EEVblog #797 - Siglent SDS1000X Review

[tautech]

The whole SDS1000X series has no option for a logic analyzer, I have no idea why they put the cutout for it on the front panel.

Neither did the Siglent SDS2000 series when it was first released, the demo models sent to dealers was an early HW version and later versions DO all have an MSO port and MSO capability as a free trial option until it is puchased and permanently enabled.

Will the SDS1000X series be available in the future with MSO? 

The 1 GSa/s and 14 Mpts does not impress me. Furthermore, the hack-ability of the Siglent gear is in question, in theory they can be hacked but no one has actually managed to do it yet.

With new beta FW for the SDS2000 series there will be additional memory depth available as an option and will Siglent enable an additional memory depth option for the SDS1000X series too? 
It wouldn't surprise me at all if they did.

[Lightages]

For me, the DS1054Z is still a better buy. The Siglent seems too buggy at this point, improperly matched sampling vs bandwidth means it is not that much better than the DS1054Z (hacked), and only two channels. So $100 less gets you 4 channels and basically the same useful bandwidth once you use two channels.....

[rf-loop]

The whole SDS1000X series has no option for a logic analyzer, I have no idea why they put the cutout for it on the front panel.

Do you think it is clever to pay new injection mold costs just after they do MSO1000X. or do separate front and/or back panel for every model. Yes they do if you pay it.
Have you ever asked how much it take money (total costs including all) when you do whole process for new mold.  There is one shot price and then there is of course price for every single the front panel they do. If you have ever designed any production you know that there need really heavy reason for do new mold.

[Wuerstchenhund]

With new beta FW for the SDS2000 series there will be additional memory depth available as an option and will Siglent enable an additional memory depth option for the SDS1000X series too? 
It wouldn't surprise me at all if they did.

I don't think we'll see that on the SDS1000X. On the SDS2000 it seems they replaced the original firmware (v1) which was a bug fest they were incapable of fixing with a port of the firmware for their new SDS1000X scopes (v2), and in the process were probably able to free up some memory which they can now offer as another option (which quite frankly should have been a free upgrade for all SDS2000, considering how long their users had to live with the shitty firmware!). It's certainly not something they designed into the scope from the start. The SDS1000X was designed with the firmware it runs on, so there won't be any similar memory savings.

I guess we'll see when the teardowns arrive, which should be able to establish how much memory there is in these things.

[rf-loop]

500uV/ on rigol 2000 is BULLSHIT. Did i say it loud enough?

B-U-L-L-S-H-I-T

 Do not even think of it as an advantage over other scopes. It is a x2 digital zoom off of 1mV/ . Nothing else happens when you set the switch to 500uS/, no any changes in the front end chain, zero, i have measured and can prove it. The bastards even did not bother to interpolate as they zoom in. They just stretch the display vertically. that is it. 

I do not know about Siglent.

I know Siglent. It is full true 500uV/div from analog and also without forced displayed BW reject or any internal hidden BW rejection... just tiny fast measurement and it looks like -3dB point related to 1MHz ref is around over 250MHz (perhaps same 275MHz what these other V/div bands what I have measured ). (I have not yet made enough tests for tell it more accurate. It is more tricky to measure than higher V/div settings.

[rf-loop]

With new beta FW for the SDS2000 series there will be additional memory depth available as an option and will Siglent enable an additional memory depth option for the SDS1000X series too? 
It wouldn't surprise me at all if they did.

I don't think we'll see that on the SDS1000X. On the SDS2000 it seems they replaced the original firmware (v1) which was a bug fest they were incapable of fixing with a port of the firmware for their new SDS1000X scopes (v2), and in the process were probably able to free up some memory which they can now offer as another option (which quite frankly should have been a free upgrade for all SDS2000, considering how long their users had to live with the shitty firmware!). It's certainly not something they designed into the scope from the start. The SDS1000X was designed with the firmware it runs on, so there won't be any similar memory savings.

I guess we'll see when the teardowns arrive, which should be able to establish how much memory there is in these things.

Did you note one major change. (and it is Same in SDS1000X !)
Rev 1.0 there was history as fast segmented memory together.
I do not put my old table about speeds and memory here.

Rev2.0 have all different just in this place.
History "mode" is not anymore at all this fast segmented memory acquiring.
History mode is now always on bacround and it is still up to 80000 acquisitions but now it works like history FIFO and you stop scope and you can see what have happend in last times and you can search waveforms (how much, depending memory and speed settings) But it works now in bacround and speed is just what is current normal visible wfm/s speed. Not anymore  max up to ~400kwfm/s (measured).
If we calculate how much there is memory needes in best (worst) case it is much and this memory need also be available quite fast.

Rev 2.0 have separate fast "segmented" memory acquisition. (even faster than Rev1.0)
It can up to 500kwfm/s (segments) speed, depending memory and timebase settings. But now maximum amount of segments is 1024. (Previously up to 80000)

This, same, is also in SDS1000X. It also have same max 1024 segments in segmented (sequence as it read in upper menu) and also in SDS1000X its max "up to" speed is 500kwfm/s (but limited 1024 segments)

 

Perhaps this change in SDS2k affect something in memory side what is now available for this option?





---------------------------
Other things related to video

on the video it looks like Dave measure burst speed when he measure normal wfm/s update rate. He did not measure average wfm/s speed what is bit lower and perhaps more recommended for claim wfm/s speed if we do not tell we are talking highest short burst speeds, and this is in SDS1000X over 70kwfm/s. (average speed need include display refresh cycle or other  system time gaps in acquisition and it include in "up to" 60kwfm/s speed. It is continuing average even if measure every trig events over second or hour.)

FFT cursors.
Just select cursors source math or channels.. and both tracking cursors may have separate source.
FFT there can use splitted or not splitted window but also there can me YT- splitted for YT zoom (this zoom have not relation to FFT, FFT use fixed samples.



Edit: segments max count is up to 80000 in SDS1000X after new  FW 1.1.2.1.1

[Helix70]

both scopes don't give each other much, and which one is the better buy really depends on what you require.

For most people I'd say 4 channels trumps 200Mhz.

Not necessarily. There are many tasks that don't require more than two channels, and where the extra bandwidth can be a real advantage.

Plus the level of maturity of the SDS1000X firmware is still unknown.

I wouldn't rush to buy one. Dave managed to find a some "turn it off and on again" bugs just in the making of this video...

Indeed, and Siglent isn't exactly known for bring products to market with mature firmware (i.e. the SDS2000). However, Rigol and the DS1000z had their own fair share of issues, and if the thread on this forum can be believed then it seems some are still waiting to be resolved.

I'd say it's not as clear cut as you think it is.

Horses for courses. If you are primarily working on sub 100 Mhz digital stuff (like me), the 1054Z with 4 channels is a no brainer. I probe the power supply, a couple of inputs and a debug output and I have my 4 channels, and I thank the lord for those extra channels most days.

If, however, you are looking at a modulated signal at or above 100 Mhz, the Siglent is going to cream the Rigol for that application. If you need above 200Mhz, keep on going to the Rigol 2000 series, and so on, and so forth. If you can upgrade the siglent to the sig gen by hacking, then I agree, there will be a standard choice of 2 DSOs for entry level, 4 channels or free sig gen, but Dave suggested this not the case, at least right now.

If you are looking for the cheapest swiss army knife DSO, the DS1054Z is hard to beat.

[Fungus]

improperly matched sampling vs bandwidth means it is not that much better than the DS1054Z

The bandwidth:sampling ratio is exactly the same as a DS1054Z. 

The only reason the signal started to distort in the video is because Dave was hitting it with a 300 MHz signal with both channels turned on.

Try a 150MHz signal into a DS1054Z with all four channels turned on and you'll get exactly the same thing.

[Fungus]

The -3 dB bandwidth of the $400 DS1000Z (same total sample rate as the new Siglent) is about 170-180 MHz, so no.

I heard the measured bandwidth was about 135Mhz (on EEVBLOG form)

1000Z Bandwidth:
Frequency      Amplitude
150 MHz         -3.0 dBm

It probably varies a bit from unit to unit (and cable).

Let's call it 150Mhz for the sake of argument.

[Fungus]

If you can upgrade the siglent to the sig gen by hacking, then I agree, there will be a standard choice of 2 DSOs for entry level, 4 channels or free sig gen, but Dave suggested this not the case, at least right now.

...except they don't cost the same. The Siglent is 50% more than a DS1054Z.

Maybe you can have 4 channels+separate 25MHz signal generator for the same price.

[tautech]

If you can upgrade the siglent to the sig gen by hacking, then I agree, there will be a standard choice of 2 DSOs for entry level, 4 channels or free sig gen, but Dave suggested this not the case, at least right now.

...except they don't cost the same. The Siglent is 50% more than a DS1054Z.

A flaw in your maths maybe.....

AFAIK the DS1054Z is US$399 and the SDS1102X is US$499, excuse me but that difference is not 50%.

[Fungus]

...except they don't cost the same. The Siglent is 50% more than a DS1054Z.

A flaw in your maths maybe.....

AFAIK the DS1054Z is US$399 and the SDS1102X is US$499, excuse me but that difference is not 50%.

Oh, I see... you meant the 100MHz SDS1102X version. Yes, that costs less.

If you can unlock the bandwidth and signal generator then it will be an interesting choice.

[rf-loop]

...except they don't cost the same. The Siglent is 50% more than a DS1054Z.

A flaw in your maths maybe.....

AFAIK the DS1054Z is US$399 and the SDS1102X is US$499, excuse me but that difference is not 50%.

Oh, I see... you meant the 100MHz SDS1102X version. Yes, that costs less.

If you can unlock the bandwidth and signal generator then it will be an interesting choice.

If look BW 1000Z and SDS1102X is comparable instead of 4 channel in Z.

But other side...  Z do not have real 500uV/div, not even DS2000.
Z do not have even 1mV/div true, not even 2mV/div

Then, Z do not have sequence mode where it can acquire up to 500000 waveform (segment)/s speed not even near.
Then Z mask test is total junk. It is totally slow if mask test result is fail.
Then Z do not have full speed backround history where X stores laast up to 80000 acquistions to history buffer all time when scope run (exept if sequence mode selected)..
Then Z do not have selectable Sinc (only it can select on/off when 4 ch in use. And it really do not work even there right. (first principle is that it need draw sinc through real unmanipulated raw sample points but here it loose reality... documented in other threas)
Then Z have smaller display. (specially example for me reading Z display is pain.)

Both have advantages and disadvantages. No one can say which is universally better. If know user individual needs then it can tell which one is better for this user.

And so on.


Attached images.

Signal and signal path is exactly same for both scopes. All is same as possible, even selfcal done just before for both.  Because Rigol did not have 50ohm inputs I use also with Siglent same external feed thru terminator.
Signal is same ans quite good quality from HP E4421B
Test made with 1mV/div because both have it.
It need also note when look noise. Part of noise is related to BW and Siglent analog BW is much more than Rigol DS1kZ.
And Rigol advertise this with many kind of superlatives about low noise...
Note with images. True display physical size in Rigol is 7" and Siglent 8"

Images tell many things.
Also sidenote, look trigger point.

[Orange]

Plus ... if people can hack/unlock the signal generator with then it's a killer device. They'll sell truckloads of them.

I've heard that Siglent have taken extra measures to ensure that's not going to happen.
I give it a week once they hit the shelves

Bad marketing strategy if you ask me.  The RIGOL DS2072A can be unlocked to 300MHZ, plus memory, plus all decoders for 891 euros including 21% TAX. And on top of that it is a 2Gs/s system.

Mind you that in Europe the SDS1202X is offered for 756 euros including TAX. This still a 1Gs/s system, and no decoding options....And probably no ways to hack the thing.
http://www.siglent.eu/sds1202x.html

[AF6LJ]

Not bad for an entry level scope.

[Hydrawerk]

Can that tiny readout font (Vpp, Freq) be enlarged?

[Lightages]

improperly matched sampling vs bandwidth means it is not that much better than the DS1054Z

The bandwidth:sampling ratio is exactly the same as a DS1054Z. 

The only reason the signal started to distort in the video is because Dave was hitting it with a 300 MHz signal with both channels turned on.

Try a 150MHz signal into a DS1054Z with all four channels turned on and you'll get exactly the same thing.

Hmmm, this seems to be what I am saying and you arguing my point for me. The DS1054Z performs (basically) the same as the Siglent with TWO channels, apples to apples. With one channel, the Sigent does have some advantage. With four channels the Siglent isn't even in the picture! So IMHO, $100 more for the Siglent with only the added usable bandwidth in one channel mode vs the option of four channels (lower bandwidth) is not $100 well spent. There are other benefits of course, but TO ME, IMHO, it is not worth it.

[rf-loop]

improperly matched sampling vs bandwidth means it is not that much better than the DS1054Z

The bandwidth:sampling ratio is exactly the same as a DS1054Z. 

The only reason the signal started to distort in the video is because Dave was hitting it with a 300 MHz signal with both channels turned on.

Try a 150MHz signal into a DS1054Z with all four channels turned on and you'll get exactly the same thing.

Hmmm, this seems to be what I am saying and you arguing my point for me. The DS1054Z performs (basically) the same as the Siglent with TWO channels, apples to apples. With one channel, the Sigent does have some advantage. With four channels the Siglent isn't even in the picture! So IMHO, $100 more for the Siglent with only the added usable bandwidth in one channel mode vs the option of four channels (lower bandwidth) is not $100 well spent. There are other benefits of course, but TO ME, IMHO, it is not worth it.

Different needs, different scope is best. Just if take example. If need fast segmented memory acquisition, what is other scope what give max even near 500kwfm/s.

If want feature where when ever you stop scope you can go back up to 80000 normal currently used wfm/s speed acquisitions...  what is other scope what give tjhis...
If you work with low signals...  true 500uV/div full freq band...  it drops out immediately this DS1000Z
Also of course tehre is many opposite things.  There is not universal good oscilloscope.  This is why user need know what he need and what he do not need. Hobbyist think more that he want all because he do not know what he need. Professional may know exatly what is project, what need and what not and he can select scope what have good these things he need and not matter what ever other features are junk or crap or just do not exists. But also this, world is not only black or white.. some professional also may try find equipment what is good for all things what perhaps need some day or just for "nice to have but never use".

[Lightages]

Different needs, different scope is best. Just if take example. If need fast segmented memory acquisition, what is other scope what give max even near 500kwfm/s.

Exactly. If you need one function or feature, then get the scope that has it. In my general hobbyist needs, 4 channels outweighs most of the other features of the Siglent. The big, and I mean big, variable here, is that the Rigol has the freely available information on how to change it to the 100MHz version with all options. In reality this is an unfair fight because the Siglent loses to the Rigol over something many people feel is an immoral or unethical modification. I do not think it is and so to me the Rigol wins for my needs.

[tautech]

Both have advantages and disadvantages. No one can say which is universally better. If know user individual needs then it can tell which one is better for this user.

And so on.

But the list grows....
Individual channel controls
400V channel inputs
and so on...

But really, despite all this banter the X series products have just been released, Dave found a few simple bugs and rf-loop seems to have found several more, isn't that what pre-release units to dealers products are for?

The unit Dave reviewed is one of these, rf-loop's unit is one too and mine should be here tomorrow. 

But forget all this hype and tell us where they are available for immediate dispatch. 

[rf-loop]

But forget all this hype and tell us where they are available for immediate dispatch. 

I do not know.

I have heard rumors that there is units for the general sales already departed from factory and on the road to Siglent's own European wholesale warehouse.

Lets hope Siglent get well (enough) documented feedback about bugs, design errors and things what need develop from sellers demo units to repair and improve FW asap although  it is much more mature than SDS2k was in this phase.
Of course this is one function of demo units what may support also some kind of delta test purposes before public sell units but also these can use for limited demonstrations. And for seller education-training purpose. It is very important that seller is specialist who also have enough knowledge and experience about equipments what he is selling. (there is of course other kind of sellers who only know how to receive payment and how to send nice box to customer. but this is other case, we need support also these "box shipper" sellers customers free... (as long as there exist these "NiceBox shippers" as sellers )

[Stephan_T]

Both have advantages and disadvantages. No one can say which is universally better. If know user individual needs then it can tell which one is better for this user.

And so on.

But the list grows....
Individual channel controls
400V channel inputs
and so on...
[...]

Which one and for what CAT rating?

As I cited here, the Rigol Manual for the DS1000Z series states:

Chapter 18 Specifications

Maximum Input Voltage (1M?)

Analog Channel: CAT I 300 Vrms, CAT II 100 Vrms, Transient Overvoltage 1000 Vpk With RP2200 10:1 probe: CAT II 300 Vrms

So as far as I understand this, the safety level for mains voltages depends on the selection of the probe switch (1x vs. 10x). How does this compare to the Siglent?

One of the most important bits of wisdom that I learned from Dave here at the EEVblog is, that the superior quality of test equipment is not accuracy or features, but the level of confidence that it provides you.
Having problems with aliasing may be just a nuisance for experts, but it could be very misleading for beginners.

[EEVblog]

FYI, none other than the Siglent CEO will be visiting the EEVblog lab in December.
Siglent fanboys can get their questions ready.

[vlad777]

 I knew that this one will have great variable intensity for your 1MHz/1KHz AM test signal.
 Because it was about time.

 edit:  And it also catches all runts and 1inM glitches.

 BTW
    Do make a video about:  sample rate needs to be 4* frequency for sinx on x .

[rsjsouza]

So far one can nitpick its features and glitches, but overall it is configuring itself to be a very nice entry level oscilloscope. Hopefully Siglent will keep up with firmware updates.