SIGLENT SDS1102CM and EasyScope

[tinhead]

In the case of the Hantek, this was most likely due to imperfectly spaced sampling due to difficulty timing 5 DACs rather than 4 on the atten. 

actually it is vice versa - Hantek does use 4 or 8 ADCs and ATTEN/Rigol/Siglent 5 or 10, but the problem is everywhere the same:
- ADCs clocked from FPGA (jitter)
- FPGA not using dedicated clock outputs (jitter)
- all VREF (ADCs) connected together and to internal VREF instead of external and stable VREF
- bad PSUs/signal path supply filtering

All these factors above have big influence, most of them (PSU and VREF can be changed, clock change need drastical redesign - cost factor!!!) can not be fixed or only when you spend a lot of $$$ (which didn't make sense because then you can buy
better scope and hope the manufacturer did good job .. which is unfortunately not always the case, you can buy 20k $ scope
and will still see distortion created by not properly aligned interlevaed ADCs).

Additionaly by design all these low budget DSOs does not have clocks aligned properly in phase - why ? well
because they can't, the FPGA i/o jitter is already that high that you can't align them. All these DSOs are using
simple trick - sample every ADC sperate -> calculate cross zero timepoint -> align the sampled data.
This works of course good on paper - but in real world already a small jitter in test signal during factory calibration
is creating distortion in the table (and again, look the factors above, they have later still influence),
sometimes (when you have good FG) you can recalibrate DSO to work a better as with original factory calibration,
the cross zero calculation is very depandant on jitter, edge rise time and timebase settigns during calibration.
I was for example to calibrate my DSO light years better than Hantek did it originaly - a extrem low jitter FG very fast rise time,
patched firmware (to run calibration in 1ns/div - yes ONE).

The best case scenario is with non-interleaved ADCs, however it costs money - actually only Rigol CA/ATTEN CAL is
using ONE dual-channel ADC (which of course is still in interleave mode when only one channel enabled, but the distortion
is a bit better because the ADC is taking care of clock phase and not FPGA) - but they have only small sample memory.
Another option is Owon SDS - same as Rigol CA/ATTEN CAL but it does have 10Mpoint memory - unfortunatley waveform update
rate is slow as hell. The next possible DSO is then already Agilent DSOX, Instek DSO3000 and so on.

Now back to low range DSOs - depends on firmware you might get fooled, a non "wobbling" waveform on screen didn't means
that the DSO is really properly working (ADCs phase aligned) - when you switch Hantek to slower screen update you
will get waveform exact such stable as on Rigol - on Rigol you can't just switch to faster update to see all these distortions
created by jitter and aliasing. So what is better ? Non of them, one setup is lying by doing avg, another is wobbling around
due all these factor above

And finally QC - when you have no luck you bought "broken" scope, maybe not 100% broken but probably some
components issue (drift, cold solder, transport issues) or lazy ppl at manufacturer side (moved cable during calibration, whatever).

All you can do is to compare with EXACT same model and EXACT the same test setup, just a quartz oscillator and piece of unknown
cable/terminator is not enough.

[Zero999]

Guess who's kicking himself right now that he didn't follow the crowd and get a DS1052E ? 

The DS1052E is only 50MHz unless you hack it, has much less memory. The 100MHz Rigol is much more expensive and there's still the short memory issue.

The best case scenario is with non-interleaved ADCs, however it costs money - actually only Rigol CA/ATTEN CAL is
using ONE dual-channel ADC (which of course is still in interleave mode when only one channel enabled, but the distortion
is a bit better because the ADC is taking care of clock phase and not FPGA) - but they have only small sample memory.

So the ADS1102CAL seems like the better 'scope if having loads of memory is unimportant.

[saturation]

tinhead, always a pleasure to read your very thoughtful, analytical and extensively technical commentary!

I agree fully with your comments but regardless on how the ADC are configured, the designs should take the jitter into account so it isn't noticeable, if not invisible, until >= the scopes rated maximum bandwidth.   

The 2 items of concern shown by the OP in videos are horizontal drifting and sampling waveform distortion at very low frequencies.  Unless this model Siglent has user adjustable sampling rate apart from the horizontal timebase adjustment, the drift and waveform distortion occurs at << 1 MHz. 

As for the Rigol 1052e,  as one of the most tested scopes on eevblog, Dave has shown in his reviews how it compares at transients against the superfast Agilent InfiniiVision series, but in this exercise it clearly need not do more to demonstrate it does not have the same issues this Siglent shows at << 1 MHz.

The only cause for some doubt about the Siglent's flaws is the signal source could be more stable to rule out its instability as throwing the Siglent off; but the simultaneously analog scope measurement at low frequencies suggests the Siglent more likely has issues.

Here is a much better test, in this case its a Tek vs a UniT scope, and the quality of the signal source is much clearer so the differences in waveform distortion is more apparent.



By comparison, the newer Atten scopes appear to be better than this Siglent and the UniT:

..you can buy 20k $ scope and will still see distortion created by not properly aligned interlevaed ADCs).
...all these low budget DSOs does not have clocks aligned properly in phase - why ? well
because they can't, the FPGA i/o jitter is already that high that you can't align them. ...a small jitter in test signal during factory calibration

... The best case scenario is with non-interleaved ADCs, however it costs money - actually only Rigol CA/ATTEN CAL is using ONE dual-channel ADC (which of course is still in interleave mode when only one channel enabled, but the distortion is a bit better because the ADC is taking care of clock phase and not FPGA) - but they have only small sample memory. Another option is Owon SDS - same as Rigol CA/ATTEN CAL but it does have 10Mpoint memory - unfortunatley waveform update rate is slow as hell. The next possible DSO is then already Agilent DSOX, Instek DSO3000 and so on.

Now back to low range DSOs - depends on firmware you might get fooled, a non "wobbling" waveform on screen didn't means that the DSO is really properly working (ADCs phase aligned) - when you switch Hantek to slower screen update you will get waveform exact such stable as on Rigol - on Rigol you can't just switch to faster update to see all these distortions created by jitter and aliasing. So what is better ? Non of them, one setup is lying by doing avg, another is wobbling around due all these factor above

And finally QC - when you have no luck you bought "broken" scope, maybe not 100% broken but probably some
components issue (drift, cold solder, transport issues) or lazy ppl at manufacturer side (moved cable during calibration, whatever). All you can do is to compare with EXACT same model and EXACT the same test setup, just a quartz oscillator and piece of unknown cable/terminator is not enough.

[Kozmyk]

My Siglent is parcelled up ready for return and I'm ordering a Rigol.
My concern now is that after all the niceties of my comms with the seller; when I asked to confirm the return address they sent me a different one from their sending address and they're asking me to mark it as a GBP10 GIFT. 
Doing that would lose me all protection if the parcel were lost or damaged.
I need it to go Insured and Signed for just in case I have to claim through PayPal if it all goes tits up.

I've voiced my concerns to them and I'm awaiting a reply.
Their business day won't start for another hour or two.
I don't mind helping them avoid a few charges but not at the risk of it disappearing and me getting shafted.

[saturation]

Good luck, Kozmyk.  Yes, the biggest problem with mail order is returning it, it can eat into the savings of buying it in the first place.  Please keep us posted with your adventures with Siglent.  Note the Rigol is not without its faults recently,  since 2011 Hellene received a unit with much high baseline noise than prior models and several users report firmware 2.5 caused some usability bugs like flickering screen and buttons not working, sometimes [ there are videos on youtube].  However, if you follow the Rigol thread on hacking it, other users are not reporting such issues.  Lastly, this week a new firmware was released 2.6, and the last known good one, 2.4 is downloadable from the Rigol website.  All the Rigol faults have discussions on eevblog.

My Siglent is parcelled up ready for return and I'm ordering a Rigol.
My concern now is that after all the niceties of my comms with the seller; when I asked to confirm the return address they sent me a different one from their sending address and they're asking me to mark it as a GBP10 GIFT. 
Doing that would lose me all protection if the parcel were lost or damaged.
I need it to go Insured and Signed for just in case I have to claim through PayPal if it all goes tits up.

I've voiced my concerns to them and I'm awaiting a reply.
Their business day won't start for another hour or two.
I don't mind helping them avoid a few charges but not at the risk of it disappearing and me getting shafted.

Yes, the Rigol is basically 50 MHz but I think its still bang for buck; but there are other scopes reviewed and championed by tinhead, marmad and others ... with superb reviews of function and usability, but they generally cost more money.  But worth looking into it if the costs difference is very small in your country balanced against their specs and bandwidth; these folks are showing the scopes are not just feature laded but well made, including the newer Atten wide screen models.  For me, the Rigol cost $360 delivered by an authorized local dealer, so even the very competitive Instek 1062A, with ~ similar bandwidth, its was  $60+ through tequipment.net over 1 year ago.  So in the end, what you end up with may take into account more than just the specs, but what deals fit you best.  To ease the problem with defects and warranty, I prefer vendors be local to make service easier.

The extra long memory is always appreciated but the scope, IMHO, had first work best in providing stable clean waveforms before anything else.

Guess who's kicking himself right now that he didn't follow the crowd and get a DS1052E ? 

The DS1052E is only 50MHz unless you hack it, has much less memory. The 100MHz Rigol is much more expensive and there's still the short memory issue.

The best case scenario is with non-interleaved ADCs, however it costs money - actually only Rigol CA/ATTEN CAL is
using ONE dual-channel ADC (which of course is still in interleave mode when only one channel enabled, but the distortion
is a bit better because the ADC is taking care of clock phase and not FPGA) - but they have only small sample memory.

So the ADS1102CAL seems like the better 'scope if having loads of memory is unimportant.

[Kozmyk]

I probably will have a go at doing the Rigol hack eventually.
But not before I've satisfied myself that it's working properly.

After all the fuss and bother it appears that I've only saved £30/$48 by buying the RIGOL from eBay rather than from the German distributor Batronix.
I was too fixated on eBay deals to realise there was a good deal much nearer home.
My mistake.
I had looked at UK prices but the ones I could find here were all much higher than Germany.

As long as this Rigol from the seller in China is Ok then I'm clear but that's the risk we take trying to save  money isn't it.

[tinhead]

The 2 items of concern shown by the OP in videos are horizontal drifting and sampling waveform distortion at very low frequencies.  Unless this model Siglent has user adjustable sampling rate apart from the horizontal timebase adjustment, the drift and waveform distortion occurs at << 1 MHz. 

when you open both Rigol E and Siglent aka ATTEN you will find identical hardware (actually ATTEN/Siglent
added only some extras for 50R termination and changed TFT controller to drive larger screen), so whatever they
did wrong is inside firmware (as Rigol was able to use exact same pice of hardware much better).
This is actually the price you have to pay for cloned clones of clones
Normally i would say "who cares, they will fix that in next fw revision", but unfortunately these products are from
the "buy and die" category, support is not existing at all. I noticed it already as i played with the older ATTEN DSO
(and was thinking about to test Siglent). This is a really unfortunate, a better firmware would made these scopes
better than the original (Rigol E) just because of the larger display.

Anyway, good luck Kozmyk, in my case was easy to return due my friends in China.

[Kozmyk]

Thanks for the wishes of good luck.

The sellers comms have been good so far, allowing for the time zone difference.
I'm just uncomfortable about this request to send Gift and low value.
I don't understand why, it's not like the seller is Importing the unit, it's a return.
Maybe they get a tax break for an export sale but have to pay it back if the unit is returned.
I don't know how it import/export rules work for Chinese sellers.

At the end of the day I'm going to go with whatever eBay/PayPal advise.
They're the ones I'd be claiming from if anything went wrong with the return.

[cseb]

Hi,

This is how I see the SDS1102CM after 2 weeks:
http://cseb.hu/siglent

Regards,
Csaba

[dmderev]

I have Atten ADS1102CAL - and my impression is that it uses the same 5x AD9288 ADCs in interleaved mode. I was mostly satisfied with the features for 1 week, until I discovered few problems on fast timebases (25ns/div, 10ns/div, 5ns/div). First, the trigger point (the crossing of the waveform and trigger level and location where the scope shows the trigger arrow) is shifted by 3-4ns. Not a big deal, but something to pay attention to.

The more serious problem is that sometimes, the scope shows the waveform shifted by 10ns right from the trigger point. This happens randomly, and looks like an overlay of two unstable waveforms. The shifted waveform is exactly the same, but simply shifted. Very annoying. This makes me think that it is a synchronization problem in FPGA - the trigger signal simply ends in the earlier memory location. This also makes me think that the fundamental sampling rate is 100MHz (10ns) and the ADCs are time-interleaved. Or at least the words stored in RAM are combined of 10 samples.

Interestingly, this behavior happens not all the time. It is possible to turn the scope on and off, and sometimes it does not show the defect. This makes me think that some logic in the acquisition section is not properly reset or initialized - otherwise all power-on cycles would have shown same result.

Did anyone see such strange behavior? Did anyone open Rigol 1102 or Atten 1102 or Siglent 1102 scopes (they should be exact clones)?

Is there any schematics available?

[techhie]

It looks like I am not the only one who has had a problem getting Easyscope 3 to talk to the ATTEN ADS1102CAL. I found the English version on the included CD as well as USB drivers, which only load on a Windows XP machine. Will not work on a Windows 7. Everything goes according to plan when plugging in the scope and starting the software then, press "connect", and it fails! Very frustrating. No reply so far from ATTEN. Maybe they are unable to diagnose and fix?

Any and all suggestions here would be much appreciated.

[Erwin Ried]

It looks like I am not the only one who has had a problem getting Easyscope 3 to talk to the ATTEN ADS1102CAL. I found the English version on the included CD as well as USB drivers, which only load on a Windows XP machine. Will not work on a Windows 7. Everything goes according to plan when plugging in the scope and starting the software then, press "connect", and it fails! Very frustrating. No reply so far from ATTEN. Maybe they are unable to diagnose and fix?

Any and all suggestions here would be much appreciated.

Did you ever find a solution? I can't connect either (sorry for replying to this very old thread, but it seems the issues still continue in this area, now with easyscope X...)