Author Topic: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown  (Read 25826 times)

0 Members and 1 Guest are viewing this topic.

Offline CharlieEcho

  • Regular Contributor
  • *
  • Posts: 59
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #75 on: January 21, 2018, 05:30:20 pm »
Whelp, these are dangerous waters for a first post! I've read this and the related threads with great interest.

I'm very interested in comparison of the Siglent 1104x-e v. Rigol 1000z options v. Micsig TO1104 head to head. I wonder if the DS1074Z Plus or DS1074Z-S Plus might be the best deals on the Rigol side right now? Those are the configurations that support the MSO upgrade and also have function generation, respectively.  (My understanding is that the Rigol function generator is a factory hardware option, as is the presence of the port for the MSO harness.) Presuming the Riglol hacks, the Siglent plus the external signal generator is within $50 of the DS1074Z-S Plus hacked to full BW (Amazon pricing). 

I'd also be interested in comparison to a  Keysight DSO1024A as a reference, which can be found used w/a limited warranty for around $800 after S/H, tax etc in the US right now. And I suppose at that point, the Siglent 11204x-e as well which is around the same price from Saelig. Perhaps that's the best absolute value if one's budget can stretch? All of these choices are within reasonable reach of the others, although dongles like the Siglent generator can be purchased later to spread out the cost, and of course adding Keysight options is prohibitively expensive.

Possibly there's a practical advantage for rank beginners with Keysight in that a lot of downloadable course material contains procedures for either Agilent/Keysight or Tek control panels & menus. And along the same lines, there might be a practical advantage for the Micsig is that it makes automotive tinkering or field use much more inviting. The latter being interesting to me as an inactive ham operator looking to get into back into that hobby with mobile SDR and embedded IOT radio stuff.

I guess there's the Instek too, but at a glance it seems to offer less overall. Possibly I'm being suckered by shiny things and community enthusiasm for Rigol and Siglent as a newbie. I haven't looked further.

A final thought: might the current Rigol promotion suggest a new series is coming after March? (But if so, it probably won't be as hackable, one would expect.)  :phew:
« Last Edit: January 21, 2018, 06:49:48 pm by CharlieEcho »
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #76 on: January 21, 2018, 06:23:24 pm »
I think that pretty much any FPGA based scope is hackable.  Most of  the T&M OEMs are using the embedded Linux toolchain provided by the FPGA vendor.
The only question is how much work is it.  My Rigol DS1102E does not appear to be running Linux, but I haven't really probed it very thoroughly.  I know that Siglent is using embedded Linux in the SSA-3000X, but don't know about the scopes.

I had the good luck to buy an Instek GDS-2072E from Amazon for $222 plus tax expressly for use as a test system for developing open source software for the GDS-2000E line.  I do NOT intend to replace all the factory software. My intent is to be able to fix bugs and add features.  So if a binary blob works, it stays untouched.  I have an MSO-2204EA as motivation and tool for the task and lots of *nix experience.

A *very* cursory inspection of the FW update files for the GDS-2000E suggests it is not difficult to create custom FW updates.
 
The following users thanked this post: CharlieEcho

Offline CharlieEcho

  • Regular Contributor
  • *
  • Posts: 59
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #77 on: January 21, 2018, 07:02:52 pm »
Yes, but that's *real* hacking. Beyond my capabilities.

 As opposed to an easy keygen that Rigol seems content to ignore. We (probably?) won't see that in a new series of Rigol scopes aimed at Siglent's offerings. Although I suppose we might just see more features included, or routinely available through promotions instead. The current 1000z series Rigol promotion is amusing that way: they're throwing in all the features w/full licensing since they'll be unlocked anyway.

Well, except the BW upgrade.  :-/O

Given the cost of the AFG options for both Rigol and Siglent, a fully standalone alternative seems like a better buy. That leaves the MSO addons / serial decoding as the other option to compare --  my understanding is that  Siglent's serial decoding has been observed to be much more responsive than Rigols  (https://www.eevblog.com/forum/testgear/open-source-lxi-tools-and-liblxi-v1-0-released-for-gnulinux/msg1394796/#msg1394796), although I don't know if that performance is unique to the 4ch 200MHz Siglent.

In a vacuum, that would suggest that the upgraded/updated Siglent is the way to go -- but again one can now snag a Keysight at the same price. Though without a full warranty, and fewer bells and whistles (if I'm reading all the product info and etc correctly). And of course smaller screen and older USB. Plus, while I don't see much discussion of that model, my takeaway from other teardowns of recent budget Keysight gear suggests it's rebadged gear anyway.  :-//
« Last Edit: January 21, 2018, 07:22:14 pm by CharlieEcho »
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 26751
  • Country: nl
    • NCT Developments
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #78 on: January 21, 2018, 07:22:37 pm »
It depends a bit on the intended use (hobby or commercial environment) and how much money you want to spend. IMHO Siglent and Rigol are not the brands to look at if you need a device which just works. Spending a bit more on a Keysight, Tektronix, R&S or a GW Instek will get you a device which works as advertised.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 28136
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #79 on: January 21, 2018, 07:23:32 pm »
Given the cost of the AFG options for both Rigol and Siglent, a fully standalone alternative seems like a better buy.
Nice to see someone who's done a good amount of homework.  :)

Yes, I've always thought a standalone the way to go if you want a fully featured AWG. Of course you also have more output capability and second channel. Plus if you go down the time nut path, the standalone's allow for a 10MHz reference input.

Quote
That leaves the MSO addons / serial decoding as the other option to compare --  my understanding is that  Siglent's serial decoding has been observed to be much more responsive than Rigols  (https://www.eevblog.com/forum/testgear/open-source-lxi-tools-and-liblxi-v1-0-released-for-gnulinux/msg1394796/#msg1394796), although I don't know if that performance is unique to the 4ch 200MHz Siglent.
Processor speed, pure and simple.

Quote
In a vacuum, that would suggest that the upgraded/updated Siglent is the way to go -- but again one can now snag a Keysight at the same price. Though without a full warranty, and fewer bells and whistles (if I'm reading all the product info and etc correctly).  :-//
The only thing I think you may have overlooked is the internal architecture, that is, the 4ch SDS1*04X-E has 2x 1Gsa/s ASCI's so in any usage configuration the lowest sampling is 500 Msa/s.
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 

Offline CharlieEcho

  • Regular Contributor
  • *
  • Posts: 59
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #80 on: January 21, 2018, 08:07:10 pm »
Appreciate the responses!

I'm far from having commercial requirements, my interest in the Keysight is mostly because it's available at the same price as the 200MHz Signlent, and that seems like a recent phenomenon. It's hard to tell from Keysight's promotional materials just how equivalent the standard features are, which is what you get for the price. But more digging on the forums here suggests that the DSO1024 is a Rigol DS1204B re-badge, so at that point I'd think I'd be better off with the more recent Rigol or Siglent in widespread use by hobbyists...

It's news to me that GWI scopes are regarded as being closer in quality to A/K or Tek than Rigol or Siglent but it's certainly something to keep in mind.

Ignoring the signal generation but retaining the MSO options means the point of comparison is the DS1074Z Plus (if I'm correct that you have to buy that factory model for hardware compatibility with the software upgrade option and the external interface). So around $520 vs. $500 for the 100MHz Siglent, or $750 for the 200 MHz. (See below.)

It can all be a bit bewildering, heh. I think I'm leaning towards the upgraded Siglent, and then maybe the tablet on a holiday sale down the road when I've learned enough for real-world experiments (presuming it continues to get favorable reviews), but I'll probably try to make myself wait for a few more reviews and comparisons to show up before I decide.
« Last Edit: January 21, 2018, 08:25:49 pm by CharlieEcho »
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16560
  • Country: 00
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #81 on: January 21, 2018, 08:21:32 pm »
Yes, but that's *real* hacking. Beyond my capabilities.

 As opposed to an easy keygen that Rigol seems content to ignore.

nb. The Rigol MSO/Plus models don't work with the easy keygen, only the base models do (ie. the plain DS1054Z).
 
The following users thanked this post: CharlieEcho, KungFuJosh

Offline CharlieEcho

  • Regular Contributor
  • *
  • Posts: 59
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #82 on: January 21, 2018, 08:24:16 pm »
Ah. Well, that's good to know! Appreciate it. That would make the DS11094Z plus the point of comparison for a Rigol that will operate at 100MHz and support the MSO upgrade. $670 list. At which point the appeal of the 200MHz Siglent is hard to ignore.

One more note, for anyone else weighing the same decision - according to this Keysight comparison chart - the Rigol's don't do CAN or LIN -- not something I saw highlighted elsewhere, although maybe I just tuned it out.

Quote
Nice to see someone who's done a good amount of homework.  :)

If I have a clue it's because of everyone's excellent blog posts and video reviews and the extensive discussions here.
« Last Edit: January 21, 2018, 08:37:24 pm by CharlieEcho »
 

Offline edgelog

  • Regular Contributor
  • *
  • Posts: 70
  • Country: se
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #83 on: January 21, 2018, 08:34:07 pm »
nb. The Rigol MSO/Plus models don't work with the easy keygen, only the base models do (ie. the plain DS1054Z).

Yes, they do (at least the MSO), with a slight change in the search string that rigup uses.
 

Offline CharlieEcho

  • Regular Contributor
  • *
  • Posts: 59
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #84 on: January 21, 2018, 08:42:10 pm »
Hmm. I will do more reading.

Although the additional serial decode modes may tip the scales for me. In which case, it's the Keysight with its smaller screen,no network support, but deeper sample rate  (2G/1G vs 1G/5M) or the Siglent with its larger screen, new internals, wifi & ethernet, and additional upgrade availability... but half the sample rate... :scared:
« Last Edit: January 21, 2018, 09:10:50 pm by CharlieEcho »
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 26751
  • Country: nl
    • NCT Developments
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #85 on: January 21, 2018, 09:26:57 pm »
Samplerate is least important. It is fine for as long as the samplerate is at least 2.5 times the bandwidth on a modern DSO.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 
The following users thanked this post: CharlieEcho

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #86 on: January 22, 2018, 12:33:49 am »
I bought one of Leo Bodnar's 40 pS  rise time pulse generators.  So just to be mean and further confuse the OP, here is a plot of the 10%-90% rise time of a plain vanilla Instek GDS-2072E. 

I bought the Instek from Amazon during one of their price reductions for $222 + tax. It is running the V 1.24 FW.  I have not made any changes.  I bought  it to test creating an open source software stack without endangering my MSO-2204EA which cost 5x as much.  But I'm still reorganizing my work area to accommodate an influx of new gear.  So the 2072E has only been powered on for an hour or two.

The measured rise time of 2.3 nS corresponds to a BW of 152 MHz, over twice the 70 MHz spec. The MSO-2204EA has a rise time of 1.70 nS which is 204 MHz, just slightly above spec.  The plot is with averaging of 256 samples.

FWIW the FFT on the Instek GDS/MDO/MSO-2000E line is quite good.  It could be better which is one of my goals.  The Rigol FFT is too short to be useful.  I can't say about the Siglent.

Leo's pulse generator comes with a plot on a Tek CSA803 showing the actual rise time of the unit you bought.  Mine is 36 pS.  For the price it is a hell of a good deal.
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #87 on: January 22, 2018, 02:04:36 am »
The measured rise time of 2.3 nS corresponds to a BW of 152 MHz, over twice the 70 MHz spec. The MSO-2204EA has a rise time of 1.70 nS which is 204 MHz, just slightly above spec.  The plot is with averaging of 256 samples.

I have noticed before that these cheap DSOs usually or always have a bandwidth which varies considerably at different vertical attenuator settings; I suspect the specifications give the worst case value.  But testing this requires a higher voltage source and a set of RF attenuators.
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #88 on: January 22, 2018, 02:37:14 am »
A s I understand a comment by Leo, a fast rise time pulse generator over 5-6 V is an ITAR regulated device.  I have  3-4 Pomona 6 dB attenuators, so I can go down, but no convenient way to go up.

I'm sure that R&S, Tek and Keysight sell instruments which meet the top end spec and are software limited.  I'm skeptical that is true for low end Chinese stuff. I suspect that if you tested a bunch of BW "hacked"  gear you'd find that it didn't meet the higher spec.  I've seen assertions about the results of "hacking", but I've never seen a rise time plot or similar.

I do have a couple of Tek 106s.  I've not tried to use one in 10-15 years, so no idea if they still work.  Also no idea where the adaptor might be.  Last time I used one was doing a post repair cal on a Tek 465 and a Dumont 1060.

[edit]  Tektronix held the analytic solution of the effects of large inputs on the attenuator section as a trade secret for a long time.  I'm not willing to go on a hunt the wumpus to find the paper, but as I recall, it was a really nasty series of Bessel functions having to do with heating changing the component values..  In any case, anyone who thinks the attenuator section of a high speed scope is easy to build doesn't have a clue.
« Last Edit: January 22, 2018, 03:18:58 am by rhb »
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #89 on: January 22, 2018, 03:23:26 am »
A s I understand a comment by Leo, a fast rise time pulse generator over 5-6 V is an ITAR regulated device.  I have  3-4 Pomona 6 dB attenuators, so I can go down, but no convenient way to go up.

I know someone who could not import some 40+ year old Tektronix gear to the US where it was made because it had tunnel diodes and he mentioned this in the documentation.  He was after the tunnel diodes to fix other Tektronix gear.  Even if ITAR does not cover it, customers officials are more likely to error on the side of prohibition and convincing them otherwise is expensive even when they are wrong.  They are not paid to not block imports and exports.

At higher voltages, the common fast reference edge generation circuits are much slower but of course 300ps is plenty fast for 100 to 200 MHz testing.  But still anything above 5 volts is tough and most are still only up to 1 volt.

Quote
I'm sure that R&S, Tek and Keysight sell instruments which meet the top end spec and are software limited.  I'm skeptical that is true for low end Chinese stuff. I suspect that if you tested a bunch of BW "hacked"  gear you'd find that it didn't meet the higher spec.  I've seen assertions about the results of "hacking", but I've never seen a rise time plot or similar.

The only example I have seen online where a hacked Rigol DS1054Z was tested properly showed a nonlinearity in the edge which I suspect came from saturation or cutoff of the preamplifer so the transient response test showing greater than 100 MHz bandwidth revealed a real problem which I would have completely failed an oscilloscope for having.

I do not see anything like that in your test but that 2ns long dip immediately after the edge is not right.  Below is an example of bad transient response in a 100 MHz Tektronix 465 which likely just needs calibration and almost good transient response in a 100 MHz Tektronix 2232.
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: EEVblog #1042 - Siglent's $499 SDS1104X-E 4CH Oscilloscope Teardown
« Reply #90 on: January 22, 2018, 03:36:42 am »
I'm planning to do an FFT of the response.  I'll have to digitize the trace Leo sent me by hand to do a proper job of extracting the transfer function, but that 4 nS wobble is an anomaly at 250 MHz.  So still damn good for a cheap scope.  At $222 + tax, an exceptional deal. I sort of feel guilty, but I'm hoping that the results of my efforts will make up for depriving someone else of the deal.

But I very much agree about how hard it is.  I got handed my head on a platter when I tried to build a fast rise piulse long ago.

Fixing the Tek and the Dumont was some of the best education I ever got.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf