Siglent SDS800X HD 12 bit DSO's

[eTobey]

While trying to move the list of the decoding (at the arrow), i realized it can not be done this way. Only that line would be selected. I havent noticed any behaviour there. The horizontal position did not jump to a point (a decoded frame), of what i would have expected then.

The manual says nothing about this, but is there a function possible with that selected frame?

[tautech]

Packet 5 is the 5th after the off-display trigger position.
Is that not clear ?

[eTobey]

Packet 5 is the 5th after the off-display trigger position.
Is that not clear ?

That is absolutely clear (it was clear before), but that was not the question.

[tautech]

Packet 5 is the 5th after the off-display trigger position.
Is that not clear ?

That is absolutely clear (it was clear before), but that was not the question.

What function do you expect ?
Timing info for packet #5 is available from the table post trigger although you aren't using a decode trigger looking for a specific bit. 

[eTobey]

Packet 5 is the 5th after the off-display trigger position.
Is that not clear ?

That is absolutely clear (it was clear before), but that was not the question.

What function do you expect ?
Timing info for packet #5 is available from the table post trigger although you aren't using a decode trigger looking for a specific bit. 

Still not an answer to my question. 

[tautech]

While trying to move the list of the decoding (at the arrow), i realized it can not be done this way. Only that line would be selected. I havent noticed any behaviour there. The horizontal position did not jump to a point (a decoded frame), of what i would have expected then.

The manual says nothing about this, but is there a function possible with that selected frame?

No.
You pan to the listed timestamp with the H Pos encoder (or drag with touch/mouse) noting the time position in the timebase tab.

[eTobey]

You pan to the listed timestamp with the H Pos encoder (or drag with touch/mouse) noting the time position in the timebase tab.

That is not exactly great. The search feature does much better in this regard.

[MathWizard]

I got the SDS804X HD last week, and it's nice so far. I'm just using it to study a single capacitor, 2BJT astable multi-vibrator, from the 5-BJT ESR meter on here. My BJT's match the calculations and sim pretty good. I'm just looking at the rise times, and w/ CH2 and CH4 on.

I'm on 20us/div, the period is about 147us. It says CH4 rise time is 2.2us avg. CH2 still seems fine, whether I'm triggering from CH2 or 4, it's about 55ns.

At 10us/div, CH4 rise time is now 230ns, and at 5us/div it gets to about what it is, 113ns.

Hmm swapping probes CH2 is now worse, but now 20us/div gets me, CH2 246ns, CH4 55ns,

But 100us/div gets me 224ns, 55ns

and 50us/div gets me 18us, and 55ns. It does it with single shot too.

Is it just the downward slope before the sudden rise? And then it misses when the slope changes sign sometimes ?

I'm just probing the emitters, and swapped probes positions, between emitters, not on the scope BNC's, from pic18 upwards.

Now I can't remember if I updated the firmware, or did a CAL right away, but what's going on with the rise time ?

[MathWizard]

here's the last 3-4 pic's, but they didn't appear again

[2N3055]

Enable measurement cursors and set them on RiseTime measurement.
They will show you where exactly scope is measuring. It might give a clue of what is going on.

There are also settings for thresholds and also measurement gates that influence measurements too..

Start with cursors.

[MathWizard]

Ok and I'll try my other siglent scopes, see what they do. I'm sure I did a cal, but I did another 1 to be sure.

The CH4 probe has 100 or 200MHz BW, the CH2 was only 70MHz too, but yeah 100us did better than 50us/div. I never tried higher yet.

[Conrad Hoffman]

Not sure of the objective but measuring a capacitor is probably better done by applying a sine wave through a resistor and measuring the current, voltage and the phase shift between them. Then you can calculate pretty much everything of interest.

[Furna]

TrueNAS Scale ElectricEel-24.10.1 -> latest version
What I don't understand is the different oscilloscope's behaviour for PNG and BIN ...
Also keep in mind I am able to move the file to the NAS after it is saved locally using the oscilloscope File Manager; so no size limit or SMB version.

Are you able/willing to capture the network traffic on your Win10 (i.e. Wireshark)?

Thank you.

Thanks.

Sure I can capture the sniffer trace, when I get a bit of time...
Will let you know. Don't worry, we'll figure it out.

Can you send me CAP files you captured?
There is an email in my profile you can send it to.
Best,

Did you get any feedback from Siglent?

[2N3055]

TrueNAS Scale ElectricEel-24.10.1 -> latest version
What I don't understand is the different oscilloscope's behaviour for PNG and BIN ...
Also keep in mind I am able to move the file to the NAS after it is saved locally using the oscilloscope File Manager; so no size limit or SMB version.

Are you able/willing to capture the network traffic on your Win10 (i.e. Wireshark)?

Thank you.

Thanks.

Sure I can capture the sniffer trace, when I get a bit of time...
Will let you know. Don't worry, we'll figure it out.

Can you send me CAP files you captured?
There is an email in my profile you can send it to.
Best,

Did you get any feedback from Siglent?

Actually I didn't have time to investigate further... Sorry. I will soon, just busy now... Will let you know..

[MathWizard]

Well with my SDS1204X-E, and both 200MHz BW probes, it also has trouble, on certain time/div, with that rise time, that follows the long decreasing voltage section. It's just curious that sometimes the greater time/div, can be more accurate.

IDK, just something to watch out for I guess. I don't remember noticing this before tho. But with both scopes doing it, I guess nothing's broken.

[awakephd]

I hesitate to add anything, since I am still very much a newbie, but ... when I sat in on an electronics course last semester, when using the lab's Keysight oscilloscope, I encountered similar issues with regard to the reported measurements changing depending on the time-base. My conclusion was that the scope wasn't necessarily measuring what I thought it was measuring, and I began to pay more attention to the trace on the screen, using the "automatic" measurements only as a rough guide. This was especially true when the curve I was attempting to measure was more complex than a "simple" square wave - not that the latter is truly simple when getting down into the ns range ...

I don't know if my conclusion is correct - again, newbie here! - but hopefully a useful observation, FWIW.

[2N3055]

These scope support Measurement Cursors. You can assign cursors to a measurement, so it will show where exactly measurement is performed from. You can also fine tune thresholds and strategies. And also measurement gate...

[awakephd]

I noted your response about the cursors, and am eager to try out that function! I'm not entirely sure what it means to "set a cursor on the rise-time measurement," but figured I would hold off on questions until I sit down and try it out. Hopefully it will be fairly obvious ...

[2N3055]

I noted your response about the cursors, and am eager to try out that function! I'm not entirely sure what it means to "set a cursor on the rise-time measurement," but figured I would hold off on questions until I sit down and try it out. Hopefully it will be fairly obvious ...

Make sure to have manual ready. But it will be logical as soon as you see it.. But we are here... Either way you'll figure it out.

[eTobey]

I hesitate to add anything, since I am still very much a newbie, but ... when I sat in on an electronics course last semester, when using the lab's Keysight oscilloscope, I encountered similar issues with regard to the reported measurements changing depending on the time-base. My conclusion was that the scope wasn't necessarily measuring what I thought it was measuring, and I began to pay more attention to the trace on the screen, using the "automatic" measurements only as a rough guide. This was especially true when the curve I was attempting to measure was more complex than a "simple" square wave - not that the latter is truly simple when getting down into the ns range ...

I don't know if my conclusion is correct - again, newbie here! - but hopefully a useful observation, FWIW.

Its a good thing to not trust the scope right away. Measurements are done on less points on some occasions. Also some measurments are wrong, as like FRLF (a different one of these) for example. And also at some occasions it shows single ns, when the accuracy is only like a few hundrets ns. And measurments can be setup wrong. But you can use cursors to show what is measured. I have made a script, that lets one quickly switch between those measurements (the cursors).

[Fluvance]

I noticed something weird in my FFT the other day. When analyzing the full spectrum to 1GHz, I see a -38dB spike at exactly 500MHz, with some smaller harmonics(?) at 250 and 750 MHz. I included a screenshot of the normal FFT capture as well as one of the peak-hold.
Does anyone else see this on their SDS800X HD? Is this the result of an unshielded internal oscillator or some artifact of the sample rate?

I have the 200MHz unlocked model. I see the exact same results with the same amplitude whether the input channel is in 20M or full bandwidth, and whether there is a probe connected or not. Grounding the probe tip does not reduce the 500MHz peak. Setting the channel to ground reduces the peaks to -100dB but they are still present above everything else. I ran a self-cal and tried with factory default settings.

[2N3055]

I noticed something weird in my FFT the other day. When analyzing the full spectrum to 1GHz, I see a -38dB spike at exactly 500MHz, with some smaller harmonics(?) at 250 and 750 MHz. I included a screenshot of the normal FFT capture as well as one of the peak-hold.
Does anyone else see this on their SDS800X HD? Is this the result of an unshielded internal oscillator or some artifact of the sample rate?

I have the 200MHz unlocked model. I see the exact same results with the same amplitude whether the input channel is in 20M or full bandwidth, and whether there is a probe connected or not. Grounding the probe tip does not reduce the 500MHz peak. Setting the channel to ground reduces the peaks to -100dB but they are still present above everything else. I ran a self-cal and tried with factory default settings.

Where to start...

1. These are artefacts from ADC and sampling process.
2. Scope input BW is officially 200 MHz, but is more like 300Mhz. Anything more than 450 MHz on the screen is not coming in from BNC in the front.
3. Which means a peak at 500 MHZ is irrelevant. Simply ignore it. Just set FFT horizontal to show up to 400 MHz. Or 495MHz. There cannot be any of your signal higher up. Front end won't let it in.
4. Let's look at the levels.  At V/div and 10:1 probe you chose, full screen signal is 80 V P-P. That drills down to cca 29dBV FS.
If we set scope to show +30dBV on top of scale, it paints a different picture. It means the 500 MHz Peak (that we can safely ignore anyways) is 67dB down from full scale of scope. 

And in 0-400MHz range in that scope settings, a noisiest spur in the noise floor is at 250Mhz with -71dBV. That is pretty much exactly 100dB down from full scale....

So if anything that is excellent results...

[Fluvance]

I noticed something weird in my FFT the other day. When analyzing the full spectrum to 1GHz, I see a -38dB spike at exactly 500MHz, with some smaller harmonics(?) at 250 and 750 MHz. I included a screenshot of the normal FFT capture as well as one of the peak-hold.
Does anyone else see this on their SDS800X HD? Is this the result of an unshielded internal oscillator or some artifact of the sample rate?

I have the 200MHz unlocked model. I see the exact same results with the same amplitude whether the input channel is in 20M or full bandwidth, and whether there is a probe connected or not. Grounding the probe tip does not reduce the 500MHz peak. Setting the channel to ground reduces the peaks to -100dB but they are still present above everything else. I ran a self-cal and tried with factory default settings.

Where to start...

1. These are artefacts from ADC and sampling process.
2. Scope input BW is officially 200 MHz, but is more like 300Mhz. Anything more than 450 MHz on the screen is not coming in from BNC in the front.
3. Which means a peak at 500 MHZ is irrelevant. Simply ignore it. Just set FFT horizontal to show up to 400 MHz. Or 495MHz. There cannot be any of your signal higher up. Front end won't let it in.
4. Let's look at the levels.  At V/div and 10:1 probe you chose, full screen signal is 80 V P-P. That drills down to cca 29dBV FS.
If we set scope to show +30dBV on top of scale, it paints a different picture. It means the 500 MHz Peak (that we can safely ignore anyways) is 67dB down from full scale of scope. 

And in 0-400MHz range in that scope settings, a noisiest spur in the noise floor is at 250Mhz with -71dBV. That is pretty much exactly 100dB down from full scale....

So if anything that is excellent results...

Very interesting! I'm still new to using FFTs with a digital scope so I wasn't sure what I was seeing or how dB levels relate, nor if higher frequency signals could even make it in. I'll keep it under 500. Thanks for the reply     

[2N3055]

I noticed something weird in my FFT the other day. When analyzing the full spectrum to 1GHz, I see a -38dB spike at exactly 500MHz, with some smaller harmonics(?) at 250 and 750 MHz. I included a screenshot of the normal FFT capture as well as one of the peak-hold.
Does anyone else see this on their SDS800X HD? Is this the result of an unshielded internal oscillator or some artifact of the sample rate?

I have the 200MHz unlocked model. I see the exact same results with the same amplitude whether the input channel is in 20M or full bandwidth, and whether there is a probe connected or not. Grounding the probe tip does not reduce the 500MHz peak. Setting the channel to ground reduces the peaks to -100dB but they are still present above everything else. I ran a self-cal and tried with factory default settings.

Where to start...

1. These are artefacts from ADC and sampling process.
2. Scope input BW is officially 200 MHz, but is more like 300Mhz. Anything more than 450 MHz on the screen is not coming in from BNC in the front.
3. Which means a peak at 500 MHZ is irrelevant. Simply ignore it. Just set FFT horizontal to show up to 400 MHz. Or 495MHz. There cannot be any of your signal higher up. Front end won't let it in.
4. Let's look at the levels.  At V/div and 10:1 probe you chose, full screen signal is 80 V P-P. That drills down to cca 29dBV FS.
If we set scope to show +30dBV on top of scale, it paints a different picture. It means the 500 MHz Peak (that we can safely ignore anyways) is 67dB down from full scale of scope. 

And in 0-400MHz range in that scope settings, a noisiest spur in the noise floor is at 250Mhz with -71dBV. That is pretty much exactly 100dB down from full scale....

So if anything that is excellent results...

Very interesting! I'm still new to using FFTs with a digital scope so I wasn't sure what I was seeing or how dB levels relate, nor if higher frequency signals could even make it in. I'll keep it under 500. Thanks for the reply 

Thank you for putting in effort to read. If you need more explanations ask..
Truth is even super expensive instruments are not perfect.  Trick is to know it well, all the flaws and sweet spots where it performs better than specifications and use to your advantage...

[rf-loop]

The following is not in any way a Siglent specific thing, it is fully a general one for all.

Digital oscilloscopes often have FFT implemented more or less well.

Quite often, their frequency scale exceeds the Nyquist limit that prevails at the time of use.

I don't understand why this is the case. (yes I know, lazy designers and program "writers" who are not also paid for better work)

There is not the slightest bit of intelligence in that way of displaying. On the contrary, it causes a lot of confusion.

If a frequency component higher than the prevailing Nyquist frequency arrives in the system for digitization, it will be folded back from the Nyquist frequency in any case. There is no any real signal on the screen above the Nyquist frequency. As all know or is good to learn.  (of course, it also applies to that the f Nyquist frequency that has been lowered by possible decimation after the AD converter.)

However, why do many oscilloscopes have these over than f Nyquist FFT representations year after year.

That stupidity should be stopped. 

These are, or try to be,  Test and Measurement instruments, not Hollywood movies fun decorations.

Ten points for company who is the first and do it also very well.


Of course, this is just one thing among many other things that may be even more important if think whole oscilloscope what is lot of more than just one FFT.

This can easily be bypassed by the user by deciding not to watch anything that exceeds the prevailing f Nyquist frequency. Just ignore. 


Byt then agan FFT
The second question is more difficult. When the sampling is decimated to very low, the analog channel of the oscilloscope does not attenuate over f Nyquist freq and they are then in many kinds of mess on the FFT screen as folded back frequencies.  Hoe user know 4MHz peak is from 4MHz signal or from 6MHz signal if sampling was 10MSa/s. Or perhaps it is 14 MHz... and so on.

High-quality FFT work requires filtering out all frequencies above f Nyquist before digitization.
If the FFT display says 10MSa/s, the input frequency components must be limited to less than 5MHs and actually in practice a little below that.   

If this kind of analog filter system were built into an oscilloscope, such an oscilloscope would cost well more than a lot. So these do not exist.

So still user need be very careful when look oscilloscope FFT so that one don't get fooled  at least too much.
Of course, there is not much truth in the oscilloscope display anyway, not in time domain and not in freq. domain. Just the sum of many different and partially unknown errors and imagined truth.