Siglent SDG1032 high level limited to low ...

[eTobey]

Hi,

i had a go before, but i guess i am missing something?

I have set up amplitude limit to 6.6Vpp. But when setting the low level to 0V and the high level to 3.3V, it does limit to level to 1.65V.

I have also tried with higher amplitude limit, and it does work above 1.65V. But this limiting looks wrong to be.

The load is 50 Ohm, and so is the setting of the load in the device.

[KungFuJosh]

You should stop creating threads for questions about gear that already has threads. You should post this question in the SDG1032 main thread.

[Fgrir]

Is your output in 50 ohm load mode?  I think the amplitude limit is always based on the unloaded voltage, so with 50 ohm the output limit will be half of what you set.

[tautech]

Is your output in 50 ohm load mode?  I think the amplitude limit is always based on the unloaded voltage, so with 50 ohm the output limit will be half of what you set.

Most likely quite correct,  however as is usual the full setup details are never shared. 

[eTobey]

Is your output in 50 ohm load mode?  I think the amplitude limit is always based on the unloaded voltage, so with 50 ohm the output limit will be half of what you set.

Sorry, i forgot to mention that i have a 50 Ohm load, and that i have set up the load as 50 Ohm.

[ebastler]

It seems pretty clear then that the voltage limits are applied assuming a non-loaded output, as Fgrir suggested. Dividing the displayed actual voltages by two -- if the user declares the load to be 50 Ohms -- only happens afterwards.

Feels like an awkward design decision to me. If we assume that a conscious decision on the designer's part was involved, and not just negligence...

[newbrain]

Feels like an awkward design decision to me.

Feels, I would say, like a safe decision.
Showing the voltage limit as if a 50 ohm load was connected could cause the user to apply double the intended maximum, if the DUT has a high impedance.

[ebastler]

Feels like an awkward design decision to me.

Feels, I would say, like a safe decision.
Showing the voltage limit as if a 50 ohm load was connected could cause the user to apply double the intended maximum, if the DUT has a high impedance.

So now the user, planning to connect a 50 Ohm load, has to set double the planned maximum voltage as his upper safety limit. And if he then forgets to actually connect his load, he will fry his DUT again.

Catch 22... I'm afraid the current design does not add more safety, it just adds more confusion.

[eTobey]

Feels like an awkward design decision to me.

Feels, I would say, like a safe decision.
Showing the voltage limit as if a 50 ohm load was connected could cause the user to apply double the intended maximum, if the DUT has a high impedance.

No? It shows a 1.65V limit, and when i remove the 50 Ohm load, and suspect a 1.65V limit, i would get 3.3V.

[newbrain]

it just adds more confusion.

The way I had understood the quite unclear statement from the OP:

I have set up amplitude limit to 6.6Vpp. But when setting the low level to 0V and the high level to 3.3V, it does limit to level to 1.65V.

was that the set limit was always the maximum (and minimum) voltage the generator was going to output, regardless of the load/DUT impedance.

So, to me, the safest choice.

Maybe it's not so, once more the OP

just adds more confusion

with their second statement:

It shows a 1.65V limit, and when i remove the 50 Ohm load, and suspect a 1.65V limit, i would get 3.3V.

So, 

[2N3055]

What is going on here?
 

[thm_w]

Wasn't the purpose of making a clipping circuit so you didn't have to screw around with these limit settings?
I've literally never heard or used a limit mode on a sig-gen until it was brought up recently.

Wonder what the bandwidth of those current limiting diodes is: https://mm.digikey.com/Volume0/opasdata/d220001/medias/docus/380/CLD20B.pdf
Probably not high enough.

[ebastler]

What is going on here?
 

We are discussing a little design issue of a Siglent product and waiting for a white knight to come to Siglent's defense. 

[ebastler]

Maybe it's not so, once more the OP

just adds more confusion

with their second statement:

It shows a 1.65V limit, and when i remove the 50 Ohm load, and suspect a 1.65V limit, i would get 3.3V.

Agree, that post does not make sense to me. Seems to describe things the wrong way round.

My understanding is that the "safety limits" always apply to the voltage the generator's output stage is allowed to provide. If a 50 Ohm load is connected, the actual maximum output voltage will be cut in half. I stand by my earlier comment that this requires a work-around if a user wants to use 50 Ohm loads, requiring them to set twice the safety limit -- and hence remains open to the risk of mishaps if the load is not actually present.

[2N3055]

What is going on here?
 

We are discussing a little design issue of a Siglent product and waiting for a white knight to come to Siglent's defense. 

None of it makes sense..
There is no design issue.
Just complete lack of understanding how it works. Even by you.

Newbrain tried to explain...

And only white knight coming to someone's defense is you defending eTobeys right to publicly post his weird questions that make no sense.
And you keep doing it systematically lately...
No matter how wrong the statements of his are.

[rf-loop]

What is going on here?
 

We are discussing a little design issue of a Siglent product and waiting for a white knight to come to Siglent's defense. 

None of it makes sense..
There is no design issue.
Just complete lack of understanding how it works. Even by you.

Newbrain tried to explain...

And only white knight coming to someone's defense is you defending eTobeys right to publicly post his weird questions that make no sense.
And you keep doing it systematically lately...
No matter how wrong the statements of his are.

It works as designed. But why there exist lack of understanding how it works.

If the user knew what the designer knew / thought, then he would (probably) understand and master its use and find that it works as designed. Including a couple of small details.

Actually, the confusion boils down to the screen where the amplitude limit is set. It can cause confusion. It lacks essential information that would explain its operation more clearly Of course, it doesn't bother us who know how it works.

Example: I set the limit there (Utility > Output Setup > Amplitude) to, for example, 10,000 Vpp.


(I know that, and many others who know it, that) it means the maximum Vpp voltage of the open BNC output port.
It means the maximum voltage out from output BNC even if load is removed. It limits all other settings for keep this true.

It need clarify even more.
 
It mean limits: Lower limit is -([Amplitude Vpp]/2)  and upper limit is +([Amplitude Vpp]/2) in BNC output without load (aka open).
[Amplitude Vpp] here is value in:  Utility > Output Setup > Amplitude

Including situations where the user has selected HiZ, 50 ohm or some other value between 50 ohm to 100 kohm as the external load value. (Yes it is naturally free adjustable)

Example if user have selected now 50 ohm load impedance and this Amplitude limit is 10.000Vpp. He can now set example max 5Vpp square wave with zero offset. In this case to 50 ohm load square wave bottom is -2.5V and top +2.5V
Now if 50ohm load breaks (open)  now output voltage is bottom -5V and top +5V. (10Vpp)
It is clear that it may be confusing if user do not know how it is designed to work and only information is in meny Utility > Output Setup > Amplitude one Vpp value without any explanation. And now also if then try use Low Level/High level setup or Level and offset it may produce confusion.
Only thing what really deen hard code to user brain is. Instead of Vpp limits it set +V limit and -V limit for Open output (HiZ)
Level values  what can then select in signal paramaters depends what is defined Load and other things what affect signal levels. These values limits are calculated so that if then load goes infinite still levels out from generator do not go over these Amplitude limit. 
it works with dBm values, with RMS values, Voltage levels, with all load values what can define for Load. with al AM modulation levels... sorry if I forget something here...

The user may also choose, for example, AM modulation. This is also important to note due to Siglent SDG's AM modulation level principle (Many other AWG do not use this principle).

With all its available modulation depths, the maximum open gate voltage can therefore be between +5V and -5V. (also note that 120% modulation depth is an option)


If the settings screen read additional information that it specifically sets the hard "open load" limits (in this example +5 V and -5 V), the user might understand it better.
(and User manual)

HiZ is used (and, as example, Amplitude limit set for 10Vpp)
Now I want to output a 0 V to 5 V square wave.   I can now set Amplitude 5 Vpp
Offset 2.5 Vdc. Now there is 0 - 5V square.

But now if want 6Vpp square but so that it is between 1V and 6V, with this setup it is not possible.  Then need to change  Amplitude limit to least 12Vpp
Because then it is -6V and +6V

It is very simply if these limits are told in this Amplitude limit setup screen (and of course also in user manual)
This is because also users do not have clairvoyant skills.

What should be improved is the information on the device's setting screen and the user's manual.


At least after you know that it applies to the open output and all load setting values and, among other things, all modulation depths.

Note: This AM modulation is included here separately because Siglent SDG follows the usual principle of RF generators. The output level is set to the level of the carrier wave. Not sum level. When adjusting the modulation depth, the carrier level does not change but sum level change and max is reached when Max available AM depth 120% is in use. So max signal level is rejected so that if user turn AM depth to 120%, then it do not break maximum Amplitude limit (if there is no output load, aka open BNC).

[ebastler]

Thanks for the detailed explanation, @rf-loop. But I think the functionality was understood before -- Fgrir had stated the key principle in reply #2, and I had followed up in #5.

While the internal workings are certainly consistent, I still don't think they are the clearest way to present "output level limiter" functionality to the user. I would consider it much more straightforward, from a user perspective, if the output limit I set is always the one the generator aims to adhere to -- i.e. when I set 50 Ohm outputs, the limiter should internally double its value, such that the voltage fed into 50 Ohms remains the nominal one set in the limiter UI. Also, I still don't see how Siglent's implementation is more fail-safe -- see my reply #7.

Anyway, this probably falls into the category of "don't break existing user habits" -- so it won't be changed even if I could convince you that a change would result in a more logical interface for new users. I rest my case.

[nctnico]

Thanks for the detailed explanation, @rf-loop. But I think the functionality was understood before -- Fgrir had stated the key principle in reply #2, and I had followed up in #5.

While the internal workings are certainly consistent, I still don't think they are the clearest way to present "output level limiter" functionality to the user. I would consider it much more straightforward, from a user perspective, if the output limit I set is always the one the generator aims to adhere to -- i.e. when I set 50 Ohm outputs, the limiter should internally double its value, such that the voltage fed into 50 Ohms remains the nominal one set in the limiter UI. Also, I still don't see how Siglent's implementation is more fail-safe -- see my reply #7.

Agreed. More likely it is a bug or oversight in Siglent's firmware. I just tried on my Tektronix AFG31000. The limits scale along with the load setting which is much more consistent as the limits are always 1 on 1 related to the amplitude and offset settings (which are scaled for the configured load impedance). It is up to the user to make sure a load with the correct impedance is connected. As Tektronix is a company with a very long history of designing high end AWGs, I trust they got it right based on decades of experience.

[2N3055]

Voltage limits will be applied the same, regardless of load settings.

Because unloaded AWG will always have same output voltage, regardless if you set it to High-Z  and 4V P-P and 50Ω and 2V P-P. It will be 4V P-P, and on 50Ω it will become 2V P-P only after you load it with 50Ω.

So if you are working with Hi impedance load (non 50Ω) or with load that is changing from 100Ω to say 1000Ω impedance, voltage on output will vary, with maximum reaching Hi-Z value unloaded.

And that is how every AWG or signal generator works, unless you are using leveled gen or some special AWG (expensive) that does leveling.

[2N3055]

Thanks for the detailed explanation, @rf-loop. But I think the functionality was understood before -- Fgrir had stated the key principle in reply #2, and I had followed up in #5.

While the internal workings are certainly consistent, I still don't think they are the clearest way to present "output level limiter" functionality to the user. I would consider it much more straightforward, from a user perspective, if the output limit I set is always the one the generator aims to adhere to -- i.e. when I set 50 Ohm outputs, the limiter should internally double its value, such that the voltage fed into 50 Ohms remains the nominal one set in the limiter UI. Also, I still don't see how Siglent's implementation is more fail-safe -- see my reply #7.

Agreed. More likely it is a bug or oversight in Siglent's firmware. I just tried on my Tektronix AFG31000. The limits scale along with the load setting which is much more consistent as the limits are always 1 on 1 related to the amplitude and offset settings (which are scaled for the configured load impedance). It is up to the user to make sure a load with the correct impedance is connected. As Tektronix is a company with a very long history of designing high end AWGs, I trust they got it right based on decades of experience.

You keep forgetting to note to beginners and those that don't know that you are comparing a 300 € AWG to a 19000 € one..
Smooth..

[nctnico]

Thanks for the detailed explanation, @rf-loop. But I think the functionality was understood before -- Fgrir had stated the key principle in reply #2, and I had followed up in #5.

While the internal workings are certainly consistent, I still don't think they are the clearest way to present "output level limiter" functionality to the user. I would consider it much more straightforward, from a user perspective, if the output limit I set is always the one the generator aims to adhere to -- i.e. when I set 50 Ohm outputs, the limiter should internally double its value, such that the voltage fed into 50 Ohms remains the nominal one set in the limiter UI. Also, I still don't see how Siglent's implementation is more fail-safe -- see my reply #7.

Agreed. More likely it is a bug or oversight in Siglent's firmware. I just tried on my Tektronix AFG31000. The limits scale along with the load setting which is much more consistent as the limits are always 1 on 1 related to the amplitude and offset settings (which are scaled for the configured load impedance). It is up to the user to make sure a load with the correct impedance is connected. As Tektronix is a company with a very long history of designing high end AWGs, I trust they got it right based on decades of experience.

You keep forgetting to note to beginners and those that don't know that you are comparing a 300 € AWG to a 19000 € one..
Smooth..

It is not 19000. It is much cheaper than that and you know it. But still the truth remains that a company with a decades worth of experience (pedigree) in the field can be trusted to implement things the right way so this is a good point of reference instead of making all kinds of wild assumptions or trying to tell people bananas are always straight.

[2N3055]

Thanks for the detailed explanation, @rf-loop. But I think the functionality was understood before -- Fgrir had stated the key principle in reply #2, and I had followed up in #5.

While the internal workings are certainly consistent, I still don't think they are the clearest way to present "output level limiter" functionality to the user. I would consider it much more straightforward, from a user perspective, if the output limit I set is always the one the generator aims to adhere to -- i.e. when I set 50 Ohm outputs, the limiter should internally double its value, such that the voltage fed into 50 Ohms remains the nominal one set in the limiter UI. Also, I still don't see how Siglent's implementation is more fail-safe -- see my reply #7.

Agreed. More likely it is a bug or oversight in Siglent's firmware. I just tried on my Tektronix AFG31000. The limits scale along with the load setting which is much more consistent as the limits are always 1 on 1 related to the amplitude and offset settings (which are scaled for the configured load impedance). It is up to the user to make sure a load with the correct impedance is connected. As Tektronix is a company with a very long history of designing high end AWGs, I trust they got it right based on decades of experience.

You keep forgetting to note to beginners and those that don't know that you are comparing a 300 € AWG to a 19000 € one..
Smooth..

It is not 19000. It is much cheaper than that and you know it. But still the truth remains that a company with a decades worth of experience (pedigree) in the field can be trusted to implement things the right way so this is a good point of reference instead of making all kinds of wild assumptions or trying to tell people bananas are always straight.

Really?

 2ch 25MHz version is 5000 USD..

And there is nothing wrongly implemented. Different is not wrong.
Just a beginner that blames lack of understanding to problem with device.
If he had that Tektronix AFG31000 of yours he would still have as many topics open just for different things.

What Siglent needs to improve is better documentation.

[KungFuJosh]

It's a silly argument to pretend that one brand's pedigree is superior to any others. All the major A brands have many areas of workflow that are unique to their brand. That's been true for decades.

[mawyatt]

Was going to post something wrt to what a "Waveform" actually represents, but then realized what Josh just mentioned, this is just a Silly Argument/Discussion anyway!!

Best,

[nctnico]

It's a silly argument to pretend that one brand's pedigree is superior to any others. All the major A brands have many areas of workflow that are unique to their brand. That's been true for decades.

It is not a silly argument. Pedigree means having incorporated decades of customer feedback in a design. It is utterly foolish to ignore that and go ahead re-inventing a square wheel. If you want to try something in a different / new way, you really need to be sure to understand your customers expectations and/or manage them. Making the documentation match the device's functionality is the first step. The existence of this thread is proof that inexperienced engineers get tripped up by these little things and can waste massive amounts of time. I've experienced far worse situations with misbehaving 'cheap' equipment ended up costing thousands of euros, so yes, I'm nitpicking at the seemingly little issues.