New Siglent SDS1202X-E oscilloscope based on Xilinx Zynq-7000 SoC architecture

[Fungus]

It takes far more rigorous and methodical testing than twiddling a few knobs to find the problems. Taking a scope through all it's functions and testing their limits takes a few days at least if you are setup for this kind of testing. You can't even capture that into a video. It would be several hours long!

Wow, you're making me feel pretty special. I found one in the first 5 minutes! And it's only 11 pages into this thread!

Nah, that just means you found one of the easy ones. There will be several that take months to find.

[zike]

[/quote]
Nah, that just means you found one of the easy ones. There will be several that take months to find.
[/quote]

Sorry, new to this forum; do we explain jokes?

Here's another fun one: "set trigger at 50%" (push trigger level knob) doesn't work correctly if trigger input coupling is set to AC or LF Reject (just puts trigger at 0.00V, irrespective of input). Kinda useless.

I'll save the next one a couple weeks so it looks like I worked for it.

BTW I found a wiring error in my Tek 545B, should I post it?

[ebastler]

Here's another fun one: "set trigger at 50%" (push trigger level knob) doesn't work correctly if trigger input coupling is set to AC or LF Reject (just puts trigger at 0.00V, irrespective of input). Kinda useless.

Could you explain that joke please? I don't get it... 


Seriously -- where else would you set a default trigger threshold in AC coupling mode?

[zike]

Could you explain that joke please? I don't get it... 
Seriously -- where else would you set a default trigger threshold in AC coupling mode?

About halfway between signal maximum and signal minimum. Same as for DC coupling. Just think of a low duty-cycle pulse train.

[tautech]

Any bugs found need be well described and with screenshots for us to pass onto the engineers.
They need to be able to replicate a bug so the more info the better.
Firmware version need be stated too please.

Post them here and/or in this thread for them to be passed on:
https://www.eevblog.com/forum/testgear/siglent-technical-support-join-in-eevblog/

[nctnico]

Here's another fun one: "set trigger at 50%" (push trigger level knob) doesn't work correctly if trigger input coupling is set to AC or LF Reject (just puts trigger at 0.00V, irrespective of input). Kinda useless.

Could you explain that joke please? I don't get it... 

Seriously -- where else would you set a default trigger threshold in AC coupling mode?

Well it would fail on a very asymetric pulse so yes, it is a bug! Needless to say I just tried it on both my MicSig and GW Instek scopes with a 1Vpp 0.1% duty cycle square wave. Both set the trigger point to 0.5V with trigger and channel set to AC.

@Tautech: maybe start a seperate bug thread just like we have for the Rigol DS1054Z?

[tautech]

@Tautech: maybe start a seperate bug thread just like we have for the Rigol DS1054Z?

Yep, right away.
Will we need two threads like for the 54Z ? 

[ellrod]

This may be obvious but does rolling, zoom, and history work for dual channels as well as single channels ? I see lots of single channel demos and pictures but very few dual channel examples.

[tautech]

This may be obvious but does rolling, zoom, and history work for dual channels as well as single channels ? I see lots of single channel demos and pictures but very few dual channel examples.

Welcome to the forum.

As there a few of these units out there now maybe a new owner can post a screenshot for you, if they don't I'll put something up but later....lots to do. 

[zike]

Good point, thanks for reminding us.

Software version: 5.1.3.8
FPGA version: 2017-03-21
Hardware version: 00-01
Product type: SDS1202X-E

Really appreciate your interaction here,

Zike

Any bugs found need be well described and with screenshots for us to pass onto the engineers.
They need to be able to replicate a bug so the more info the better.
Firmware version need be stated too please.

Post them here and/or in this thread for them to be passed on:
https://www.eevblog.com/forum/testgear/siglent-technical-support-join-in-eevblog/

[rf-loop]

This may be obvious but does rolling, zoom, and history work for dual channels as well as single channels ? I see lots of single channel demos and pictures but very few dual channel examples.

Of course all works also with two channels: Waveform history buffer, Segmented memory acquisition, rolling, zoom. Difference is that samplerate is half and memory is half (max 500MSa/s for both channels simultaneously and max 7M for both channels simultaneously.)
FFT it can run only for one channel but of course also then two channels can be in use but user select  Ch1 or  Ch2 for FFT. If use "Split Screen" or "Full Screen" mode then both channels in time basis can see together with selected channel FFT (also if two channels is in use FFT max length is 10M sample.) If use "Exclusive" mode then full height FFT alone is displayed.

Then about this channels input impedance selection. Scope do not have internal 50 ohm but it have 50ohm/1M selection in menu.   This selection must NOT remove. This selection is needed or can use example  for FFT  so that user can use external 50ohm and this need also tell to scope system that user is using 50 ohm termination. So that measurement can do also for power units (dBm @ 50ohm)

I Hope they repair 50 ohm setting in menu system so that if user use external 50 ohm termination it also can use levels as dBm what is common and mandatory in every single spectrum analyzer.

[Glowtape]

So I got one of these things to debug my RC stuff and all. First thing I noticed is that the voltage level differential in the cursors doesn't scale with probe attenuation. I have them set to 10X and what measures otherwise like 3.8V over here shows as delta-Y of 380mV. Nice one.

[zike]

Re: phony 50 input select yes, it would be wonderful if the thing could stand for a spectrum analyzer, but that would take a lot more than convenient display units. The FFT mode correctly offers dBVrms units. If you know that you have installed an external terminator (and why), you should also be able to add 13. In the meantime, channel input "Impedance" select is clearly indicating real physical input impedance, and is simply wrong.

Couple more strange behaviors other owners may wish to confirm:

- In "math" you can select channels 1 and 2 as operands ("Source") A and B. But if you scroll to select operations it *sometimes* spontaneously changes both operands to channel 2.  When it does this, the soft keys at screen bottom still display the channel assignments originally selected, but the on-screen "MATH" text shows the erroneous operation. Based on signals I fed in,  that indeed appears to be what it's displaying. Manually re-selecting the channel for one of the Sources appears to resolve it for the moment; but if you then change the math operator again, it's a crapshoot.

- The manual (page 133) indicates one can do math against internally saved Reference waveforms. But only channel 1 and channel 2 are offered as operands, even if Ref A and Ref B have been saved with valid waveforms.   

- Under "acquire" menu page 2/2 there is a softbutton "Acq Mode" appearing next to the "Interpolation" select. This gives two options, "Slow" and "Fast". "Slow" appears to make the screen update jerky.   Nothing shown in the manual (although mystery button does appear in Figures 12 and 13).  Onscreen help says:  "Fast: Enable hardware accelerated view. Slow: Disable hardware accelerated view."  Any clues why this option is offered, or how it might be useful?

Once again:
Software version: 5.1.3.8
FPGA version: 2017-03-21
Hardware version: 00-01
Product type: SDS1202X-E

[boggis the cat]

Do you own this scope?

Not yet.  What 'scope did your "potential customer" choose?

[rf-loop]

Re: phony 50 input select yes, it would be wonderful if the thing could stand for a spectrum analyzer, but that would take a lot more than convenient display units. The FFT mode correctly offers dBVrms units. If you know that you have installed an external terminator (and why), you should also be able to add 13. In the meantime, channel input "Impedance" select is clearly indicating real physical input impedance, and is simply wrong.

Yes, how it is now is of course wrong. Who have told it is not wrong. No one.
I hope they add user selectable 50ohm setting for display dBm @ 50ohm with FFT.  Of course all educated and experienced users know why.  Do you think it is wrong when SDG can set for Hi-Z or user selectable load impedance. Not more complex to use 50 ohm / 1M selection than set probe multiplier.  Butr this need name and/or other way arrange so that user clearly understand there is not internal 50 ohm. (if these kind of  equipments settings need be monkey protected, how hell they can set probe multiplier....if not this)

What is real pity is that this model do not have 50 ohm internal circuits as example "big brother" 1000X have and it have dBm, Vrms and dBVrms selections (depending input selection) But big brother also do not have  frequency scale bug as X-E now have. Let's hope this scale error in X-E is repaired asap. Edit (  ) : This need more investigation = perhaps my fault or it happen only in some very special situations .

[GeorgeOfTheJungle]

Ufffff, is the X-E another case of a not fully cooked yet firmware?

[zike]

Ufffff, is the X-E another case of a not fully cooked yet firmware?

I'm an old fart, so I don't know customs nowadays; the process is now very public, and "the sky is falling" seems to be a way to get attention. Certainly it's no novelty that marketing ships a product before engineering has quite had time to get the bugs out. 'Twas ever thus;  I've restored plenty of classic HP and Tek gear where the "change & revision notices" section is thicker than the original service manual.

So far I'm impressed with the update speed, user interface (well, except buggy FFT/math...) and price:performance ratio. Some solid and innovative engineering in there, for sure. The manual isn't even that bad, a few errors and lots of translation artifacts but mostly clear and to the point.

Whether the growing bug list is truly "excessive" for so new a product is not so clear.  The answer is in how Siglent responds. Have they been notably unresponsive in earlier product rollouts? Again, I'm new to this ecosystem...

Zike

[nctnico]

Whether the growing bug list is truly "excessive" for so new a product is not so clear.  The answer is in how Siglent responds. Have they been notably unresponsive in earlier product rollouts? Again, I'm new to this ecosystem...

Siglent is notorious for releasing products with incomplete firmware and then taking years to fix the problems. Chances are not all problems get fixed. If there are problems in your scope you can't live with then return it and don't get fooled into waiting until Siglent comes up with a fix. There is a reason Siglent is so cheap...

[tautech]

Ufffff, is the X-E another case of a not fully cooked yet firmware?

I'm an old fart, so I don't know customs nowadays; the process is now very public, and "the sky is falling" seems to be a way to get attention. Certainly it's no novelty that marketing ships a product before engineering has quite had time to get the bugs out. 'Twas ever thus;  I've restored plenty of classic HP and Tek gear where the "change & revision notices" section is thicker than the original service manual.

So far I'm impressed with the update speed, user interface (well, except buggy FFT/math...) and price:performance ratio. Some solid and innovative engineering in there, for sure. The manual isn't even that bad, a few errors and lots of translation artifacts but mostly clear and to the point.

Whether the growing bug list is truly "excessive" for so new a product is not so clear.  The answer is in how Siglent responds. Have they been notably unresponsive in earlier product rollouts? Again, I'm new to this ecosystem...

Zike

Thanks Zike for your report and words of support for the X-E.
Some minor shortcomings are not unexpected in a new product these days and IMO despite all the hype and speculation this model has built for some months (first seen at Hamburg show last year) most operations seem very functional. Indications that there's been a few FW revisions before release are in the FW #: 5.1.3.8.
All the brands these days seem to offer new products to the marketplace with something overlooked, yes all brands.

We'll link the engineers to your findings for them to resolve the issues and if you find anything else please offer it here or in the Siglent support thread.

In reply to your responsiveness question, the fastest bug remedy I've seen was one reported by a member here, I sent it to the factory same day and in 2 weeks it was fixed in the next FW release.

[ebastler]

In reply to your responsiveness question, the fastest bug remedy I've seen was one reported by a member here, I sent it to the factory same day and in 2 weeks it was fixed in the next FW release.

That qualifies as a partial answer, with an ever-so-subtle bit of spin put on it.
What was the slowest response you have seen?

[tautech]

In reply to your responsiveness question, the fastest bug remedy I've seen was one reported by a member here, I sent it to the factory same day and in 2 weeks it was fixed in the next FW release.

That qualifies as a partial answer, with an ever-so-subtle bit of spin put on it.
What was the slowest response you have seen?

I try to follow up with the factory on all bug reports but life sometimes gets in the way and I overlook some until I am reminded about them. When this happens I send another request with the date that I sent the first and this normally gains results. With popular products it does normally take very long.
Some indication can be seen in the frequency of new FW releases on the Siglent FW webpage.
http://siglentamerica.com/gjjrj.aspx?id=15&page=2

[nctnico]

In reply to your responsiveness question, the fastest bug remedy I've seen was one reported by a member here, I sent it to the factory same day and in 2 weeks it was fixed in the next FW release.

That qualifies as a partial answer, with an ever-so-subtle bit of spin put on it.
What was the slowest response you have seen?

Make that a lot of spin! IIRC that was about a spectrum analyser. I think the slowest response must be the Siglent SDS2000 which was released at the end of 2013 and still has issues in the current firmware according to this posting: https://www.eevblog.com/forum/testgear/siglent-sds2000-new-v2-firmware/msg1077062/#msg1077062. That adds up to over 3.5 years.

[TheDefpom]

Interesting that you quoted a bug I found... the post was about them fixing the counter display!

As I hadn't had the scope that long at the time, and I had only reported the bug about a few months or so earlier they actually fixed it pretty quickly !

The counter was a lot further out than that, and don't forget firmware cannot correct for a slight internal oscillator variance.

[zike]

Me again. OK this one's crossed the line.

Setup to look at a functioning I2C link (arduino-LCD display). Assigned data and clock to ch.1 and 2,  linked to trigger, requested format ASCII,  hit "list" to show decode; table is drawn but shows alert symbol top left and  "No data" . OK so I set something up wrong...

but  scope is now unresponsive! 

No soft or hard buttons affect operation, just triggers repeatedly on my serial test line (or auto triggers if I unplug probes).

Power button doesn't even work; had to unplug from mains to reboot.

Took a video documenting ineffective key presses etc.,  but it's unclear how to share. Here's a screen still, in any case.

EDIT: Haven't managed to replicate problem in the last 20 minutes.

[tautech]

Me again. OK this one's crossed the line.

Setup to look at a functioning I2C link (arduino-LCD display). Assigned data and clock to ch.1 and 2,  linked to trigger, requested format ASCII,  hit "list" to show decode; table is drawn but shows alert symbol top left and  "No data" . OK so I set something up wrong...

but  scope is now unresponsive! 

No soft or hard buttons affect operation, just triggers repeatedly on my serial test line (or auto triggers if I unplug probes).

Power button doesn't even work; had to unplug from mains to reboot.

Took a video documenting ineffective key presses etc.,  but it's unclear how to share. Here's a screen still, in any case.

EDIT: Haven't managed to replicate problem in the last 20 minutes.

You can paste the link in a post or send it to me via PM, either way the factory will get to see it when they start work in ~90 minutes.

Interesting, never had a freeze, anybody else ?