Siglent SDG1032 as bad as rigol DG812?

[eTobey]

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

Only properly competent people need apply.

Really? How do they determine the proper competence?

[eTobey]

I have something that could really be interesting for siglent:

I am thinking about the possibility, of relays switching off after device power off, so that the voltage of a higher voltage rail is put onto a lower voltage rail. Obviously it does not do damage immediately, but i could over time.

[nctnico]

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?

Siglent is doing that but it doesn't work at all as the test coverage is too limited. I did some in-depth oscilloscope reviews in the past and going through the functions 'quickly' takes nearly a work-week worth of test time already. Doing a full functional test by going through a test plan with a decent coverage manually likely takes 3 to 4 weeks full-time for a modern DSO. And each firmware release needs a full re-test. Where it comes to complex pieces of equipment, you really need to do automated testing in order to verify existing and new functionality before releasing firmware with confidence that it works to a satisfactory level. Testing also reveals issues which can be listed under 'known issues' which can be fixed later or never.

[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?

Siglent is doing that but it doesn't work at all as the test coverage is too limited. I did some in-depth oscilloscope reviews in the past and going through the functions 'quickly' takes nearly a work-week worth of test time already. Doing a full functional test by going through a test plan with a decent coverage manually likely takes 3 to 4 weeks full-time for a modern DSO. And each firmware release needs a full re-test. Where it comes to complex pieces of equipment, you really need to do automated testing in order to verify existing and new functionality before releasing firmware with confidence that it works to a satisfactory level. Testing also reveals issues which can be listed under 'known issues' which can be fixed later or never.

Yeah, indeed i know how tedious it can be to test firm/software. But if someone who hasnt got a glue, and this scope for only a few weeks, can find dozends of bugs and issues, then, there is room for improvement. 

[eTobey]

I have made some more investigation:
I have put a piece of piezo on the BNC connector, to have a exact time (more or less) on the scope, when the relais are switching.

On the picture "Turn off spike 2" you can see how it looks when turning it off. The green trace is the piezo, that captures the "shockwaves".

The picture "2.3V to 3.3V 50Ohm" shows just a change of the outputvoltage (when relais do switch).

I used the change of the voltages, to see exactly, how the green trace looks when relais switch. It has a characteristic of 2 bumps ("2-3V bumbs.png"). I believe, that these 2 bumbs are actually the contact hitting each other (the opposite not connected ones).

This together with the capture of the turn off spikes ("first bump on turnff.png"), lets me conclude, that this is not a case of a relais "failing" to early, but a chip that would control the voltages of the output.

Furthermore, i made another discovery, that one may find good to know: in track mode, the channels do not switch simultaneously. Funny thing is, that you otherwise even cant have different load outputs. ("channel 2 leads.png")

Ps. Yes, it looks a bit weird to do all that investigation, but i just choosed to investigate its flaws, instead destroying something because of these flaws. 

[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.

You are changing a subject by introducing truisms.

I'm not introducing truisms. You are just so horribly wrong by being blinded by what you perceive as digital signals. You really need to seperate your thinking towards considering the output signal of an AWG as multiple components. For starters: A) a waveform and B) a DC offset. These are two very different things. When inverting the waveform, the phase is shifted by 180 degrees. The DC offset stays the same. On a decent AWG like the Tektronix AFG31000, you'll see that the waveform gets inverted but not the offset. As Tektronix has decades of experience with designing AWGs, you can rest assured they got this basic feature implemented correctly. What the SDG1032 is doing is inverting both the waveform and the offset which is the wrong thing to 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.

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.

I'm not introducing truisms. You are just so horribly wrong by being blinded by what you perceive as digital signals. You really need to seperate your thinking towards considering the output signal of an AWG as multiple components. For starters: A) a waveform and B) a DC offset. These are two very different things. When inverting the waveform, the phase is shifted by 180 degrees. The DC offset stays the same. On a decent AWG like the Tektronix AFG31000, you'll see that the waveform gets inverted but not the offset. As Tektronix has decades of experience with designing AWGs, you can rest assured they got this basic feature implemented correctly. What the SDG1032 is doing is inverting both the waveform and the offset which is the wrong thing to do.

No it is not.
It behaves differently.
I will repeat for 3rd time: Function is called INVERT POLARITY.
It inserts inverting unity buffer in signal chain.   If you don't know what that is, you need to go read..
Phase shifting signal by 180° makes no sense for single signal. It is still the same. Only frame of reference changes and that makes sense only compared to second channel, and for that you have 180° phase delay command.

What you have on Tek does not create fully differential signal in analog sense. In analog sense differential pair is fully inverted including DC component of input signal. If you subtract both, you get same signal as positive one but twice the amplitude. Otherwise you get DC component subtracted out. Which is not the original signal.

What you are talking about is what digital differential drivers do: they have differential signal riding on the common mode DC offset. And that common DC offset is unwanted and gets suppressed. And I agree the way your Tek works makes it simple to do that.

And since this is something useful to you you that makes it the only way that is right...
How typical of you.

What Signal did wrong is that when you invert signal it won't show voltages as negative in GUI.
That is a mistake.

And to make sure I don't disagree that the way your Tek work is useful for some type of work. You can achieve same thing on Siglent now by Inverting and entering negative offset.

But being able to choose which way it should work would be nice.
And I did report this to Siglent, documentation errata, GUI not showing negative voltages when inverted, and also a request to add the way your Tek works as additional operational mode.
Which if they manage to add that mode, it will be able to do both what Tek can and what Tek cannot do...

[KungFuJosh]

As Tektronix has decades of experience with designing AWGs, you can rest assured they got this basic feature implemented correctly. What the SDG1032 is doing is inverting both the waveform and the offset which is the wrong thing to do.

100% correct! But missing qualifications. Tek didn't set any standards that other brands are beholden to. What Tek did is what is expected by Tek users. They don't dictate the behavior of Siglent or any other brand. It's not right or wrong, it's brand specific design choices.

[nctnico]

As Tektronix has decades of experience with designing AWGs, you can rest assured they got this basic feature implemented correctly. What the SDG1032 is doing is inverting both the waveform and the offset which is the wrong thing to do.

100% correct! But missing qualifications. Tek didn't set any standards that other brands are beholden to. What Tek did is what is expected by Tek users. They don't dictate the behavior of Siglent or any other brand. It's not right or wrong, it's brand specific design choices.

There are things like industry standards as set by the manufacturers who pioneered the type of instrument. Stick to those standard and people will have no problem using your equipment.

Have you ever noticed that most (if not all) range selectors on auto-ranging handheld multi-meters operate clock-wise from the off position into the measurement mode and start with measuring voltage? Why would that be? Which standard dictates that?

[2N3055]

As Tektronix has decades of experience with designing AWGs, you can rest assured they got this basic feature implemented correctly. What the SDG1032 is doing is inverting both the waveform and the offset which is the wrong thing to do.

100% correct! But missing qualifications. Tek didn't set any standards that other brands are beholden to. What Tek did is what is expected by Tek users. They don't dictate the behavior of Siglent or any other brand. It's not right or wrong, it's brand specific design choices.

There are things like industry standards as set by the manufacturers who pioneered the type of instrument. Stick to those standard and people will have no problem using your equipment.

Have you ever noticed that most (if not all) range selectors on auto-ranging handheld multi-meters operate clock-wise from the off position into the measurement mode and start with measuring voltage? Why would that be? Which standard dictates that?

Cultural biases.  Didn't you notice all knobs have same layout of zero being left and clockwise to max.  And voltage is first because that was traditionally most used function.

There are many de facto standards, but they are not mandatory. And also not universally applicable always.

[mawyatt]

Have you ever noticed that most (if not all) range selectors on auto-ranging handheld multi-meters operate clock-wise from the off position into the measurement mode and start with measuring voltage? Why would that be? Which standard dictates that?

Cultural biases.  Didn't you notice all knobs have same layout of zero being left and clockwise to max.  And voltage is first because that was traditionally most used function.

There are many de facto standards, but they are not mandatory. And also not universally applicable always.

Also, wouldn't want a handheld that rotated first to ohms, then diode, then current, then voltage   

Best,

[Mortymore]

The "Invert" in my old analog generator works as nctnico described, as does some Tektronix, Tenma, Kenwood, and others that I know of.

My old Fordham generator can be seen here.
https://www.eevblog.com/forum/testgear/siglent-sdg1000x-waveform-generators/msg5504575/#msg5504575

However, Siglent is in its right to do whatever they please. What's not right is that that the generator does not behave as advertised in the manual and as represented on screen, posing a potential risk of destruction of a DUT by inverting the DC offset. It doesn't matter if one is used to Tektronix way of invert or not. The displayed image is wrong and it's up to Siglent to decide what they really meant by invert polarity, but it's an urgent fix, be it at the image and manual or the outputted signal. They must match.

[KungFuJosh]

There are things like industry standards as set by the manufacturers who pioneered the type of instrument. Stick to those standard and people will have no problem using your equipment.

Have you ever noticed that most (if not all) range selectors on auto-ranging handheld multi-meters operate clock-wise from the off position into the measurement mode and start with measuring voltage? Why would that be? Which standard dictates that?

IIRC HPAK pioneered signal generator type TE before their competitor Tek started making any. Both brands do things differently, it's not new information.

Like you said, most, not all. There are many touch screen meters now that don't use a dial at all. Or just use button arrays like the DE-5000. Use what you like, and what you're comfortable with, but don't expect everybody to do the same thing. If everything was the same, only the least expensive devices would sell.

If I get new TE, I RTFM and learn what's different about it. Just going from the 3045X/3055 to the SDM3065X required some learning for me. It's fine...as long as the manual makes sense. 😉

[eTobey]

Another issue:
Low level set to 1mV is about right, but setting it to 0 mV comes out at about -40mV. One step further on the knob, and *boom* the magic smoke of the DUT is gonna escape.



Behaves like a 50$ chinese thing at some points.

[Njk]

Thank you for not using red font to announce your findings