Siglent special offers on the SDS2000X series

[nctnico]

And dare we say thanks for your constructive input for ongoing product development.
Not likely, there's been none.   

I'm still waiting for RF-loop's comparison between the SDS1000X/2000X versus other brands when it comes to the advanced features like digital channels, decoding, decoding on digital channels, FFT, math, processing speed, network connectivity, etc. Pretty pictures is one thing but every time I bring up comparing the advanced features then suddenly they are not relevant or there is an akward silence. So man up and show how much the firmware has been improved by Siglent!

[rf-loop]

Other deal breaker for me is the lack of a proper, pre decimation peak detection mode, if I listened carefully Performa's review set.

This must be some kind of misunderstanding of the wrong interpretation.
Least in SDS2000 (2304) I can not see any problem in peak detect. (and I believe SDS2000X is least same or better.)

Conclusion:

Peak detect works absolutely perfect in roll mode and generally for time bases >20ms/div.

Below that, it seems there would still be room for improvement. As it is now, the peak values are acquired from simply decimated data according to the current sample speed, instead of replacing the decimation by a peak detect mechanism.

As it is now, if we had a perfect 1ns wide pulse we would be able to trigger on it but still might not see anything on the screen.

Big question: Since it is possible to do it in an optimal way for timebases >20ms, why on earth has it to be different for faster timebases?

There is some kind of bug in peak detect as Performa01 have noted.

This is based to my testings and valid only for SDS2000 without X.

For reliable display in peak detect mode.
It looks like it happend with 200 - 500MSa/s with some memory lenght.
There is available guite lot of memory settings and until this bug is repaired I recommend to avoid 200-500MSa/s speeds.
This can do by selecting amount of memory. There is 1.4k to 70M (in SDS2000) memory settings. If your t/div setting with memory you have selected give samplerate between 200, 250, 400, 500MSa/s just select amount of memory so that speed is 100MSa/s or below or 1GSa/s or more.  I can not find situation where I can not do this. Of course, this is my personal recommendation and it is temporary until the problem is resolved.

What ever speed in peak mode trigger itself is rock solid reliable if peak width what cross over trigger level  is 1ns or more. 
I can perhaps imagine how this happend there after ADC but without real knowledge how it trig and how it pick up data for display pathway it can only do more or less clever guesswork.

I have just tested it using all t/div speeds with pulse what scope see around 0.5ns to 2ns width and setting trigger level where is is just bit over 1ns   Trigger counter do not loose pulses and what ever t/div I can select memory so that samplerate is 100MSa/s or less or 1GSa/s or more. It need note that with full memory there do not even need peak detect between 5ms/div to 1ns/div range because 5ms/div can do 1GSa/s and 2ms/div to 1ns/div all is 2GSa/s.  Below 5ms/div if want peak detect work need drop memory so that samplerate drops to 100MSa/s or less and/or go to more slow timebases down to max. Including also roll mode.

[Wuerstchenhund]

As the last FW for both 2000 and 2000X series was released early Dec last year and Siglent have taken detailed notice of Performa01's analysis of the later V2 FW, (mostly in Jan, at least those you are quoting) it would be prudent to expect further tweaks (some based on Performa01's findings) for the 2000X series.

You mean like the kind of tweaks they did for the SDS2000 in the two years since the scope has been released, many based on user findings?   

And dare we say thanks for your constructive input for ongoing product development.

So I guess you're agreeing then that the SDS2000 (amongst other Siglent products) are still in development state and haven't yet reach a status where they can be considere mature enough for being released to the market? 

Not likely, there's been none.   

I believe myself (and many others) have given Siglent way more feedback for their products (in my case the SDS1000, the SDS1102CML and the SDS2000) than a manufacturer that charges 100% for a product where maybe 80% of the specified functionality works correctly or at all at release really deserves. Not that it was of much use, considering how long it took them to even implement some of the simpler fixes.

But of course I'm happy to offer to extensively test any new Siglent product and provide an extensive list of bugs and potential feature improvements, however I fear my hourly charge-out rate might be a bit too high for them (only a moron would do it for free). 

I guess instead of fixing the current mess first they rather throw new bug-ridden products on the market, like that new SSA3000X, which again comes with hillarious bugs like having the screenshot file selection window captured in that screenshot itself 

You really can't make that shit up 

I'm still waiting for RF-loop's comparison between the SDS1000X/2000X versus other brands when it comes to the advanced features like digital channels, decoding, decoding on digital channels, FFT, math, processing speed, network connectivity, etc. Pretty pictures is one thing but every time I bring up comparing the advanced features then suddenly they are not relevant or there is an akward silence. So man up and show how much the firmware has been improved by Siglent!

  Exactly.

[nctnico]

Other deal breaker for me is the lack of a proper, pre decimation peak detection mode, if I listened carefully Performa's review set.

This can do by selecting amount of memory. There is 1.4k to 70M (in SDS2000) memory settings. If your t/div setting with memory you have selected give samplerate between 200, 250, 400, 500MSa/s just select amount of memory so that speed is 100MSa/s or below or 1GSa/s or more.  I can not find situation where I can not do this. Of course, this is my personal recommendation and it is temporary until the problem is resolved.

How long is temporary? Peak detect never worked right on the SDS2000(X) for over 2 years now!

[rf-loop]

I guess instead of fixing the current mess first they rather throw new bug-ridden products on the market, like that new SSA3000X, which again comes with hillarious bugs like having the screenshot file selection window captured in that screenshot itself 

Your SSA3000X comment is valid for this?
Current Version: 07.03 | Published?2016-03-29

Today this is history but..
I remember my old R&S Spectrum analyzer what price was one house when it was new.
There was nearly fatal class bug. If with narrow range RBW (least if was using this narrow continuously adjustable RBW) adjust too fast freq span or center freq whole equipment crash (freeze) and can not go out until boot agen. All user settings loosed and default setup loaded.  Even with latest firmware what have less bugs what previous. Never fixed.

But, it is clear that Siglent (as many others also) need develop better practices for FW beta/final test before product is accepted to public sell.
There need be independent test group who know how equipments are used in real works in customers side and never they can same persons who do FW programming. If programmer is who do these important tests it is big major class mistake! Programmer is many times blind for his own mistakes exept in very very early phase basic things!  Also testers need have lot of experience. Saving 1 amount of money in this phase costs easy 100 or thousend time more.  Whole working culture need develop. (And least I believe it is going on but also time is running...)

[Wuerstchenhund]

I guess instead of fixing the current mess first they rather throw new bug-ridden products on the market, like that new SSA3000X, which again comes with hillarious bugs like having the screenshot file selection window captured in that screenshot itself 

Your SSA3000X comment is valid for this?
Current Version: 07.03 | Published?2016-03-29

I don't know. But the thing is that such blatantly obvious bugs should not be in a release product.

Today this is history but..
I remember my old R&S Spectrum analyzer what price was one house when it was new.

What model was that?

There was nearly fatal class bug. If with narrow range RBW (least if was using this narrow continuously adjustable RBW) adjust too fast freq span or center freq whole equipment crash (freeze) and can not go out until boot agen. All user settings loosed and default setup loaded.  Even with latest firmware what have less bugs what previous. Never fixed.

I wonder what model that was, because I know pretty much all the R&S SAs back to the old FSEA Series, and while I've seen quite a few bugs in them, I can't recall that anything that serious would have been left unfixed.

But, it is clear that Siglent (as many others also) need develop better practices for FW beta/final test before product is accepted to public sell.
There need be independent test group who know how equipments are used in real works in customers side and never they can same persons who do FW programming. If programmer is who do these important tests it is big major class mistake! Programmer is many times blind for his own mistakes exept in very very early phase basic things!  Also testers need have lot of experience. Saving 1 amount of money in this phase costs easy 100 or thousend time more.  Whole working culture need develop. (And least I believe it is going on but also time is running...)

I fully agree with you here, however I think it's clear now that this unlikely to happen. Siglent isn't a new business, they're doing this for a while, and I'm sure they must know that pretty much everything comes out with bugs that should not be in a release quality product. It's not that they don't know any better, they willingly decide to forego improving their development methods, very likely in a means to keep costs down.

The sad thing is that this is actually a real shame, because I believe their hardware (with some exceptions) is actually pretty good. For the SDS2000, my only complaint with the hardware would be that the multi-function encoder knob was placed too close to another knob (was it Trigger Level?). That's it. Everything else is down to software. If they were to do software properly, they had very good chances to climb up the T&M food chain to become a reputable brand which can then charge higher prices. But they're pretty much pissing away that potential with their software.

[rf-loop]

Other deal breaker for me is the lack of a proper, pre decimation peak detection mode, if I listened carefully Performa's review set.

This can do by selecting amount of memory. There is 1.4k to 70M (in SDS2000) memory settings. If your t/div setting with memory you have selected give samplerate between 200, 250, 400, 500MSa/s just select amount of memory so that speed is 100MSa/s or below or 1GSa/s or more.  I can not find situation where I can not do this. Of course, this is my personal recommendation and it is temporary until the problem is resolved.

How long is temporary? Peak detect never worked right on the SDS2000(X) for over 2 years now!

This time it can say it really did not work (V1 in SDS2k). Today, situation is very different. If I use SDS2000 it never fails in peak detect. I can extremely easy avoid just these some memory/samplerate combinations where this problem exist if peak what need detect is ns class. (these selections combinations can also turn off so that user can not do these settings, also this method for repair it is just ok if want fast repair.)
I do not try tell it is ok. But, when user know his equipment current limit he never trop to this trap what is extremely easy jump over (until it is repaired)  Previous peak detect FW repair was totally in different class.
Also this bug is not in HW.  But all know to do list have also some kind of priority order.

[rf-loop]

There was nearly fatal class bug. If with narrow range RBW (least if was using this narrow continuously adjustable RBW) adjust too fast freq span or center freq whole equipment crash (freeze) and can not go out until boot agen. All user settings loosed and default setup loaded.  Even with latest firmware what have less bugs what previous. Never fixed.

I wonder what model that was, because I know pretty much all the R&S SAs back to the old FSEA Series, and while I've seen quite a few bugs in them, I can't recall that anything that serious would have been left unfixed.

5GHz FSB  but not remember anymore exactly version.
(but it is true that this crash did not happend easy, only in some rare settings combinations + very fast turn adjusment knob, but still wonder it because mostly R&S (what I have used) equipments have been really in top-class or "state of art" class.)

[rf-loop]

I fully agree with you here, however I think it's clear now that this unlikely to happen.

I think opposite. Not as black and white but I believe they are developing. Of course this process is hard and perhaps also they have done mistakes and need learn.  Perhaps also in some phase some peoples have been too optimistic and believe that some thing is "easy to do" and after then reality come in... 

What I have talked with Siglent, and what experience I have with them  I can see continuous trend to better in many areas.
Also I have some kind of imagine what kind of recrytion process they have in China going on and it also may give signal that something is going on in product development. (some times it can see in Siglent (real) chinese internet sides).

[Wuerstchenhund]

5GHz FSB  but not remember anymore exactly version.

(but it is true that this crash did not happend easy, only in some rare settings combinations + very fast turn adjusment knob, but still wonder it because mostly R&S (what I have used) equipments have been really in top-class or "state of art" class.)

Well, the FSB was, together with the FSA, amongst the first spectrum analyzers R&S made, and that was way back in the '80s. I think the FSB was R&S' equivalent to HP's 8568B.

It's been a very long time when I last saw an FSB but if I remember right these things didn't even have firmware upgrades in the sense as we have today (the firmware was spread around the boards in various EPROMs, although some part of the user software was on a MFM hard drive?). R&S probably did fix the bug in a later firmware version, but a firmware upgrade at that time would have probably meant board swaps plus re-alignment which I'm sure wouldn't have come for free.

Back then we mostly had HP SAs (pretty much 8566A/Bs), and I remember that some of these had some minor bugs which wasn't present in newer production units, and when the affected wenit in for repair (which wasn't that uncommon) HP seemed to replace the firmware in those units as well.

Thanksfully, that's not much an issue today.

[rf-loop]

Other deal breaker for me is the lack of a proper, pre decimation peak detection mode, if I listened carefully Performa's review set.

This must be some kind of misunderstanding of the wrong interpretation.
Least in SDS2000 (2304) I can not see any problem in peak detect. (and I believe SDS2000X is least same or better.)

Conclusion:

Peak detect works absolutely perfect in roll mode and generally for time bases >20ms/div.

Below that, it seems there would still be room for improvement. As it is now, the peak values are acquired from simply decimated data according to the current sample speed, instead of replacing the decimation by a peak detect mechanism.

As it is now, if we had a perfect 1ns wide pulse we would be able to trigger on it but still might not see anything on the screen.

Big question: Since it is possible to do it in an optimal way for timebases >20ms, why on earth has it to be different for faster timebases?

Here is attached small table where can see problematic settings combination with PeakDetect mode.
NOTE: This is NOT SDS2000X, this is tested using SDS2000!
I have tested it now with every settings what can see in table.

Also with these problematic settings trigger itself work rock solid. But some times detected pulse is very attenuated on the display. If all is ok, trigged 1ns or more wide pulse must be least trigger level height.

(ADD: Note for  table:  Peak detect works  also with these orange colored settings but do not meet specifications >1ns minimum. In these areas it also trigger rock solid down to 1ns glitches.  Problem is peak draw. In table orange areas it looks like that reliable display need >2 - >5ns glitch width, depending decimated samplerates)





Also just two random images from tests...

[pxl]

At 1ms, how does a 50ns pulse look like with 14k? (e.g. at rate 10kHz)

[rf-loop]

At 1ms, how does a 50ns pulse look like with 14k? (e.g. at rate 10kHz)

Not exactly as you want but I find these. If neeed later can do as you want at 10kHz.
In these images pulse is rectangle and rise/fall 6ns and middle width 50ns. first near 500Hz and two other  well under 1Hz

[pxl]

Thanks, it seems good, indeed. The problem really edge case then.

[nctnico]

The problem really edge case then.

No. Nothing is more confusing than a signal showing at one time/div settng and dissapearing on another. It makes you doubt the measurements. Try it and you'll see for yourself...

[rf-loop]

The problem really edge case then.

No. Nothing is more confusing than a signal showing at one time/div settng and dissapearing on another. It makes you doubt the measurements. Try it and you'll see for yourself...

It can correct easy. Siglent can shut off these "illegal" settings combinations in PeakDetect mode.

[rf-loop]

At 1ms, how does a 50ns pulse look like with 14k? (e.g. at rate 10kHz)

Here image using 50ns pulses with 100us period (10kHz) and PeakDetect 1ms/div, 14k memory (in SDS2k what ever memory lenght is selected or currently in use,  whole lenght is visible, no overlap = what is captured is what you see, even with 70Mpoints lenght. This is Siglent solution, others can do how they want.)

(Note for previous table:  Peak detect works  also with these orange colored settings but do not meet specifications >1ns minimum. In these areas it also trigger rock solid down to 1ns glitches.  Problem is peak draw. In table orange areas it looks like that reliable display need >2 - >5ns glitch width, depending decimated samplerates)

Note agen, this is SDS2000   not SDS2000X !

[willd1971]

Attention UK buyers.  The SDS2000X promotion is now available from Labtronix here:

https://labtronix.co.uk/drupal/shop/oscilloscope/sds2000x.htm

Labtronix can drop-ship these oscilloscopes direct from Siglent EU, takes no more than 5 days - we always aim to give the best service and the best prices!

<***UPDATE***>
Delivery time has increased to approximately 2 weeks due to high demand - Sorry.
Hopefully this will settle down after the initial period of the offer and more stock will be on-hand
</***UPDATE***>

All the best

William

[mark-r]

Attention UK buyers.  The SDS2000X promotion is now available from Labtronix here:

Brilliant, I'll order one soon 

[papousek]

@ Siglent
i can not find official seller in Japan,,,please help

[F5D]

Hi guys, I am new on this forum. Fantastic resource of information! I am mainly an audio guy, collecting synthesizers etc., but do my own maintenance and modifications to gear, where I need test equipment. I am also looking for a new scope to keep up-to-date. At work I use more expensive stuff. This Siglent series looks really promising, especially with the offer price.

I would also like to hear what are the actual differences in use between the SDS2000X and SDS1000X series? Is the 2000-series really faster drawing on the screen with same settings? I just noticed an interesting thing browsing the brochures. For 1000-series, the highest sensitivity level listed is 500 uV/div, whereas for the 2000-series it is only 1 mV/div. This must be a mistake, right? The 2000X should reach the same noise performance as the 1000X? Also, would like to know how the enhanced resolution (eres) works with the latest firmware, how does it affect the update rates etc. And lastly, the FFT-mode, is it useable or not. I mainly need frequency range 0 - 500 kHz for audio equipment.

[Wuerstchenhund]

Hi guys, I am new on this forum. Fantastic resource of information! I am mainly an audio guy, collecting synthesizers etc., but do my own maintenance and modifications to gear, where I need test equipment. I am also looking for a new scope to keep up-to-date. At work I use more expensive stuff. This Siglent series looks really promising, especially with the offer price.

Welcome to the forum!

The scope looks promising but I'd strongly recommend to read the threads about the SDS2000 and other Siglent gear as 'looking promising' doesn't mean the promises get fullfilled. Siglent has a very poor track record in terms of firmware bugs, and the SDS2000 (the predecessor of the SDS2000X, with the difference being down to cosmetics and sample memory size) still has issues even after more than 2 years on the market. I had a SDS2204 myself, and it was truly by far the worst scope I ever had (and I had a lot!).

Therefore my advice, do your research, don't believe the marketing that tells you about all the great functions you get, make sure if you buy a Siglent scope that you can easily return it as many functions may not work correctly or at all.

[rf-loop]

Is the 2000-series really faster drawing on the screen with same settings?

Yes and no. Mostly yes.
Highest wfm/s speed for SDS2k(X) is 140kwfm/s.  Highest speed 60kwfm/s  with same settings.
But fastest speed exist with different t/div setting. SDS2k(X) highest speed is using 5ns/div and SDS1kX max speed exist with 50ns/div. There is more or less differencies if look different settings combinations but mostly SDS2k(X) is faster than SDS1kX. With some settings difference is marginal. Of course if we set 50ms/div and there is 750ms capture lenght in both. In this case of course difference is very minimal if any.

I just noticed an interesting thing browsing the brochures. For 1000-series, the highest sensitivity level listed is 500 uV/div, whereas for the 2000-series it is only 1 mV/div. This must be a mistake, right?

This is not mistake. SDS2k(X) do not have 500uV/div sensitivity and SDS1kX have and it is not digital vertical zoom, it is real.
Also in this forum you may find some tests etc.

Also, would like to know how the enhanced resolution (eres) works with the latest firmware, how does it affect the update rates etc.

It slows wafm/s speed lot of. If I say "typical" (both models) is that with ERES wfm/s speed is always <25wfm/s and mostly something roughly like 10wfm/s.  Also, maximum acquisition lenght using ERES is 14k. Eres is more hard process than "High-Res" in many scopes what use simple box car average.

And lastly, the FFT-mode, is it useable or not. I mainly need frequency range 0 - 500 kHz for audio equipment.

Yes, it works and can use but...
 it is far away what true audio work need. For this purpose - not for serious use.
First thing where is drops out is frequency resolution what is really rough due to small amount of sample points used. Other thing is speed. It is slow. Third is dynamic range. This FFT can use to some very limited purposes only or just example in basic schools,  kids can take lesson number one about what is FFT with it and play rest of lesson time in classroom before this lesson one hour ends. FFT is these scopes weakest point.

But need note that of course from scope can get out raw sampledata and give it to example Matlab or what ever.

[tautech]

Hi guys, I am new on this forum.

Welcome.

I would also like to hear what are the actual differences in use between the SDS2000X and SDS1000X series? Is the 2000-series really faster drawing on the screen with same settings?

Little apparent difference, the GUI on each series is very similar.
The predominant differences are at least double the sampling speed of the 2000X series, significantly more memory depth and faster update rates.

I just noticed an interesting thing browsing the brochures. For 1000-series, the highest sensitivity level listed is 500 uV/div, whereas for the 2000-series it is only 1 mV/div. This must be a mistake, right?

No mistake and in each case at full BW; ie. not BW limited at maximum sensitivity.

The 1mV sensitivity range was added to the 2000 series (previously 2mV BW limited) with release of the V2 FW some months back as a beta, but now officially adopted as "shipped" FW for 1000X, 2000 and 2000X series.

The 2000X should reach the same noise performance as the 1000X?

Why?
Higher BW of the 2000X series will display more apparent noise.

Check out Daves vids why:
http://www.eevblog.org/video/EEVblog601-WhyDigitalScopesAppearNoisy-640x360.m4v
http://www.eevblog.org/video/EEVblog610-WhyDigitalScopesAppearNoisyPart2-640x360.m4v

Also, would like to know how the enhanced resolution (eres) works with the latest firmware, how does it affect the update rates etc. And lastly, the FFT-mode, is it useable or not.

Much other information can be found in the below thread where the FW that was developed for the 2000X and 1000X series was also released for the 2000 series, the earlier HW version of the 2000X range.
https://www.eevblog.com/forum/testgear/siglent-sds2000-new-v2-firmware/


As there is not a lot of real info, (although plenty of conjecture  ) on the new 2000X range other than just these few on EEVblog that you've probably found already:
https://www.eevblog.com/2016/03/31/eevblog-864-siglent-sds2000x-series-oscilloscope-teardown/
https://www.eevblog.com/forum/testgear/siglent's-new-product-sds2000x-series/
https://www.eevblog.com/forum/testgear/sds2000x-japan-review/

Dave might add to these if he does a review on the 2304X that he did the teardown on. Let's hope he takes more care in any forthcoming review than he did for the 1000X.

Much of the "in depth" analysis of these Siglent's has been done by rf-loop and Performa01, members that physically have these DSO's and have shared their real findings.

[rf-loop]

The scope looks promising but I'd strongly recommend to read the threads about the SDS2000 and other Siglent gear as 'looking promising' doesn't mean the promises get fullfilled. Siglent has a very poor track record in terms of firmware bugs, and the SDS2000 (the predecessor of the SDS2000X, with the difference being down to cosmetics and sample memory size) still has issues even after more than 2 years on the market. I had a SDS2204 myself, and it was truly by far the worst scope I ever had (and I had a lot!).

Therefore my advice, do your research, don't believe the marketing that tells you about all the great functions you get, make sure if you buy a Siglent scope that you can easily return it as many functions may not work correctly or at all.

Yes you had old version and at this time it was really bad. But repeating and repeating your feel from history... it looks like some kind of fixation to this your sad case so that all where read Siglent is like red flag for bull.

Both models, current version of  SDS2304 and SDS1202X-S is here on my table and every potential customer in Finland can come here and test himself as much as want before any deal.  This is also what I always recommend before buy.  Also here in Finland is customers who have these equipments and perhaps also they can tell what they thing after they have get some experience with these.