Siglent SDS2000X Plus

[maginnovision]

For hobbyists I think they'd take the 4 channel rigol over the 4 channel siglent. 40% price difference. They should have gone the rigol route and just software locked the channels.

[supperman]

I think it is simple.. the hardware is the same on all models.

I think it is just marketing:

1. They don't want the 2000x plus to step on the 5000x but they also don't want to loose sales to Rigol.
2. They don't want to be in the upgrade business.. so this is probably a one off.. till a 500 comes out.
3. They are still working out software issues at 500mhz but feel confident the hardware can do it.

It is also possible (like with the 5000x) that initial hardware has some issues.. and they sort them into bandwidth buckets depending on how pronounced it is. This probably means that the 100Mhz can still do 450Mhz just fine if it does not make the cut.. If this is the case it will improve quickly as they work on production issues..

My 2 C's

[supperman]

For hobbyists I think they'd take the 4 channel rigol over the 4 channel siglent. 40% price difference. They should have gone the rigol route and just software locked the channels.

I think another reason the 100Mhz can be "upgraded" to 500Mhz... to cash in on the hobby market.. and beat out Rigol.

[thm_w]

Canadian pricing here:
https://www.transcat.ca/brand/siglent-store/oscilloscope
https://www.transcat.ca/brand/siglent-store/oscilloscope/siglent-sds2104x-plus-100mhz-4ch-digital-oscilloscope ($1447 USD)
Datasheet: https://www.transcat.ca/media/pdf/siglent-SDS2000X-Plus-Datasheet.pdf
Rigol is $999 or 909 for 2 probes.

So again our prices are higher than the Euro prices (1200 EUR = $1336 USD).

Interleaved, 2Gs/s with 2 channels, 1GS/s with 4 channels on. Which means a realistic bandwidth of ~400MHz for all four channels. Still not bad.

If you have the budget and want low noise (80uV rms noise at 500MHz), its good value, assuming it can be hacked.

[Martin72]

Rigols´benefits are the 8GS/s and the waste amount of memory or/and math functions, also using 4 math functions at the same time, also the 2 channel wavegenerator, also waste amount of trigger functions.
I´m waiting for the next fw update, should be come end of this month.
If it´s too disappointing, I´ll have a look to the 2000 plus...

[supperman]

I think it is simple.. the hardware is the same on all models.

I think it is just marketing:

1. They don't want the 2000x plus to step on the 5000x but they also don't want to loose sales to Rigol.
2. They don't want to be in the upgrade business.. so this is probably a one off.. till a 500 comes out.
3. They are still working out software issues at 500mhz but feel confident the hardware can do it.

It is also possible (like with the 5000x) that initial hardware has some issues.. and they sort them into bandwidth buckets depending on how pronounced it is. This probably means that the 100Mhz can still do 450Mhz just fine if it does not make the cut.. If this is the case it will improve quickly as they work on production issues..

My 2 C's

I spoke too soon.. looking at the data sheet gives us another clue.. You can upgrade up from any speed.. 100 to 200. 200 to 350 and 350 to 500mhz. No mention of any restrictions. There is actually a code to go from 100 to 350 for 2 channel scopes.
Did they forget to mention you can't get more than one upgrade.. or is it fine..

[supperman]

Bandwidth upgrade, 350 to 500 MHz, 4 channel models
$1,399.00

https://siglentna.com/product/bandwidth-upgrade-350-to-500-mhz-4-channel-models-2/

[tautech]

Bandwidth upgrade, 350 to 500 MHz, 4 channel models
$1,399.00

https://siglentna.com/product/bandwidth-upgrade-350-to-500-mhz-4-channel-models-2/

Yes but in comes with this limitation:
SDS2000XP-4BW05 Upgrade SDS2354X Plus bandwidth to 500 MHz (Only two channels)

Study reply #21 earlier in this thread.

[ResistorRob]

Finally after years of waiting someone came out with a scope that actually has the features I want at a price I can afford. The Rigol MSO5000 looked amazing but it's dim screen was a deal breaker. I heard this is fixed... can anyone confirm this?

I'm wanting a 100MHz, 4 Channel touchscreen scope. Between this new Siglent and the Rigol MSO5000 does anyone know what the key differences are? If it was your $1,400 which one would you buy? That is like 3 weeks of pay for me, so I need to get this right on the first purchase!

[thm_w]

Finally after years of waiting someone came out with a scope that actually has the features I want at a price I can afford. The Rigol MSO5000 looked amazing but it's dim screen was a deal breaker. I heard this is fixed... can anyone confirm this?

I'm wanting a 100MHz, 4 Channel touchscreen scope. Between this new Siglent and the Rigol MSO5000 does anyone know what the key differences are? If it was your $1,400 which one would you buy? That is like 3 weeks of pay for me, so I need to get this right on the first purchase!

Screen was never an issue for me and others. But without seeing it in person its hard to say if it works for you. It was improved.

If your budget is $1,500 then get this, if $1,000 then get the Rigol. I would also consider waiting until siglent has been hacked.
The main differences I can think of: Siglent has a lower noise front end,  10" vs 9" screen, 50 ohm. Rigol: hdmi, higher sample rate, dedicated channel knobs. But its going to depend on yourself to decide, don't get others to choose for you.

[supperman]

Bandwidth upgrade, 350 to 500 MHz, 4 channel models
$1,399.00

https://siglentna.com/product/bandwidth-upgrade-350-to-500-mhz-4-channel-models-2/

Yes but in comes with this limitation:
SDS2000XP-4BW05 Upgrade SDS2354X Plus bandwidth to 500 MHz (Only two channels)

Study reply #21 earlier in this thread.

That is a sample rate restrictions.. 2Gs for 2 channels.. 1Gs for all 4..

[thm_w]

That is a sample rate restrictions.. 2Gs for 2 channels.. 1Gs for all 4..

So how well do you think 1Gs/s is going to work at 500 MHz?

[rf-loop]

But what is the purpose of history mode? I've seen this on a Yokogawa oscilloscope and R&S has it as well but I don't really see a particular use for having a history. If you want to acquire segments then you might as well turn on segmented recording.

Every serious DSO has history mode – not only Yokogawa and R&S, but of course also Keysight and LeCroy, the latter even have WaveScan, a very powerful tool that can do much more than a simple search.

Segmented recording is a special mode for optimizing memory usage and minimizing blind time for events with irregular occurrence. It is set up on purpose and has no real time screen update.

No. The HP/Agilent/Keysight scopes I have owned & used so far didn't have history mode. By default they use full memory. And it isn't true segmented mode is purely intended to offer fast acquisition of repeating events. For that deep memory is equally useful because you won't save much memory anyway. Some DSOs even have a fast & slow segmented recording mode. In slow mode the progress is shown on screen. In fast mode you don't see anything until the number of segments if captured or press 'stop'.

"Some DSO,s even have...."   yes Siglent have.
Fast segmented acquisition mode and slow mode. Fast mode user need set and select for use (sequence acq).  Siglent slow segmented mode do not need turn on, it is always on and working background continuously (also it do not affect wfm/s speed  and it do not record image frames, it record true acquisitions with full ADC data and time stamps to segment memory (aka history) fifo).

Is it so that word "History" turn your blind spot on or what... only drawback is that this Siglent slow speed segmented acquisition mode (= normal mode ) do not show segment number when it is capturing and user can not set number of segments before stop. It run current run state maximum amount of segments fifo continuously.  (Perhaps Siglent can add this feature if there is enough demand. This is perhaps useful in cases where slow trigger events and want see segments on the screen before total is reached and handled.  Other ways need look empty screen some minutes or hours and only see segment counter progress.  )

Time ago one professional was looking new scope and I ask if he need also segmented memory or history function.
No, I do not need, also my old scope have not. I have never needed this kind of things.

Can I show you still what they are and some simple examples how to use these?

After one hour and tens of questions and some kind of explanations I ask if he still think these features are "no need" for his new scope.  Answer was clear, I do not want any new scope without least this slow sequence feature.

[nctnico]

What is your point? By your own definition history mode isn't a replacement for slow segmented recording.

[edigi]

So how well do you think 1Gs/s is going to work at 500 MHz?

The trouble that 1Gs/s does not work well even for 350 or for 200 MHz signals. I mean non-sinusoidal signals.
So while based on the period you may have the false impression that it will fit to the Nyquist limit but simply in many cases the spectral content of the signal can easily exceed that.
If that happens you have aliasing, and as a consequence what you see on the display is not what the signal looks like in reality.
(Having strong input filtering does not help much since that also removes spectral content thus distorts the signal).
E.g. how far would you go with a Square signal with 1Gs/s? Maybe somewhere 20-25 MHz? How fast signals modern MCUs or CPLDs/FPGAs can generate?

[nctnico]

So how well do you think 1Gs/s is going to work at 500 MHz?

The trouble that 1Gs/s does not work well even for 350 or for 200 MHz signals. I mean non-sinusoidal signals.
So while based on the period you may have the false impression that it will fit to the Nyquist limit but simply in many cases the spectral content of the signal can easily exceed that.
If that happens you have aliasing, and as a consequence what you see on the display is not what the signal looks like in reality.
(Having strong input filtering does not help much since that also removes spectral content thus distorts the signal).

You are going down the wrong path here. If you want to look at a signal with high frequency components you need an oscilloscope with enough bandwidth. Otherwise the anti-aliasing filter will chop off the higher frequency components and show what is left. But either way you always need to think about how the oscilloscope + probe will affect the signal you are looking at so you are not chasing ghosts.

[rf-loop]

@Performa01: Does Siglent have any plans to get rid of the automatic memory depth setting? For me the automatic memory depth adjustment is a total productivity killer. It annoyed the hell out of me during the short time I had the SDS2000.

The short answer is no, sorry.

What engineers want to do is capture a record that represents a certain timespan. What is easier and feels more natural:

•   Select the main timebase for the required total timespan and then adjust the zoom timebase for closer inspection of details
or
•   Calculate the required memory depth from the required timespan and current sample rate, dig into the Acquisition menu to select the appropriate memory depth (provided it could be selected with similar fine granularity as the timebase) and then have to scroll back and forth through the buffer in order to inspect the entire acquisition
?

You (and Siglent) forgot one very simple use case: trigger on part of a signal and when that looks OK change the timebase to look at the rest of the signal (which was outside the screen during capture). I use this regulary during design verification. And this is much easier to achieve than you think: just set the oscilloscope to the maximum memory depth and leave it at that. I have my oscilloscopes set to the maximum memory depth 99.9% of the time. It is very handy to be able zoom out or pan left/right to examine the signal which was outside the screen. With Siglent's automatic memory depth I'd need to change the time base for each capture twice which is just time consuming and irritating. And why not have the maximum memory depth available? Otherwise the memory just sits there idle anyway.

First this "I have my oscilloscopes set to the maximum memory depth 99.9% of the time."
With 200Mpts and 2GSa/s one acquisition takes 100ms. Are you still sure you want keep oscilloscope nearly always, 99.9%, set for max memory depth.  Why need zoom out, also you can use it (zoom out) only after stop in certain scopes.
If I need this kind of, I will turn zoom mode on and look whole memory and details at once and pan zoomed window where ever need. Most of scopes do not even loose vertical resolution using this method. It is just better to learn and adapt for use windowed zoom as "normal" mode. And more when also scopes start get more vertical height and resolution screens.

I do not want scope what have lot of visual blind time, hidden data,  at runtime. If I do work what need long capture and still details I turn windowed zoom on.  I remember well this time when example Owon did not even have windowed zoom, it have only full screen zoom alone. I have never used so frustrating system. And your working method is just like using full screen zoom, your runtime window is zoomed part of whole acquisition and you can not see rest at all in runtime, you need stop scope. And most captured trace out from screen and then full hassle to find some things there and always stop scope and pan and zoom.
Using scope where whole capture is always visible without overlapp and then zoom window for detail. You have always knowledge about whole capture and where your details window is related to whole trace and this all can do runtime and stopped.
Why need think all need  do as always before have done. Human can learn new modes of operations. But also of course one may like tulips and one may like fish.  And yes, I remember this first reaction after I have adapted this other system what you expolain... it was first some "shock" when I meet this Siglent principle. How hell I can do my works now, how I can find rest of traces... I zoom out and no...nothing...  how this can use. Perhaps this is some kind of bug...or what hell. (There was also full memory available outside of screen in Siglent older models, also as in many many other scopes.
But after then I start use it and today I do not at all miss this method at all.. more like vice versa, I wonder why many still do these. But it need learn so that it is part of your muscle memory and thinking process. Just like changing from algebraic calculator to calculators rolls royces, HP RPN and after familiar, then do not want use any others anymore. But I know, there is peoples who can not or want not use or learn to use RPN.

Lets hope Siglent stay calm and continue this selected road.
What is captured need show and decode (except if decode max data amount is reached.) You many times ask why Siglent do not decode full memory. (It do but somehow you don not want accept it)  But I can ask, why many scopes mostly do not display full currently used memory length at runtime... when they start make scopes what can display all what is captured and runtime, wit details.

With these user need stop scope for look this data. With Siglent whole data is visible and can look also runtime, including details, even when capture length is hundreds of megasamples.

Also always bacround working slow segmented acquisition is part of this total. If we always capture your full memory length there is no enought amount enough fast memory and processing power for history buffer. Of course data can cut and use only displayed part

[Fungus]

I think it will be an interesting battle!  Of course you should compare with 5074 since it's the lowest we can get.

If the Siglent isn't hacable then it's going to be MUCH more expensive for equivalent features to a hacked Rigol.

[rf-loop]

So how well do you think 1Gs/s is going to work at 500 MHz?

The trouble that 1Gs/s does not work well even for 350 or for 200 MHz signals. I mean non-sinusoidal signals.
So while based on the period you may have the false impression that it will fit to the Nyquist limit but simply in many cases the spectral content of the signal can easily exceed that.
If that happens you have aliasing, and as a consequence what you see on the display is not what the signal looks like in reality.
(Having strong input filtering does not help much since that also removes spectral content thus distorts the signal).
E.g. how far would you go with a Square signal with 1Gs/s? Maybe somewhere 20-25 MHz? How fast signals modern MCUs or CPLDs/FPGAs can generate?

Oscilloscopes BW is defined for sinewave.
100MHz non sinusoidal signal is not 100MHz sinusoidal signal and oscilloscope for 100MHz BW is not for this (depending need of accuracy and true signal).
This have been clear to all educated and or experienced oscilloscope users tens of years. As long as I can remember backwards. But still this kind of talk pops up frequently and continuously. Why. 100MHz what is not 100MHz and then 100MHz BW is not enough...
Who do not still know what is oscilloscope BW "definition" and about digital scopes who do not still understand what Nyquist and Shannon have told and also what HPs Chris Rehorn have told about digital scopes.  Even with old time with analog scopes we know what is oscilloscope BW meaning. Still some wonder why 100MHz scope can not display 100MHz square waves.
1GSa/s is enough for 350MHz sinewave with normal fast Sinc interpolation when we are talking about normal use of scope with its accuracy class. It is some times  not at all enough even 1Hz square wave, some times 1GHz BW is not enough for 1Hz square and we need talk what BW need if need measure real fast risetime with some good accuracy ... it may surprice.

[nctnico]

Lets hope Siglent stay calm and continue this selected road.
What is captured need show and decode (except if decode max data amount is reached.) You many times ask why Siglent do not decode full memory. (It do but somehow you don not want accept it)  But I can ask, why many scopes mostly do not display full currently used memory length at runtime... when they start make scopes what can display all what is captured and runtime, wit details.

You and Performa4 keep missing the point that the way Siglent works makes operating the scope extremely cumbersome. I will never ever buy an oscilloscope which can't be set to use full memory all time. Same for decoding only what is on screen. And I'm 100% sure that once I show you two the benefits you will fully agree with me (just like your 'customer' initially thought segmented recording wasn't needed). It just takes looking outside the Siglent box.

[edigi]

If you want to look at a signal with high frequency components you need an oscilloscope with enough bandwidth. Otherwise the anti-aliasing filter will chop off the higher frequency components and show what is left.

That anti-aliasing filter doesn't have too much room between 350 MHz and 500 MHz so it must be pretty good, right?
So what frequency square signal would you go max. with this setup?

[edigi]

It is some times  not at all enough even 1Hz square wave, some times 1GHz BW is not enough for 1Hz square and we need talk what BW need if need measure real fast risetime with some good accuracy

Sure, rise time determines the needed bandwidth but let's try to be a bit more practical here and assume that the user wants to check signals on MCUs, CPLDs/FPGAs so just everyday stuff.

[tv84]

I think it will be an interesting battle!  Of course you should compare with 5074 since it's the lowest we can get.

If the Siglent isn't hacable then it's going to be MUCH more expensive for equivalent features to a hacked Rigol.

ALL Siglent models are as "upgradeable" as Rigol's. If that wouldn't be the case then we should not compare them with a "hacked Rigol".

Nevertheless, it seems Siglent is pretty confident in its product given the price it chosed to it's starting model 2104X+ in comparison with the 5074.

I say that it will be "interesting times" because all will have plenty of factors to discuss once again in the bang/buck realm. You know pretty well the DS1004Z / 10004X-E war. Now, it will be the 5000 / 2000X+ war...

For Rigol's sake, it better release a good FW in the near future...

[nctnico]

If you want to look at a signal with high frequency components you need an oscilloscope with enough bandwidth. Otherwise the anti-aliasing filter will chop off the higher frequency components and show what is left.

That anti-aliasing filter doesn't have too much room between 350 MHz and 500 MHz so it must be pretty good, right?
So what frequency square signal would you go max. with this setup?

For a square wave (=average digital signal / clock) it is nice to have the fundamental + 2 harmonics (3x and 5x fundamental frequency) visible in order to asses things like signal levels. Fundamental + 1 harmonic is cutting it close. So a 350MHz oscilloscope should be usable on a 100MHz clock but from 70MHz and onwards you won't be able to make much of the shape. If you don't care much about the shape (for example only the frequency or time relation with another signal) then you can go all the way up to 350MHz.

[Martin72]

For Rigol's sake, it better release a good FW in the near future...

Definitely, otherwise I´ll change back to siglent.

Funny thing:

https://www.batronix.com/versand/oszilloskope/Siglent-SDS2104X-plus.html#yoReviews

Have a closer look to the display, the picture must be from the 5000 - or it´s magic having 2.5GSa/s on a 2GSa/s scope….