decode options) but should I need digital channels then I'll probably just get one of the USB logic analyzers.

Reminding you the SDS2000 series until Feb15 2015 offer of buy 1 option, get 2 more free.

[nctnico]

IMHO it depends on what you find important. Most second hand scopes don't have serial protocol decoding and digital inputs which is why I bought a new oscilloscope. It just sucks that in the low costs range you appearantly have to trade protocol decoding + digital for proper signal analyses functions.

That's true, serial decode severely limits the choice for 2nd hand scopes to a few newer scopes. The Agilent MSO6k comes to mind but these scopes are still pretty expensive. LeCroy X-Stream scopes (with the exception of the WaveSurfer 400) can do serial decode for a wide range of protocols (serial, CAN, USB, Firewire, Ethernet, ARINC/MIL-1553 and so on) as a software option, but unless the scope in question already comes with the required option it's hardly a financially viable way (the options are still available from LeCroy but aren't cheap).

And if you need MSO capabilities (digital inputs) then it gets even worse.

At the moment I can live with serial decode only (my WavePro has all the decode options) but should I need digital channels then I'll probably just get one of the USB logic analyzers.

But those are not convenient to use and usually have limited memory. The past few days I have been using my SDS2204 for debugging a CPLD design and scrolling through / magnifying the digital signals using rotary knobs is very handy.

[Mark_O]

And if you need MSO capabilities (digital inputs) then it gets even worse.

At the moment I can live with serial decode only (my WavePro has all the decode options) but should I need digital channels then I'll probably just get one of the USB logic analyzers.

Reminding you the SDS2000 series until Feb15 2015 offer of buy 1 option, get 2 more free.

And also reminding everyone (again) that it's not really a generalized Buy 1, Get 2 Free deal. 

It is specifically, buy the $299 8-channel LA option, and get the $188 FG function generator and $188 DC decoder options, for free.  A $376 savings.  The protocol decoder option was already a win over the Rigol options, because Rigol breaks them up into multiple pricey chunks.  While Siglent bundles all the decoders into one affordable package.  And, for a limited time, they're even cheaper.  I.e., 55% off on that option bundle.