SDS800X HD Bug Reports + Firmware

[tautech]

I though the problem with the SCPI export was in the list...

SCPI export of decoded frames/bytes is incomplete:
There are 78 bytes that are captured, and that appear in the list on the scope, but the export does cut it short.

I tested it well with NI VISA, it seems that the web is limiting the return of data. You can use Python to read the data and save it.
My opinion is that the web is not a good analysis tool, such as obtaining decades of megabytes of data, Python is a good way to process data.

There are other ways to do such things when you get to know your scopes better. 
 
BIN files that can easily be converted to CSV with the onboard file convertor tool and SMB File Share can make the best use of the available connectivity.

[2N3055]

I though the problem with the SCPI export was in the list...

SCPI export of decoded frames/bytes is incomplete:
There are 78 bytes that are captured, and that appear in the list on the scope, but the export does cut it short.

I tested it well with NI VISA, it seems that the web is limiting the return of data. You can use Python to read the data and save it.
My opinion is that the web is not a good analysis tool, such as obtaining decades of megabytes of data, Python is a good way to process data.

There are other ways to do such things when you get to know your scopes better. 
 
BIN files that can easily be converted to CSV with the onboard file convertor tool and SMB File Share can make the best use of the available connectivity.

Both BIN files and SCPI have their own merits.

SCPI is used when you don't want to save file, open it and parse it before being able to analyse it.
With SCPI you pull data directly into your program's memory...

BIN is useful to send data to others and save it for later.
SCPI is good for online analysis.

[ebastler]

There are other ways to do such things when you get to know your scopes better. 

That sounds like a lame excuse for a feature that is not quite working as advertised (i.e. SCPI delivering incomplete results over the web interface). Does "know your scope better" mean "know how to work around the scope's little defects" here? Or was it just meant as a little condescending side note?

[KungFuJosh]

There are other ways to do such things when you get to know your scopes better. 

That sounds like a lame excuse for a feature that is not quite working as advertised (i.e. SCPI delivering incomplete results over the web interface). Does "know your scope better" mean "know how to work around the scope's little defects" here? Or was it just meant as a little condescending side note?

Has anybody more experienced confirmed the SCPI issue? Perhaps somebody with a better network setup, and possibly using something better than a web browser to run the SCPI? There's too many variables, especially user error, to automatically assume there is actually an issue.

[ebastler]

Has anybody more experienced confirmed the SCPI issue? Perhaps somebody with a better network setup, and possibly using something better than a web browser to run the SCPI? There's too many variables, especially user error, to automatically assume there is actually an issue.

I had reproduced it with FW 1.1.3.3. I haven't re-tried with the new FW 1.1.3.6 yet, and am away from my scope at the moment. But the mangled data look the same as before in eTobey's recent screenshot (post #73); so the results appear unchanged when requesting data over the SCPI web interface.

[eTobey]

There are other ways to do such things when you get to know your scopes better. 
 
BIN files that can easily be converted to CSV with the onboard file convertor tool and SMB File Share can make the best use of the available connectivity.

So many features... i would be satisfied with less features, but then flawless working ones. Quality over quantitiy. 

[KungFuJosh]

So many features... i would be satisfied with less features, but then flawless working ones. Quality over quantitiy. 

Easy! Stop using the messed up features and pretend they don't exist. 😉

[eTobey]

Found an issue: Opposite gate gets set to screen border
When changing gate A, while gate B is outside of the screen (zoomed in on gate A), gate B gets set to the time postion at the border of the current screen.

Edit:
i thought its an obvious issue, and the description was sufficient to explain the issue, but it seems i was wrong, so i try different:

- Capture a waveform
- Set gate cursors at 10us and 100us
- Zoom in on cursor at 10us so that cursor at 100us is out of screen
- change  cursor from 10us to like 9.9us
- zoom out, and observe, that the other cursor is not at 100us anymore

[2N3055]

Found an issue: Opposite gate gets set to screen border
When changing gate A, while gate B is outside of the screen (zoomed in on gate A), gate B gets set to the time postion at the border of the current screen.

Another bug in your brain.
There is nothing after the end of the screen..
You cannot set cursor to data that doesn't exist.

[2N3055]

Also this SCPI thing.

Stop saying SCPI bug.
SCPI works perfectly when used as SCPI.

Problem is that web interface cannot handle lots of data.
It is not supposed to be used to grab 2MB of buffer data, but for simple commands and status.

It might be a Java component problem in a browser.

Will report but grabbing data from scope this way is abusing poor browser.

[KungFuJosh]

Also this SCPI thing.

Stop saying SCPI bug.
SCPI works perfectly when used as SCPI.

Problem is that web interface cannot handle lots of data.
It is not supposed to be used to grab 2MB of buffer data, but for simple commands and status.

It might be a Java component problem in a browser.

Will report but grabbing data from scope this way is abusing poor browser.

Exactly why I suggested a "more experienced" user try SCPI commands in something besides a browser.

[eTobey]

Another bug in your brain.
There is nothing after the end of the screen..
You cannot set cursor to data that doesn't exist.

That does not sound, that you tried it yourself, but i am in good faith, that you just did not understand, so i updated my post.
https://www.eevblog.com/forum/testgear/sds800x-hd-bug-reports-firmware/msg5520580/#msg5520580

[ebastler]

Exactly why I suggested a "more experienced" user try SCPI commands in something besides a browser.

electronics hobbyist has done that, a few posts above. It seems clear that only the SCPI-over-HTTP (if that is how it's done for the web interface?) is affected.

[eTobey]

Also this SCPI thing.

Stop saying SCPI bug.
SCPI works perfectly when used as SCPI.

Problem is that web interface cannot handle lots of data.
It is not supposed to be used to grab 2MB of buffer data, but for simple commands and status.

It might be a Java component problem in a browser.

Will report but grabbing data from scope this way is abusing poor browser.

Exactly why I suggested a "more experienced" user try SCPI commands in something besides a browser.

I can confirm, that SCPI works better with python.

[KungFuJosh]

Exactly why I suggested a "more experienced" user try SCPI commands in something besides a browser.

electronics hobbyist has done that, a few posts above. It seems clear that only the SCPI-over-HTTP (if that is how it's done for the web interface?) is affected.

It's not a bug of the scope, it's limitations of the web browser as explained well above by 2N3055.

I can confirm, that SCPI works better with python.

Nice work!

[Furna]

It's not a bug of the scope, it's limitations of the web browser as explained well above by 2N3055.

Well, I still do not own the scope yet but ...

I would think twice before blaming a modern browser; Java is not used in browsers since a few years; Javascript is a completely different beast.
Browsers *are* capable of handling a lot of data; just open Youtube home page and continue to scroll down.
You can watch a 4K movie and that could be a 100Mbit/sec of data.
It's the interaction between the web server and the browser and I would say the web server was not designed with that scope in mind.

I totaly agree for some kind of jobs Python could be better but if I am not wrong here we are speaking about 78 byte?
78byte are 78 characters ... I will not count the characters illustrated in eTobey's post but I guess they are more than 78.
Even 78kB is a small amount; a modern web page's HTML is more tahn that.

IMHO issue should be definitely investigated.

[ebastler]

It's not a bug of the scope, it's limitations of the web browser as explained well above by 2N3055.

That was not an explanation but a hypothesis, and an unfounded one. If I recall correclty, when I reproduced the problem a month (?) ago I could do that with very small datasets, just a dozen bytes or so.

[KungFuJosh]

It's not a bug of the scope, it's limitations of the web browser as explained well above by 2N3055.

That was not an explanation but a hypothesis, and an unfounded one. If I recall correclty, when I reproduced the problem a month (?) ago I could do that with very small datasets, just a dozen bytes or so.

Read what he said again:

Problem is that web interface cannot handle lots of data.
It is not supposed to be used to grab 2MB of buffer data, but for simple commands and status.

How does your small dataset test NOT support the hypothesis that the browser console is the issue? Problem with too much data, no problem with smaller data sets.

Try connecting with SCPI through any other means than the web browser and see if you have trouble. The browser console wasn't designed for the use you're trying. Could it be improved? Probably. That doesn't make it a bug if you're trying to use it beyond the capabilities it was designed for. Everybody knows the SCPI web console was only created to "improve" the scopes. 😉😉

[2N3055]

It's not a bug of the scope, it's limitations of the web browser as explained well above by 2N3055.

Well, I still do not own the scope yet but ...

I would think twice before blaming a modern browser; Java is not used in browsers since a few years; Javascript is a completely different beast.
Browsers *are* capable of handling a lot of data; just open Youtube home page and continue to scroll down.
You can watch a 4K movie and that could be a 100Mbit/sec of data.
It's the interaction between the web server and the browser and I would say the web server was not designed with that scope in mind.

I totaly agree for some kind of jobs Python could be better but if I am not wrong here we are speaking about 78 byte?
78byte are 78 characters ... I will not count the characters illustrated in eTobey's post but I guess they are more than 78.
Even 78kB is a small amount; a modern web page's HTML is more tahn that.

IMHO issue should be definitely investigated.

If you want to play with browsers capabilities you can do that without scope.
This is tangential discussion about mating rituals of Africa swallows..

eTobey claims something is wrong with SCPI on the scope.
There is not, because if you use propper SCPI client, there are no errors.

So problem is with SCPI client component on the web page.

And while technically that is a bug (in web interface) that is definitely abuse of simple SCPI component that was designed to be simple interface for simple commands.. It is used for getting scope ID, viewing or installing options and such.

Using that small window to get data from the scope is not very smart or usefull on many levels.
Firstly, it was never meant to be used for that.
Second, why on earth you want to do that?
You by hand type in SCPI command and then you copy result to clipboard and paste it into some editor to save to disk... Err, no thank you.

If you want to pull SCPI data you pull it directly into your program. Use SCPI library (one of many) from NI Visa (officially supported) to many Python direct ones...
If you want to download data to your PC, map a folder on your PC to a scope (server and client is supported) and dump data directly as a file to a disk....

Do not make a mistake, it will be reported and if they can debug/improve/upgrade (choose your favorite word, so nobody feels excluded) they will. But this is not very big problem. It is not mainstream use of this component. So don't expect priority overnight fix.

If some SCPI expert here (there seem to be many) test deeper and finds real problems when SCPI is used with proper SCPI client on other side, please let us know.

[eTobey]

UI lags big time:

Its the second time today, that the UI started to become laggy to the point, you had to restart the scope. Might not be introduced with the new firmware, but just in case, i post my setup, and want to make others aware of it.

[Furna]

My last post on this specific topic.
Let me state again: blaming the browser without further investigation is just gambling; browsers are updated every few *days* by big (bigger than Siglent) organizations squashing bugs evry hour.
IMHO the bug is between the SCPI engine and the web server; or in the webserver itself. Both components are *inside* the scope and are not updated if nobody signal Siglent about a bug.
I understand the bug will be signaled to Siglent.
End of my post
;-)

[2N3055]

My last post on this specific topic.
Let me state again: blaming the browser without further investigation is just gambling; browsers are updated every few *days* by big (bigger than Siglent) organizations squashing bugs evry hour.
IMHO the bug is between the SCPI engine and the web server; or in the webserver itself. Both components are *inside* the scope and are not updated if nobody signal Siglent about a bug.
I understand the bug will be signaled to Siglent.
End of my post
;-)

Yes, the web component is LOADED from the scope. But COMMUNICATES from and LIVES (executes) in browser on your PC.
As it is, it is a remote connected client to SCPI server.

So it is NOT a SCPI server bug but that specific webpage component bug.
That is what I'm saying.

Nobody is blaming the BROWSER itself (as in Chrome or Edge...).

Since you seem to be very committed to very fine detail you should appreciate that I'm just trying to be very specific as to what the problem actually is.
In the end without good, correct and accurate specification of error there is no correct fix, but searching for a problem in a wrong place...

[eTobey]

So it is NOT a SCPI server bug but that specific webpage component bug.

I have analysed the traffic. The result says something different. The browser shows exactly what is coming in over LAN.

[2N3055]

So it is NOT a SCPI server bug but that specific webpage component bug.

I have analysed the traffic. The result says something different. The browser shows exactly what is coming in over LAN.

Are you always this obnoxious?
Who cares !!
It is a bug only with trying to get that data from web interface. Siglent will figure it out. I don't plan to analyze it in any detail. Not my job.

Your bombastic claim (as all your claims are, with titles like some stupid Youtube channel trying to get attention) is "SCPI BUG !!!"
It is not..

SCPI does not have problems when used via NI VISA as is officially supported by manufacturer (2000 individual packets ):

[eTobey]

You cant just acknowledge the facts, cant you? There is no need to get upset about facts...