Siglent SDG1000X Waveform Generators

[TurboTom]

100% identical on SDG6000X. That's actually what's to be expected on this family of AWG with an almost identical U/I. That behaviour isn't "nice" but at least it's showing on the screen what it's doing  .

[pope]

Forgive me if this has already been discussed but I scanned trough all 9 pages and didn't find any info....

I read on another forum that the SDG1032X (and probably all SDG1000X series) produces DC offset at the output even if you set the DC offset to zero VDC.

Can someone confirm whether that's the case or not?

Many thanks! 

[rf-loop]

Forgive me if this has already been discussed but I scanned trough all 9 pages and didn't find any info....

I read on another forum that the SDG1032X (and probably all SDG1000X series) produces DC offset at the output even if you set the DC offset to zero VDC.

Can someone confirm whether that's the case or not?

Many thanks! 

Yes of course there is offset.

In whole world there do not exist ideal perfect instruments in real life - exept in kids school books.
Also it is told in data sheet and also it is more told in public service manual. You have not read them.

Question is about how much. 
If there is offset what is too much then it need repair or recalibrate in Siglent service center (or just put up with it).
How much offset your SDG1032X have (after it is warmed/stabilized)

My individual 32X. With Waveform DC selected and level set for 0.000V  CH1 output is roughly around  -0.63mV and CH2 -0.48mV  after 30 minute warming in 23°C room. (meas with cal checked Keithley 2015THD)

So my SDG is well inside specifications  +/-3mV   @0.000Vdc  set.

Attached DC Output verification record table from SDG1000X service manual. Measure and fill table.

[pope]

Thanks for your answer.

You're right. I haven't read the manual. Probably because I don't own this AWG yet. I'm still debating whether I should get an 1000x or an 2000x since I only work with analog audio and I already have a couple of (old) function generators that can do up to 2MHz. Any advice on that?

In the comment I'm referring to, the user said that he measured around 80mV on one output and 60mV on the other. This is well out of specs.

[2N3055]

Thanks for your answer.

You're right. I haven't read the manual. Probably because I don't own this AWG yet. I'm still debating whether I should get an 1000x or an 2000x since I only work with analog audio and I already have a couple of (old) function generators that can do up to 2MHz. Any advice on that?

In the comment I'm referring to, the user said that he measured around 80mV on one output and 60mV on the other. This is well out of specs.

That person has a damaged AWG or is doing something wrong..

[ScottW]

Just wondering if this is normal/expected behavior for SDG1062X. 

Both images are 30MHz square wave.  All equipment/cables same in both tests, just changing SDG1062X amplitude from 1Vpp to 2Vpp.

Is the difference in the waveform to be expected?

Equipment is:
SDG1062X - Ch1 - Output set 30MHz Square and 50Ohm
SDS1204X-E
Pamona RG-58 patch cable with Cal-Test 50Ohm pass-through terminator

[tautech]

It's the scope.
Gibbs ears.
https://en.wikipedia.org/wiki/Gibbs_phenomenon

Some acquisition settings can reduce them.

[ScottW]

It's the scope.
Gibbs ears.
https://en.wikipedia.org/wiki/Gibbs_phenomenon

Not sure scope acquisition (Gibbs) explains what I'm seeing.  The wave looks good at 1Vpp, but suddenly has noticeable overshoot when moving from 1.0Vpp to 1.1Vpp.   Coincidentally, there is an audible "CLICK" from the SDG1062X when moving between 1.0Vpp and 1.1Vpp (Transformer tap relay? Amplifier switchover?).

Here are two more images.  One at 1.0Vpp, then one at 1.1Vpp (seen by scope as 1.23V, due to the overshoot).  Notice the change in overshoot that appears with that 0.1Vpp increase.  This does NOT happen anywhere below 1.0Vpp.  It occurs abruptly when moving from 1.0Vpp to 1.1Vpp -- exactly when the relay "CLICK" is heard from the SDG1062X.  [Clarification:  The "Click" and "Waveform Change" happens when moving between 1.9Vpp and 2.0Vpp when output is HiZ, or moving between 1.0Vpp and 1.1Vpp when output is at 50Ohm.]

Ideas?

[HendriXML]

I've used the AWG as a precision voltage source using an 6.5 digit multimeter meter and 2 awg outputs and some resistors. And the voltage ranging (relay click) had some slight jumps/drops. Making it harder to get to the target voltage with small adjustments. It would have been nice if one could leave it in the high voltage range and thus disable auto ranging. (Would be handy for testing your case as well.)

Your case seems like and overshoot. Maybe it's a property of the output stage when supplied with a higher voltage?

[ScottW]

Maybe it's a property of the output stage when supplied with a higher voltage?

Perhaps, as it definitely seems related to the voltage range threshold (relay click).  But I'm not yet sure whether what I'm seeing with the waveform is "characteristic of the model", or if my particular SDG1062x has a problem.  If what I'm seeing is characteristic of the model, I guess I can live with it.  Hoping another SDG1062x owner can confirm one way or the other.

[tubularnut]

Output from SDG1062X into both an SDS1204X-E and SDS2504X-Plus

Taken at 1.000Vp-p and 1.001Vp-p, which is where my relay switches. Though earlier today it switched when going from 0.999 to 1.000 ??

[ScottW]

Output from SDG1062X into both an SDS1204X-E and SDS2504X-Plus

Taken at 1.000Vp-p and 1.001Vp-p, which is where my relay switches. Though earlier today it switched when going from 0.999 to 1.000 ??

Thanks!  Those show the same thing I'm seeing (though with less magnitude), and on a 2504X+ as well.

I did a self-cal on the scope and swapped in a different cable and terminator.  That improved the picture a bit, so PART of it was the scope depiction. 

Probably not a big deal, but there is a definite overshoot that appears exactly at (and above) the "relay click" heard between 1.000V and 1.001V.

Here are my results, at 30MHz, 1.000Vpp and 1.001Vpp, after running a scope self-cal and swapping cables.

[HendriXML]

When decoding a SCPI response this kinda bothers me:

Code: [Select]

Sending: C1:BaSic_WaVe?
Received: C1:BSWV WVTP,PULSE,FRQ,500HZ,PERI,0.002S,AMP,5V,AMPVRMS,2.5Vrms,OFST,6.5V,HLEV,9V,LLEV,4V,DUTY,3,WIDTH,6e-05,RISE,1.68e-08S,FALL,1.68e-08S,DLY,0.0002

What is wrong IMO?

• HZ should be Hz
• S (siemens!) should be s
• Vrms should -in my view- just be V
• Delay, Width should also have an unit of s
• Duty should have an unit of %
• Values and units are normally separated by a space, but that I find less of a annoyance

Why?
Because if it's consistent and conform SI rules one can use generic methods of decoding such quantities. Otherwise it's in my view better not to use units at all.

[dorkshoei]

Hi.

I'm pondering buying a 1032X (tangent,  is $319 retail minus the 6% eevblog discount the cheapest I'm going to find it? Or is there anyone currently cheaper)

What is the current status of the 1032X hack? 

I see BillB's original post with telnet instructions.

I see a post from wally2q asking if there is an updated set of instructions but it didn't seem to be answered.

I see reference to this thread

I'm seeing posts about bricked units and skipping some steps.   I also seeing several posts by @rf-loop getting irritated with people for not reading the instructions which seems rather harsh given that it's not exactly clear.

Is it simply a case of following Bill B's original instructions plus (since it seems a later firmware removed the known telnet password) using the zip from "How to open a telnet session in a Siglent when the root password is unknown" thread to obtain a login?

And one final thing please;  cat $flames > /dev/null && echo ":-)"

[Hexley]

I also seeing several posts by @rf-loop getting irritated with people for not reading the instructions which seems rather harsh given that it's not exactly clear.

Actually, rf-loop posted a very helpful link in reply #74 above. This expands a bit on BillB's procedure and makes a useful guide. I've used the procedures described in that link a couple of times, most recently to install 1.01.01.33R1B6 back when it was released in 2020.

[dorkshoei]

I also seeing several posts by @rf-loop getting irritated with people for not reading the instructions which seems rather harsh given that it's not exactly clear.

Actually, rf-loop posted a very helpful link in reply #74 above. This expands a bit on BillB's procedure and makes a useful guide. I've used the procedures described in that link a couple of times, most recently to install 1.01.01.33R1B6 back when it was released in 2020.

Are you referring to this:  https://www.eevblog.com/forum/testgear/siglent-sdg1000x-waveform-generators/msg2876058/#msg2876058

I can't really make much sense of it.   

Regardless, re-posting a summary of the procedure (since what you reference is 6 pages, 140 posts and almost 2 full years old) is not a bad idea IMHO.

It sounds like you've been following the thread for a while, so I'm sure it all makes more sense.    For someone coming to it for the first time now, it is less clear. 

[Hexley]

Here is the address from the hyperlink he posted: [Edit - stale link. Use the actual link at reply #74, which has been updated.]
After you download all the files there, check the step-by-step instructions in the document "SDG1000X-MOD-Part-I-II-V33R1.PDF" and that should help make sense of it all.

[bdunham7]

Here is the address from the hyperlink he posted: https://app.box.com/s/srj3uev2s9wy78qj8d4jzi0j5ocgkmr9
After you download all the files there, check the step-by-step instructions in the document "SDG1000X-MOD-Part-I-II-V33R1.PDF" and that should help make sense of it all.

Very nice.  Do you know of a similar one-stop current resource for the SDG2000X models?

[Atomillo]

Hi:

For some time now, I've needed to count pulses at a low rate and to show that measurement in a computer. It obviously isn't worth a specially dedicated frequency counter, so I tried to use the Frequency Counter incorporated into my SDG 1062X.

I do not see a Totalize option and neither is it mentioned in the manual. In the bottom of the screen there is a Num value, but it doesn't seem to count pulses. If I connect a 1kHz square wave, it correctly measures it's parameters, but the Num value increases one unit with about 3 seconds or so.

Is there any way to use the frequency counter just as a counter?

Many thanks!

[ScottW]

I do not see a Totalize option and neither is it mentioned in the manual. In the bottom of the screen there is a Num value, but it doesn't seem to count pulses. If I connect a 1kHz square wave, it correctly measures it's parameters, but the Num value increases one unit with about 3 seconds or so.

Is there any way to use the frequency counter just as a counter?

I don't think there's a way to count individual pulses with the 1062x (though I'd love to be wrong).

I think the "Num" value is essentially showing the number of "analysis periods" (i.e., quantity of gate time periods over which the events have been analyzed) to arrive at the displayed value.  I don't see any way to change that gate time on the 1062x -- looks like it is fixed at about 2 seconds.

[Atomillo]

That seems indeed to be the case.
I wonder how easy it would be to count pulses. Might be it's just a slight firmware change. Being able to change the game time would also increase the functionality tremendously.
As is, the frequency counter seems more like an afterthought.

[cncjerry]

I have a 1032x (1062x).  I can align the phase using the utility menu but it doesn't track even when coupled.  At 30Mhz it is off by 7 degrees and by 60Mhz it is off by 14 degrees.  I can adjust the phase etc. but you can't do that easily when sweeping.  Is there another setting or something I am missing?  I bought it to use with the SDS 1104x-e BodeII plot function. I can fix it using PC based software to intercept the sweep setup but I bought it to avoid having to use the PC for Bode plots.

[rf-loop]

I have a 1032x (1062x).  I can align the phase using the utility menu but it doesn't track even when coupled.  At 30Mhz it is off by 7 degrees and by 60Mhz it is off by 14 degrees.  I can adjust the phase etc. but you can't do that easily when sweeping.  Is there another setting or something I am missing?  I bought it to use with the SDS 1104x-e BodeII plot function. I can fix it using PC based software to intercept the sweep setup but I bought it to avoid having to use the PC for Bode plots.

Some kind of default case. If use BodePlot so that only one generator channel is in use and signal is splitted for Ref and DUT, there need care that signal pathways travel times are equal (it need calibrate by adjusting cable lengths without DUT). Still some small phase shift depending frequency.

But then if use SDG and PB so that SDG one channel is for Ref and other cannel is for DUT.
In this case SDG is in tracking mode.
SDG [Utility] »» [CH Copy/Coupling] »» [Tack ON] (there can also adjust CH2 - CH1 PhaseDev). Usually we keep this 0 as default as it is. But it need of course understand that this is fixed phase shift, independent of frequency but with higher frequencies some error)

As you told there is other thing.  Around 7° at 30MHz  and around 14° at 60MHz in your case.
In my system it looks roughly around same.

It is better to think this based to time (phase debends frequency), not phase angle.
If look 7° phase shift at 33.333 ns period (30MHz) it is 0.648 ns
And 14° phase shift  at 16.666 ns period (60MHz) it is 0.648 ns

This time is total time shift and may be different with different individual instruments (oscilloscope + signal generator).
Also it looks like this is not so simple. It is also not exactly fixed time, it depends bit about frequency due to many reasons.

This time error need compensate externally. Example using cables where this error is compensated using cable lengths.
After that, some residual error cannot be fully compensated in any very simple way.

In pictures.

Pictures
SDG CH1 to Ref (oscilloscope Ch1)
SDG CH2 to Dut (oscilloscope Ch2)
SDG in tracking mode.

PodePlot Log sweep, freq 500kHz - 60MHz, level
Cables M17/084 (RG223) . Oscilloscope inputs 1M (not terminated 50ohm!)
2 equal length roughly 80cm  and one ~13cm longer



picture 1
Roughly equal cable lengths.
(SDG [Utility] »» [CH Copy/Coupling] »» [PhaseDev] »» CH2-CH1 PhaseDev -0°)


picture 2
Cable CH2 to Ch2 is ~13cm longer for compensate time error
After then there still exist some residual time error (displayed as phase error) what is bit more complex and depends freq.
(SDG [Utility] »» [CH Copy/Coupling] »» [PhaseDev] »» CH2-CH1 PhaseDev -0°)


picture 3
cables as picture 2 but then also SDG tracking PhaseDev adjusted for -5°
SDG [Utility] »» [CH Copy/Coupling] »» [PhaseDev] »» CH2-CH1 PhaseDev -5°
As can see it add fixed phase shift.

[cncjerry]

I guess I will use a resistive splitter when I need two driven channels.  It bugs me though because if the unit has two DDS chips then they can be synced to the same clock and this phase shift problem would be avoided.  I did test everything using a splitter and the problem is in the generator.  I can lengthen a cable based on the phase shift and velocity of the cable.

The alternative is to use a table or curve fit and drive the source through the PC like I do with my HP and other DDS units. But that is a hassle. 

Thanks for confirming the shift, RF-loop.

[gf]

Some kind of default case. If use BodePlot so that only one generator channel is in use and signal is splitted for Ref and DUT, there need care that signal pathways travel times are equal (it need calibrate by adjusting cable lengths without DUT). Still some small phase shift depending frequency.

Just wondering: Does the Bode Plot feature of the scope not support response normalization, in order that the effect of anything in the signal path which is not part of the DUT (cables, adapters, fixtures, probes) can be compensated? (For a network analyzer, such a feature would be taken for granted.)