Siglent SDG1032 as bad as rigol DG812?

[Fgrir]

On the SDG2042X using Pulse waveform instead of Square seems to behave better for this case.  Ran it multiple times and they always come out looking good.

[eTobey]

On the SDG2042X using Pulse waveform instead of Square seems to behave better for this case.  Ran it multiple times and they always come out looking good.

So you have the same issue (wrong pulses) on the SDG2042X ?

I am currently trying to test something. I want to have channel 2 as a trigger, and the other channel to test. I found only "track" to work in this case, but you cant have different loads on those. 

( Yes it does not matter in my case, but there are likely some cases...)

[eTobey]

On the SDG2042X...

Could you try something?
- Set up square wave.
- set high 3.3V
- set low 0V
- Invert the signal

Does it then go below 0V?

[Fgrir]

On the SDG2042X...

Could you try something?
- Set up square wave.
- set high 3.3V
- set low 0V
- Invert the signal

Does it then go below 0V?

Yes, on the SDG2042X inverting the output seems to be always relative to 0V, not relative to the Offset voltage setting as specified in the user manual.  So the 3.3V/0V setting gives -3.3V/0V levels when inverted.  It doesn't seem to matter if you use High/Low levels or Amplitude/Offset settings.

[2N3055]

On the SDG2042X...

Could you try something?
- Set up square wave.
- set high 3.3V
- set low 0V
- Invert the signal

Does it then go below 0V?

Yes, on the SDG2042X inverting the output seems to be always relative to 0V, not relative to the Offset voltage setting as specified in the user manual.  So the 3.3V/0V setting gives -3.3V/0V levels when inverted.  It doesn't seem to matter if you use High/Low levels or Amplitude/Offset settings.

Inverting signal means exactly that: inverting it. Meaning multiplying amplitude values with -1 in analog domain.
It is equivalent to scope invert function.
It does not perform digital logic NOT function.
AWG is not logic gate. It can go negative and positive values, by design.
It is a good thing.

[Fgrir]

Yeah, I'm not saying it's good or bad to invert about 0V or the Offset voltage, but the instrument behavior does not agree with what the user manual says it should do.

[2N3055]

Yeah, I'm not saying it's good or bad to invert about 0V or the Offset voltage, but the instrument behavior does not agree with what the user manual says it should do.

User manual is wrong. I will report it.

Polarity flip is symmetrical around the 0V. It is multiplying voltage with -1.
So signal that goes from 1V to 2V and back becomes signal that goes from -1 to -2.

Good catch.

[nctnico]

Yeah, I'm not saying it's good or bad to invert about 0V or the Offset voltage, but the instrument behavior does not agree with what the user manual says it should do.

User manual is wrong. I will report it.

Polarity flip is symmetrical around the 0V. It is multiplying voltage with -1.

The device is definitely wrong. There are lots of differential signalling protocols (LVDS, CAN and RS485 for example) which have a DC offset. With the way the SDG1000 is working right now, you can never generate such signals. IOW: the DC offset should be applied to both outputs in the same way, only the AC part of the signal should be inverted.

[2N3055]

Yeah, I'm not saying it's good or bad to invert about 0V or the Offset voltage, but the instrument behavior does not agree with what the user manual says it should do.

User manual is wrong. I will report it.

Polarity flip is symmetrical around the 0V. It is multiplying voltage with -1.

The device is definitely wrong. There are lots of differential signalling protocols (LVDS, CAN and RS485 for example) which have a DC offset. With the way the SDG1000 is working right now, you can never generate such signals. IOW: the DC offset should be applied to both outputs in the same way, only the AC part of the signal should be inverted.

Don't arbitrarily invent things.
AWG is analog device, not digital pulse generator.
Inverting analog signal is just that: inverting it, as in put it through unity gain inverting amplifier.

We could argue that a request could be made to Siglent to enhance AWGs with adding a choice to cater for digital signals.
Adding option Logic NOT, for instance.  So you would have "Normal", "Inverted" (both would be same as now) and "Logic NOT" that would invert signal in a logic NOT manner (inverting it but inside current voltage envelope as set).
I don't think that is a bad idea.  But certainly not a default behavior of an AWG.

[Fgrir]

I really shouldn't get in the middle of which mode is right or wrong, but I think my preference is to invert about the offset voltage.  In a perfect world I think there would be a control to let you do either.  A quick check with the function generator on my RTB2004 shows that its invert control works relative to the offset voltage, so there is at least one other example of a function generator that would invert +3.3/0V to 0V/+3.3V.

If you think the behavior of the SDG generators are correct, then they would also need to change their GUI representation of the signals when they are inverted.  I have to say that this GUI is often times not accurately representing the active settings, but for the invert control the GUI clearly agrees with the user manual.

EDIT: Fixed the invert attachment

[2N3055]

I really shouldn't get in the middle of which mode is right or wrong, but I think my preference is to invert about the offset voltage.  In a perfect world I think there would be a control to let you do either.  A quick check with the function generator on my RTB2004 shows that its invert control works relative to the offset voltage, so there is at least one other example of a function generator that would invert +3.3/0V to 0V/+3.3V.

If you think the behavior of the SDG generators are correct, then they would also need to change their GUI representation of the signals when they are inverted.  I have to say that this GUI is often times not accurately representing the active settings, but for the invert control the GUI clearly agrees with the user manual.

EDIT: Fixed the invert attachment

I'm not arguing that sometimes to someone because of work they do some options seems more logical.


Full name of that function is "Invert polarity".
Which explains it.
Like I said, if you are doing analog work, inverting polarity will accomplish same thing as adding inverting unity gain buffer in signal chain. You set one channel to track another and then invert one and you have fully differential signals.

RTB does what it does because it's dynamic range is already very low so they did that...

I reported documentation error to Siglent.
I also asked to consider enhancing this with more user control.
More user control is good.
GUI representation is more symbolic. It shows the shape but not offset.

We will see what they think.
In meantime, documentation needs to follow the reality.

[nctnico]

Yeah, I'm not saying it's good or bad to invert about 0V or the Offset voltage, but the instrument behavior does not agree with what the user manual says it should do.

User manual is wrong. I will report it.

Polarity flip is symmetrical around the 0V. It is multiplying voltage with -1.

The device is definitely wrong. There are lots of differential signalling protocols (LVDS, CAN and RS485 for example) which have a DC offset. With the way the SDG1000 is working right now, you can never generate such signals. IOW: the DC offset should be applied to both outputs in the same way, only the AC part of the signal should be inverted.

Don't arbitrarily invent things.
AWG is analog device, not digital pulse generator.

Hint: digital signals do not exist in the real world! If you want to generate any of the signalling protocols I listed before (especially for introducing analog domain errors), an AWG is the tool for that purpose.

[Fgrir]

GUI representation is more symbolic. It shows the shape but not offset.

Except the GUI does show the offset, and it disagrees with what is coming out of the output.

But I agree that the main important thing is that documentation, GUI and output all agree.

[2N3055]

Yeah, I'm not saying it's good or bad to invert about 0V or the Offset voltage, but the instrument behavior does not agree with what the user manual says it should do.

User manual is wrong. I will report it.

Polarity flip is symmetrical around the 0V. It is multiplying voltage with -1.

The device is definitely wrong. There are lots of differential signalling protocols (LVDS, CAN and RS485 for example) which have a DC offset. With the way the SDG1000 is working right now, you can never generate such signals. IOW: the DC offset should be applied to both outputs in the same way, only the AC part of the signal should be inverted.

Don't arbitrarily invent things.
AWG is analog device, not digital pulse generator.

Hint: digital signals do not exist in the real world! If you want to generate any of the signalling protocols I listed before (especially for introducing analog domain errors), an AWG is the tool for that purpose.

You are changing a subject by introducing truisms.
Of course that digital signals are analog signals. We are calling signals digital when we are referring to signals as used in various digital circuits, where that implies some sort of logic signals that have two defined voltages for LO/HIGH states and transitions between the two are made fast.
Some kind of square wave ish signals that have defined low and high voltages.

General purpose AWG is not specialized generator for generating LVDS, CAN and RS485 or that sort of data. Especially inexpensive one that costs same as cheap analog function generators few years ago. With little care you can create them but AWG are not specialized for that. Digital pattern and pulse generators are made for that.

Some AWG are made to do that too. SDG7000A has differential outputs and does the stuff you are saying. But it is expensive.

[2N3055]

GUI representation is more symbolic. It shows the shape but not offset.

Except the GUI does show the offset, and it disagrees with what is coming out of the output.

But I agree that the main important thing is that documentation, GUI and output all agree.

You are correct it show something, but it was not how I expected it so I didn't even realize. Thank you for that. Yes, it should show and document what it does.

[eTobey]

Either way: The low level that is set up, is exceeded. So its not just an error in the manual. 

[2N3055]

Either way: The low level that is set up, is exceeded. So its not just an error in the manual. 

You are wrong. Read again.

[TurboTom]

Interesting: DG800/900/2000 keep the offset / limits and invert only the signal amplitude.

[Njk]

That really make sense. With the inverted output setting, it's simple to generate a diff. signal manually (by tapping the Copy menu item) or automatically (with the track mode feature)

[thm_w]

Interesting: DG800/900/2000 keep the offset / limits and invert only the signal amplitude.

From the manual (Dg800 pro, which is the same behavior BUT it doesn't show on the UI like the siglent which would be nice):

[Njk]

I presume the intention was to show something like this

[eTobey]

Either way: The low level that is set up, is exceeded. So its not just an error in the manual. 

You are wrong. Read again.

I am talking about the SDG1032X that i got. It does not work like in the manual and i would have expected. And it does not correspond with the settings that there are.

Do they actually test these models at all?

[nctnico]

Either way: The low level that is set up, is exceeded. So its not just an error in the manual. 

You are wrong. Read again.

I am talking about the SDG1032X that i got. It does not work like in the manual and i would have expected. And it does not correspond with the settings that there are.

Do they actually test these models at all?

No they don't. While testing is very easy to automate using scripting. Just hook up an oscilloscope to the AWG and have a PC run through a bunch of tests which verify functional behaviour using SCPI commands send over the network. This setup does take effort to make but once that is done, the benefits are huge as testing firmware hardly takes any manual labour. It would greatly improve the quality of Siglent firmware where it comes to introducing new bugs in existing software and verifying functionality of new features.

[eTobey]

No they don't. While testing is very easy to automate using scripting.

But then they could just give some people some scopes to test for free?

I mean i would not mind doing some work for them in exchange ;-)

[tautech]

But then they could just give some people some scopes to test for free?

Only properly competent people need apply.