Siglent SDG1020 square signal "dirty"

[Xyphro]

Hi!

Did anybody ever have a close look at the Siglent SDG1020 square wave output at low frequencies?

Settings: 1 Hz, Square, 50% duty. Zoom in to the rising or falling edges. Output 1 (right BNC connector).

Looks like a bug in the DDS section (see attached picture).

Output 2 is free of this problem.
Other waveforms look fine at Output 1...

It could be a feature which is useful at sine wave output, but useless and nasty for square wave output: Dithering of the phase accumulator for low frequencies?!?

Best regards,

Kai

[rf-loop]

It have discussed also earlier.

Your image do not look like 50% duty.
Out from 50% this situation is more bad than 50%.

It is not DDS problem and/or FW problem at all.

Square wave is special case.  It is derived from internal sinewave using PECL comparators. (and this principle is not only Siglent special)
 This make square wave function different as all other functions.

On some HW's PECL comparator hysteresis setting is not optimum and in some cases hysteresis setting resistor value  change may give some amount better result.

Of course there may be some other problem in some individual unit if this signal pathway to comparator is noisy (example some component failure what is related to this signal noise)
Now if use low frequency and specially other than 50% duty this small noise may lead to wrong comparator "switch" specially if comparator hysteresis window is too narrow.

It is explained littlebit more here.

And here.
If need better signal quality specially for low frequencies  it is better to use pulse than square.

[Xyphro]

tak, rf-loop!

I just also noticed, that the pulse mode is clean and was wondering why.

[bmwnomad]

Is this the same thing you guys are talking about? 

I have an SDG1025 and was going to send it back under warranty for a replacement, but if they are all like that then maybe I'll have to live with it.

Attached are scope snapshots.  Yellow are channel 1, 1khz square wave and blue is channel 2, 1khz square wave.  I set it on infinite persistance for a few seconds to show the glitches.

Channel 2 is fine, as is channel 1 pulsed output(not shown), it's just channel 1 square wave that looks like crap.

Steve

[rf-loop]

This difference between channels is now strange.

[Xyphro]

Yes, it is the same issue.

Output 2 is also fine with my generator, while Output1 shows this behaviour.

[bmwnomad]

Thanks, no use in returning it then if I'll just get another one with the same issue.

Steve

[Bored@Work]

Thanks, no use in returning it then if I'll just get another one with the same issue.

Steve

But returning it could be useful when you want to get your money back. And next time the resident Siglent sockpupets tell you how great Siglent/Atten quality is, you know the truth.

[Wuerstchenhund]

But returning it could be useful when you want to get your money back. And next time the resident Siglent sockpupets tell you how great Siglent/Atten quality is, you know the truth.

The problem is that the alternatives in this price range are hardly better. And to be honest, the general hardware quality of the Siglent SDG1000 Series is not *that* bad, and if they would sort out the jitter issue then this could be a decent low-cost AWG for hobby use.

Of course Siglent should be deeply ashamed to release a product with such a design flaw on the market, and they should fix the issue for the owners of these generators (which won't happen).

[rf-loop]

I have tested SDG1020  + some SDG1010
All are same HW version (I forget all data to workshop) and same latest FW version .27

SDG1010 have no any kind of sign about this issue. Also glitch level in square wave is less than before. These works just ok.

SDG1020 is totally as from different world in quick test. But all tests are not ready yet.

CH1 square wave (not other func) very bad timing jitter and very bad glitches as someone show here before. This is problem specially in low frequencies. Higher frequencies better. (naturally due to circuit principle)

And more:

CH2 have much better square wave. (not as good as SDG1010 and not as good as SDG1025 what I have tested before. Difference is not big.

But then CH2 have other issue in unit what I have tested..

Square, ramp, sine etc. All fails but looks like only if use 50ohm and highes voltage band. Specially it can most easy look with square or pulse. After signal rise it looks like go flat but just after corner it sart "oscillate".

Oh well, if you look it with slow oscilloscope it looks just ok or like only "noise level in signal top is more than bottom. Take fast oscilloscope and ... oops.
Signal top (not bottom) have high level some kind of oscillation with high frequencies. Something like over 200MHz!

This is first time I have tested SDG1020. (becouse Teledyne do not have this version), they have 1010, 1025 and 1050 only.  (2012, 2022, 2052)

What is this problem, I do not know anything more now.
(I do not sell this 1020 model at all before whole this issue is solved)

Please, who have this issue in SDG1020 model:
HW version, FW version and whole serial number exept last 3 digits  as xxx.
(I think posibility that this issue is in some manufacturing lot.)

[rf-loop]

Here SDG1020 some tests using Square wave.
Serial SDG00001130xxx
HW:02-00-00-21-24
FW: 1.01.01.27

Also compared to SDG1010 (not only one unit)
same serial (exept xxx)
same HW
same FW

also I have earlier tested other SDG1010 and SDG1025  and not at all this problem what I can see in tested SDG1020. SDG1020 tests looks just as someone others have also find.

Connection to scope 50ohm coaxials and 50ohm terminators
SDG1020 CH1 glitches in images.
If use more low frequency there  can see  also full random width level pulses near rising edge (not picture here)

There is also problem in CH2 if use 50ohm load adn if use over 1Vpp. High frequency oscillating in signal upper levels. This oscillation can see also in other waveforms. (280MHz! oscillation with this setting)

pictures filename also explain settings.

[commongrounder]

I checked this issue on my SDG1005 (FW..27), just to add another model number to the list.  I set up my generator and scope as rf-loop did, and I have the same problem with glitches on low frequency square waves on channel 1 only. It only begins to clear up above about 10khz on my unit.  Channel 2 is very stable with only a little noise at the bottom of the waveform (no HF oscillations).  I set up my scope for infinite persistence and triggered off of channel 2 to show the severe timing jitter between the channels, over 70us peak.  About 30 second capture time.  Would not want to use dual channel square wave mode for any phase related work!  Pulse is definitely better.   Oh well.  Better to find this stuff out and understand the limitations now, rather than banging your head against the wall trying to figure out a crazy output from a DUT.

[rf-loop]

I checked this issue on my SDG1005 (FW..27), just to add another model number to the list.  I set up my generator and scope as rf-loop did, and I have the same problem with glitches on low frequency square waves on channel 1 only. It only begins to clear up above about 10khz on my unit.  Channel 2 is very stable with only a little noise at the bottom of the waveform (no HF oscillations).  I set up my scope for infinite persistence and triggered off of channel 2 to show the severe timing jitter between the channels, over 70us peak.  About 30 second capture time.  Would not want to use dual channel square wave mode for any phase related work!  Pulse is definitely better.   Oh well.  Better to find this stuff out and understand the limitations now, rather than banging your head against the wall trying to figure out a crazy output from a DUT.

Timing jitter is different case. (there you see now channel - channel total. (both channels have just individual random jitter. Also it is good to know that most of manufacturer  tell rms jitter. And it is just different case becouse jitter is gaussian random distribution)

Check channels so that look generator CH1 alone. Then look CH2 alone.
Timing jitter is specified: 0.1% of period(typical,1kHz,1Vpp)
And there is not specified lower freq jitter.

50Hz period is 20ms. Jitter (typical) 20us. Jitter from "0" line (ideal time)  to positive direction 20us and to negative direction 20us.
It can measure so that trig rising edge. Go to next rising edge and look timing jitter.
(also this is NOT channel to channel jitter)

These glitches is best to look so that trig rising edge and keep peristence on, in this case timing jitter do not overwrite on oscilloscope screen these glitches.

If need less jitter specially with low frequencies it is better to use "Pulse" function and not "Square".  Pulse have 8ns p-p jitter  independent of pulse freq.
It is not so user friendly becouse if set duty, it do not keep %. All time need adjust duty agen if change freq.

 
(SDG5000 serie is totally different animal. )

SDG1000 series Square principle:

[commongrounder]

Yes, indeed I got off on a tangent about jitter, which, while different than the glitches showing up from the sine to square conversion, do combine to aggravate the problem.  My unit does have about +/- 25us peak (worst case) jitter at 50hz (1vp-p into 50ohm scope termination), observing the next rising edge after the triggered edge.  I did my best to set the trigger point above the top of the glitches to avoid having that influence the observed jitter.  It is clearly visible even when observing several wave repetitions on the screen (on my Rigol DS4012, anyway).  CH 2 has +/- 10us under the same conditions, and none of the glitching.

Thank you for your thoughts on this, rf-loop.

[nixxon]

Here is a screenshot of my DSG1020 @ 1 kHz square wave 50% 4 Vpp (high Z) straight coax (50 ohm) after almost 90 minutes persistent display logging on my rigol DS1102. Nothing special or close to the horror images in previous post(s).

Analog Ch  State   Scale    Position   Coupling  BW Limit  Invert
CH1        On      100mV/    2.00V     DC        Off       Off

Analog Ch  Impedance   Probe
CH1        1M Ohm      1X

Time    Time Ref    Main Scale    Delay
Main    Center      5.000us/      -25.00000us

Trigger  Source      Slope    Mode      Coupling     Level    Holdoff
Edge     CH1         Rising   Auto      DC           -1.40V   500ns

Acquisition    Sampling    Memory Depth    Sample Rate
Normal         Realtime    Normal          50.00MSa   

[nixxon]

The difference between 500mV/Div and 200mV/Div on my Rigol DS1102E is strange though. Can anyone explain why I get such different low level trace on the very same signal from my Siglent SDG1020 (as in my previous post)?

500mV/Div
Analog Ch  State   Scale    Position   Coupling  BW Limit  Invert
CH1        On      500mV/    2.00V     DC        Off       Off

Analog Ch  Impedance   Probe
CH1        1M Ohm      1X

Time    Time Ref    Main Scale    Delay
Main    Center      50.00us/      -250.0000us

Trigger  Source      Slope    Mode      Coupling     Level    Holdoff
Edge     CH1         Rising   Auto      DC            264mV   500ns

Acquisition    Sampling    Memory Depth    Sample Rate
Normal         Realtime    Normal          5.000MSa   

200mV/Div
Analog Ch  State   Scale    Position   Coupling  BW Limit  Invert
CH1        On      200mV/    2.00V     DC        Off       Off

Analog Ch  Impedance   Probe
CH1        1M Ohm      1X

Time    Time Ref    Main Scale    Delay
Main    Center      50.00us/      -250.0000us

Trigger  Source      Slope    Mode      Coupling     Level    Holdoff
Edge     CH1         Rising   Auto      DC           -800mV   500ns

Acquisition    Sampling    Memory Depth    Sample Rate
Normal         Realtime    Normal          5.000MSa

[nixxon]

This is strange. Stepping up in 0.1 Vpp increments from 3.0 --> 4.0 Vpp and 2.00 --> 4.00 Vpp respectively:

[nixxon]

My Rigol was reaching bodily temperatures. That could explain it all 

(actually all the hotspots on the Rigol are the shiny, low emissivity BNC Connectors (+ test signal/Ground rightmost), so I guess the higher temperature is actually reflected from the overhead 5 spot 12V Halogen light bulbs)

Edit: Just as rf-loop suggested, the BNC connectors actually were warmer than the rest of the front panel. There was no significant IR reflection from the Halogen lights above. The panel just around the BNC connectors is slightly warmer than the rest of the front panel, due to thermal radiation and thermal conduction, and maybe even thermal convection (above the BNC connectors).

I have added an IR close up of the BNC connectors.

[rf-loop]

Can you tell your SDG1020  HW version, FW version and serial number SDGnnnnnnnnxxx (no need last three digits (xxx)
In your pictures can not see any strange from SDG.
(but test is very limited)

--------

This temperature in Rigol input BNC is normal. Analog front end convert electric to heat ( smile )) and you see then BNC connectors temp rise. This front end circuit is hot inside oscilloscope. Some chips there produce lot of heat (also without any signal), and it is normal.



In picture you have  200mV/div. Signal level is example 4Vpp. It means your signal peak to peak is 20div. Then you add around 10div offset and you think now signal bottom is on the oscilloscope center line - but it is not ideal machine. With some levels it reacxh its limits. (I do not remember numbers)

Are you sure your Rigol analog front end can perfect handle this. No, it can not. It is going to its limits. It is perhaps out from its linear working area.
My opinion is that nothing is wrong in your Rigol. It works as it is designed. You need know your instruments and they limits.

(I have tested Rigol long time ago and I do not remember exatly these numbers but I remember well how it looks in signal. And this looks like just same what I remember (in my head image memory). (your signal top is out of range what it can handle and it affect also to bottom part of signal what is visible on screen)

It can say, here test equipment and device under test is swapped. Wagons drag horse. With SDG1020 you now test Rigol but originally you think you test SDG using Rigol.

In this test you find Rigol front end "features".   

[nixxon]

Thank you, rf-loop, for much useful info and for calming me down regarding my equipment.

It seems my SDG1020 is OK at low freq. square Waves (compared to the horrific examples shown above in this thread).

Here is some more info about my unit:

Original configuration:
 Software version: 1.01.01.20
 Hardware version: 01-00-00-17-23
 Serial number: SDG00002120XXX

Previous configuration (after 1.01.01.23 update):
 Software version: 1.01.01.23
 Hardware version: 01-00-00-19-23
 Serial number: SDG00002120XXX

Current configuration (after 1.01.01.27 update):
 Software version: 1.01.01.27
 Hardware version: 01-00-00-21-23
 Serial number: SDG00002120XXX

[rf-loop]

Thank you about this information. Your SDG is older and without more knowldge it can also suspect that problem is in more new HW version.
I find this problem in quite fresh version from factory.

It is very nice if some others who have SDG1020 problem (showed in this thread) give this version information also.

[commongrounder]

Mine, as noted previously, is an SDG1005, but I will contribute my unit info for what it is worth.
From the info screen:
Boot Strap number 14
SW 1.01.01.27
HW 02-00-00-21-24
SN SDG00004120xxx

Thank you rf-loop for your interest in this issue!

[BlueLaser]

I have had similar problems and noted them down on my post:  https://www.eevblog.com/forum/testgear/siglent-sdg1020-arbitrary-fungen-first-impressions-and-a-possible-signal-glitch/  Do note that there are some modifications you can make to the board including replacement of resistors and addition of capacitors.  I found that rf-loop's advice of using the pulse function has resulted in signals clean enough for me to use for my application.  Since I did not need the extra frequency range--it worked out in my case. 

[commongrounder]

BlueLaser:  I did see, and appreciate, your post with the modifications that can be done.  Thanks!  I hesitate to crack the case until my warranty is done, however, to cover myself if the thing decides to die completely!  Pulse mode works for most of my applications as well, thankfully.

[rf-loop]

I have had similar problems and noted them down on my post:  https://www.eevblog.com/forum/testgear/siglent-sdg1020-arbitrary-fungen-first-impressions-and-a-possible-signal-glitch/  Do note that there are some modifications you can make to the board including replacement of resistors and addition of capacitors.  I found that rf-loop's advice of using the pulse function has resulted in signals clean enough for me to use for my application.  Since I did not need the extra frequency range--it worked out in my case.

What is your HW version?

It is also possible this issue is different.
This issue is not just same what I have seen before. There is something same but then also very different.