Oscilloscope Zoom Out Quirk

[nctnico]

Oh you mean split screen? Yeah they dont really do a good job of that either.. you just get the smaller horizonal box at the top you cant vertically resize to choose how much each pane is using

Yes, zoom mode = split screen.

[Elasia]

Oh you mean split screen? Yeah they dont really do a good job of that either.. you just get the smaller horizonal box at the top you cant vertically resize to choose how much each pane is using

Yes, zoom mode = split screen.

Eh.. the way it is now its hard to really call it split screen with independent timebases... their intent is more like the top is just a view finder from how it works as is but yeah thats something else that would be nice to have

[2N3055]

Please read all of my posts on this topic. I have explained very clearly why and how zoom mode is NOT the solution. And if you think carefully about it then you'll also see that by using zoom mode to force the memory length you sacrifice the zoom function (delayed / dual timebase) as well. Also, don't get stuck in old grudges and be blinded by them (same problem Tautech has). Currently Siglent is missing a memory management feature which is easy to add and it seems they are working on it right now.

I know you keep saying it but what we are talking about doesnt even modify siglents UI zoom mode as it is.. im lost why you keep saying zoom is sacrificed?

I'm now using zoom mode as is and it works fine to zoom out and see more data on screen, i mean thats kind of its job right?  What do you think zoom mode is suppose to be? I'm a bit lost on that as well

In some cases you might want to look at a signal at two positions / different time bases (this is called dual timebase on analog oscilloscopes). With auto memory length the memory depth will be set to the amount needed to fill the screen with the longest time/div setting. However if you want to force the oscilloscope to use more memory you'll need to set the time/div longer. But in that case you likely won't be able to see the details in a signal. So without being able to manually force the memory depth you can use the zoom mode to force deeper memory OR see two parts of a signal at the same time. IOW: With auto memory depth you can't have both deep memory and see two parts of the same signal at the same time.

Honestly, I don't understand what are you saying.



How is this not dual timebase?  How is this not detail from bigger picture?

[nctnico]

You choose an example in which you can still see both signals on an oscilloscope with very little memory. Set the bitrate to 1Mbit/s (100 times higher) with a much higher packet rate as well and try to do the same with the top zoom window set to 5ms/. You'll see a contigous colored band. If you set the zoom window to 50us/ in order to see the packets a Siglent scope will shorten the memory depth and there is nothing you can do about it. You have a choice between either having long memory but not being able to see the packets in the zoom window OR seeing the packets but not have deep memory.

[2N3055]

You choose an example in which you can still see both signals on an oscilloscope with very little memory. Set the bitrate to 1Mbit/s (100 times higher) with a much higher packet rate as well and try to do the same with the top zoom window set to 5ms/. You'll see a contigous colored band. If you set the zoom window to 50us/ in order to see the packets a Siglent scope will shorten the memory depth and there is nothing you can do about it. You have a choice between either having long memory but not being able to see the packets in the zoom window OR seeing the packets but not have deep memory.

No it won't. I showed you example from my Pico where I sampled 100 MPoints in 200ms and shown it zoomed in to 20ns /div, zoom factor of  1M X. Yes, one million times.
Tautech posted one image above, where he got 50 MPoints at 20 ms/div (200 ms total) and is showing details at 20 us/div.

This is how it works but you don't seem to understand. On main timebase you grab big chunk of data, and zoom shows a subset of that, in such a detail that is only defined by sampling rate. Which will be limited by memory size.

But zooming in won't change anything from how timebase and sampling is set, only what are you looking at.

So you enable 100 MPoints, set timebase to use all 100 MPoints, and you can zoom in to every single point...
And zoom will not shorten or change anything.

What are you saying is SIMPLY NOT TRUTH. Just look at the overwhelming evidence shown here..

[rf-loop]

........... IOW: With auto memory depth you can't have both deep memory and see two parts of the same signal at the same time.

In this previous extremely simple "single toggle button" solution this is not possible of course if need two separate zoomed in parts displayed. But also in your explanation this full deep memory is available only after stop. Edit: clarified: Runtime you can see only these two zoomed in parts with more or less limited positioning. (I use also "zoomed in" thinking for this full window zoom mode what is just how many (most) scopes work.  You may have 10M  sample memory and 1Gsa/s and your fastest 2ns/div "timebase" zoomed in (full screen zoom in) 10 div wide window is just example 20ns slice from this example 10000000ns long acquistion length. And what happen when turn "timebase" only zoom in factor change.
Now if your time base is 20ns you can select zoom 20ns and move it to other position in this long memory.
But how if you then have moved it but you suddenly want zoom it out. many scopes do not accept it. You need first change this "main timebase" aka zoom factor aka horizontal scale.

Full memory lenght to display + dual independent zoom in to details like in some ancient good analog scope is this. 
Digital scope can do it and also in runtime whole memory length display is possible to see or shut off and  these two separate "zoomed in" traces can draw overlaid with free time position related to each others and or each others based fixed difference where all are user adjustable including also freedom for set fixed trigger time position in acquisition memory. Three timebase with dual independent delayed timebases (timebase is bit fun word with digital scopes but Im from analog scopes golden era).  Of course all can do, even Rolls Royce. But there is markets and how much some features really add sales and more importantly profit to maker.... all is nice to have but money talks.
But I want scope what can display full current memory length always when/if I want on the screen and many scopes can not.  Even if length is 1G I want see whole trace without hiding parts out from display except cases when I do not care this. If I want cdetails today it is best there is user settable windows available with different scales overlaid on screen and full freedom for user to position and adjust these scales and time relations to each others and trigger-.

[rf-loop]

You choose an example in which you can still see both signals on an oscilloscope with very little memory. Set the bitrate to 1Mbit/s (100 times higher) with a much higher packet rate as well and try to do the same with the top zoom window set to 5ms/. You'll see a contigous colored band. If you set the zoom window to 50us/ in order to see the packets a Siglent scope will shorten the memory depth and there is nothing you can do about it. You have a choice between either having long memory but not being able to see the packets in the zoom window OR seeing the packets but not have deep memory.

What is real reason for these lies?

This is handled many many times including document images etc. Only what Sig do not have is hardware decode with pros and cons. Yes it is slower to decode and also some limits with decode result length but also it can decode without real time live signal just afterwards..

[nctnico]

You choose an example in which you can still see both signals on an oscilloscope with very little memory. Set the bitrate to 1Mbit/s (100 times higher) with a much higher packet rate as well and try to do the same with the top zoom window set to 5ms/. You'll see a contigous colored band. If you set the zoom window to 50us/ in order to see the packets a Siglent scope will shorten the memory depth and there is nothing you can do about it. You have a choice between either having long memory but not being able to see the packets in the zoom window OR seeing the packets but not have deep memory.

No it won't. I showed you example from my Pico where I sampled 100 MPoints in 200ms and shown it zoomed in to 20ns /div, zoom factor of  1M X. Yes, one million times.
Tautech posted one image above, where he got 50 MPoints at 20 ms/div (200 ms total) and is showing details at 20 us/div.

This is how it works but you don't seem to understand. On main timebase you grab big chunk of data, and zoom shows a subset of that, in such a detail that is only defined by sampling rate. Which will be limited by memory size.

But zooming in won't change anything from how timebase and sampling is set, only what are you looking at.

So you enable 100 MPoints, set timebase to use all 100 MPoints, and you can zoom in to every single point...
And zoom will not shorten or change anything.

What are you saying is SIMPLY NOT TRUTH. Just look at the overwhelming evidence shown here..

Sorry, but again you are not understanding the problem; you should try the steps I wrote on a Siglent or Lecroy oscilloscope and you'll see the limitations for yourself. You think you achieve the same but you don't. It is true there are many routes to Rome. You keep insisting the taking the long route to Rome is that same as taking the short route to Rome. That simply isn't true. It is no my fault you can't see the short route to Rome.

[jemangedeslolos]

You choose an example in which you can still see both signals on an oscilloscope with very little memory. Set the bitrate to 1Mbit/s (100 times higher) with a much higher packet rate as well and try to do the same with the top zoom window set to 5ms/. You'll see a contigous colored band. If you set the zoom window to 50us/ in order to see the packets a Siglent scope will shorten the memory depth and there is nothing you can do about it. You have a choice between either having long memory but not being able to see the packets in the zoom window OR seeing the packets but not have deep memory.

No it won't. I showed you example from my Pico where I sampled 100 MPoints in 200ms and shown it zoomed in to 20ns /div, zoom factor of  1M X. Yes, one million times.
Tautech posted one image above, where he got 50 MPoints at 20 ms/div (200 ms total) and is showing details at 20 us/div.

This is how it works but you don't seem to understand. On main timebase you grab big chunk of data, and zoom shows a subset of that, in such a detail that is only defined by sampling rate. Which will be limited by memory size.

But zooming in won't change anything from how timebase and sampling is set, only what are you looking at.

So you enable 100 MPoints, set timebase to use all 100 MPoints, and you can zoom in to every single point...
And zoom will not shorten or change anything.

What are you saying is SIMPLY NOT TRUTH. Just look at the overwhelming evidence shown here..

Sorry, but again you are not understanding the problem; you should try the steps I wrote on a Siglent or Lecroy oscilloscope and you'll see the limitations for yourself. You think you achieve the same but you don't. It is true there are many routes to Rome. You keep insisting the taking the long route to Rome is that same as taking the short route to Rome. That simply isn't true. It is no my fault you can't see the short route to Rome.

I think there are people full of good will who try to understand with the support of screenshots.

So I believe the best thing to make yourself understood is that you make a capture with your manual memory depth capable R&S or GW Instek and you do the same with your Lecroy in a scenario where clearly demonstrates the difference between the two memory management implementation.
So that everyone understands without ambiguity.

I know it takes time, but it could reduce the 100-year war and save millions of lives.

[tv84]

Dave's video already shows the "problem" explicitly (at least for my knowledge level).

The rest of the rumblings are different ways of explaining the thing (with more or less expert details) but, in the end, after all the juice has been squeezed, all come down to the same.

Those that disagree with this analysis, have misunderstood the simplicity of the bad/wrong (strike the one that makes you itchy) implementation that Dave's has showed in the video.

[jemangedeslolos]

Dave's video already shows the "problem" explicitly (at least for my knowledge level).

The rest of the rumblings are different ways of explaining the thing (with more or less expert details) but, in the end, after all the juice has been squeezed, all come down to the same.

Those that disagree with this analysis, have misunderstood the simplicity of the bad/wrong (strike the one that makes you itchy) implementation that Dave's has showed in the video.

Yes, I think everyone has understood the problem since Dave posted the video.
But this is a problem that has been discussed here for years, often on the initiative of nctnico when we talk about the operation or the choice of an oscilloscope but rarely with a demonstration as explicit as Dave's one.

Despite this, there still seems to be misunderstandings between some here and on the real impact on real cases.
It would be great if nctnico could show us a case that we can solve with a GW Instek or an R&S and that we cannot with a Siglent or a Lecroy or with or with more steps or less efficiency.

[maginnovision]

It's not about can or can not, it's about the time it takes. You can decode serial data even without the scope doing it. It just takes longer.

[2N3055]

You choose an example in which you can still see both signals on an oscilloscope with very little memory. Set the bitrate to 1Mbit/s (100 times higher) with a much higher packet rate as well and try to do the same with the top zoom window set to 5ms/. You'll see a contigous colored band. If you set the zoom window to 50us/ in order to see the packets a Siglent scope will shorten the memory depth and there is nothing you can do about it. You have a choice between either having long memory but not being able to see the packets in the zoom window OR seeing the packets but not have deep memory.

No it won't. I showed you example from my Pico where I sampled 100 MPoints in 200ms and shown it zoomed in to 20ns /div, zoom factor of  1M X. Yes, one million times.
Tautech posted one image above, where he got 50 MPoints at 20 ms/div (200 ms total) and is showing details at 20 us/div.

This is how it works but you don't seem to understand. On main timebase you grab big chunk of data, and zoom shows a subset of that, in such a detail that is only defined by sampling rate. Which will be limited by memory size.

But zooming in won't change anything from how timebase and sampling is set, only what are you looking at.

So you enable 100 MPoints, set timebase to use all 100 MPoints, and you can zoom in to every single point...
And zoom will not shorten or change anything.

What are you saying is SIMPLY NOT TRUTH. Just look at the overwhelming evidence shown here..

Sorry, but again you are not understanding the problem; you should try the steps I wrote on a Siglent or Lecroy oscilloscope and you'll see the limitations for yourself. You think you achieve the same but you don't. It is true there are many routes to Rome. You keep insisting the taking the long route to Rome is that same as taking the short route to Rome. That simply isn't true. It is no my fault you can't see the short route to Rome.

I DID show you that I accomplish the same with Picoscope.. With 1 million factor zoom.. Simultaneously capturing 100 Mpoints of data with 500 MS/sec and at the same time looking at it at trigger point (or any other point relative to trigger) at 20 ns /div.... Here it is 10 ms/div timebase , looking at it at 1,2 us/div, while simultaneously decoding 100 Mpoints worth of data sampled at 1GS/sec.



If I wanted to speculate I would simply think that you don't know how to use zoom properly, so you didn't set timebase to long enough before pressing zoom.
Or that you did try that on some previous Siglent scope that DID behave like this, and had some weird behaviour. 
That would explain why are you so adamant at repeating what is proven (with images) that is not true.

Entering zoom mode doesn't change any of parameter of sampling that is set previously with max mem depth and timebase. It simply opens new view to buffer with separate time axis on which it plots smaller subset of data on different time axis from timebase.

[tv84]

It would be great if nctnico could show us a case that we can solve with a GW Instek or an R&S and that we cannot with a Siglent or a Lecroy or with or with more steps or less efficiency.

That is another matter and goes deeply into each other's way of using a scope. Based on all the information in this forum it's perfectly clear that it will never be a consensus in the way people use a so generic instrument as a scope is. And, specially, people with lots of experience and accustomed to use it in a certain way.

Of course, it's not only about the people's way but also the type of signal analysis and/or glitches that are being searched that influence the best way to handle the scope.

If Siglent corrects this "quirk" i'm sure it will not be only nctnico that will be happy  . That will benefit even the current Siglent advocates! And some/many of them have yet to understand that benefit.

[2N3055]

Dave's video already shows the "problem" explicitly (at least for my knowledge level).

The rest of the rumblings are different ways of explaining the thing (with more or less expert details) but, in the end, after all the juice has been squeezed, all come down to the same.

Those that disagree with this analysis, have misunderstood the simplicity of the bad/wrong (strike the one that makes you itchy) implementation that Dave's has showed in the video.

What Dave have shown is good explanation how Siglent scope works. I agree that they should add manual memory management if they can, just because. It can't hurt.
I'm not disagreeing with Dave's findings.
I'm fine with the fact that Nico uses something and it works for him.

My problem with Nico's interpretation is that he keeps saying things that are simply not true... I show example with 1 million zoom factor that disproves his statements, he says that if I chose different timebase it would be different. It wouldn't. I know how my scope works, I'm the one that use it on daily basis. He says Siglent scopes can't show zoomed data without losing detail few posts down from image from Elasia that clearly shows  it can do it.. I really don't care much for that.

I know I can be very stubborn sometimes. Not very often, but I simply cannot stand that incorrect is being sold as correct by repetition, false argumentation and misdirection.
It ain't truth. Simple as that.  I can stop, but it will never be truth. Never.

[Someone]

Still rolling on...  different scopes word differently! OMG! None of them is perfect and there isn't one true way to manage memory. Let the users pick what they want, rather than insisting others should blindly believe your ways are best.

[nctnico]

Dave's video already shows the "problem" explicitly (at least for my knowledge level).

The rest of the rumblings are different ways of explaining the thing (with more or less expert details) but, in the end, after all the juice has been squeezed, all come down to the same.

Those that disagree with this analysis, have misunderstood the simplicity of the bad/wrong (strike the one that makes you itchy) implementation that Dave's has showed in the video.

What Dave have shown is good explanation how Siglent scope works. I agree that they should add manual memory management if they can, just because. It can't hurt.
I'm not disagreeing with Dave's findings.
I'm fine with the fact that Nico uses something and it works for him.

My problem with Nico's interpretation is that he keeps saying things that are simply not true... I show example with 1 million zoom factor that disproves his statements, he says that if I chose different timebase it would be different. It wouldn't. I know how my scope works, I'm the one that use it on daily basis. He says Siglent scopes can't show zoomed data without losing detail few posts down from image from Elasia that clearly shows  it can do it.. I really don't care much for that.

I know I can be very stubborn sometimes. Not very often, but I simply cannot stand that incorrect is being sold as correct by repetition, false argumentation and misdirection.

And it doesn't occur to you that you are simply misunderstanding what I'm trying to point out? I know I'm not the best one to explain things so it is not entirely your fault. I don't know how I can get you to see the subtle nuances you are missing -again my shortcoming-. You using other oscilloscopes besides one from Siglent doesn't help either because the behaviour this thread is about is very specific for how auto memory length works. However saying people tell lies because you don't understand their point is not the smartest move to make.

[Elasia]

Still rolling on...  different scopes word differently! OMG! None of them is perfect and there isn't one true way to manage memory. Let the users pick what they want, rather than insisting others should blindly believe your ways are best.

 

[Someone]

Dave's video already shows the "problem" explicitly (at least for my knowledge level).

The rest of the rumblings are different ways of explaining the thing (with more or less expert details) but, in the end, after all the juice has been squeezed, all come down to the same.

Those that disagree with this analysis, have misunderstood the simplicity of the bad/wrong (strike the one that makes you itchy) implementation that Dave's has showed in the video.

What Dave have shown is good explanation how Siglent scope works. I agree that they should add manual memory management if they can, just because. It can't hurt.
I'm not disagreeing with Dave's findings.
I'm fine with the fact that Nico uses something and it works for him.

My problem with Nico's interpretation is that he keeps saying things that are simply not true... I show example with 1 million zoom factor that disproves his statements, he says that if I chose different timebase it would be different. It wouldn't. I know how my scope works, I'm the one that use it on daily basis. He says Siglent scopes can't show zoomed data without losing detail few posts down from image from Elasia that clearly shows  it can do it.. I really don't care much for that.

I know I can be very stubborn sometimes. Not very often, but I simply cannot stand that incorrect is being sold as correct by repetition, false argumentation and misdirection.

And it doesn't occur to you that you are simply misunderstanding what I'm trying to point out? I know I'm not the best one to explain things so it is not entirely your fault. I don't know how I can get you to see the subtle nuances you are missing -again my shortcoming-. You using other oscilloscopes besides one from Siglent doesn't help either because the behaviour this thread is about is very specific for how auto memory length works. However saying people tell lies because you don't understand their point is not the smartest move to make.

You've continually made "can't be done" statements without putting all the conditions/assumptions required to back that up.

Between you and Dave its just another fanboy rant at this stage.

[nctnico]

I stated the exact conditions several times in several way so you can try it yourself. But be aware that this is expert equipment use. Like pro-athletes who need shoes that have a tiny, insignificant looking tweak that makes all the difference between losing and winning. Anyway, there is not much need for further discussion. It appears Siglent is working on adding new features to the memory management so you can force full memory depth. At least they see and understand the value and that is all that matters. (Hopefully) we end up with more oscilloscopes to choose from so mission accomplised.

As tv84 put it so elegantly: That will benefit even the current Siglent advocates! And some/many of them have yet to understand that benefit.

Or let's pull in a quote from Albert Einstein. He wasn't so subtle:
Great spirits have always encountered violent opposition from mediocre minds. 

[Elasia]

Great spirits have always encountered violent opposition from mediocre minds. 

Story of my life... i swear i spend more time just stirring up hornet nests lol

but noooo we cant do xyz! ... bullshit         

[jemangedeslolos]

It would be great if nctnico could show us a case that we can solve with a GW Instek or an R&S and that we cannot with a Siglent or a Lecroy or with or with more steps or less efficiency.

That is another matter and goes deeply into each other's way of using a scope. Based on all the information in this forum it's perfectly clear that it will never be a consensus in the way people use a so generic instrument as a scope is. And, specially, people with lots of experience and accustomed to use it in a certain way.

Of course, it's not only about the people's way but also the type of signal analysis and/or glitches that are being searched that influence the best way to handle the scope.

If Siglent corrects this "quirk" i'm sure it will not be only nctnico that will be happy  . That will benefit even the current Siglent advocates! And some/many of them have yet to understand that benefit.

Yes it's true that it's a bit off topic, but Dave's video shows that a problem discussed here in tens of threads for several years can be summed up in a single video.

I understood ( I hope... ) the test nctnico wants to make to show that the automatic management on the Siglent doesn't allows (for example) to look at a high bitrate packet in both the top zoomed window and the bottom window AND then look at another position WITHOUT having to recapture.
This is an example he found, I think, to make it clear that the use of the zoom function on the Siglent doesn't completely resolve the lack of manual memory management.

Now we are approaching 20 pages with some misunderstandings that still persist.

All this blabla to say that until we haven't some sort of standart test plan / benchmark to see how different oscilloscopes work with exactly the same settings and signals,  I still think that a screenshot or a small video or a gif can be much more explicit than dozens replies more or less well written with more or less diplomacy 

[tautech]

show that the automatic management on the Siglent doesn't allows (for example) to look at a high bitrate packet in both the top zoomed window and the bottom window AND then look at another position WITHOUT having to recapture.

Yet they do that without fault !
All the primary timebase record can be panned through from within the zoomed window.
And Navigate to any time position in the capture too.

An example of the required setup posted in another thread:

[Gandalf_Sr]

Yet [Siglent scopes] do that without fault !

Huh! I thought the simple conclusion that Dave came to was that Siglent scopes don't give users the option to manually make some of the vast amount of memory available to enable the ability to zoom out - furthermore Tautech actually told us that Siglent were working on the problem - or did I miss something?

Here we go again, the Siglent dealer chiming in to try to convert us non-believers that Siglent already does what we need, we just have to drink more Koolaid and watch while Tautech walks across water without sinking deeper than his ankles.

FFS, can't we just give it a break?  As jemangedeslolos says, Dave has summed everything up in a simple video.

 

[jemangedeslolos]

Yet they do that without fault !
All the primary timebase record can be panned through from within the zoomed window.
And Navigate to any time position in the capture too.

An example of the required setup posted in another thread:

I know my english is bad but if you try to decrypt, it is not quite exactly what ntcnico suggests.
You can only see beautiful colors on the top window but you cannot compare this signal with the signal zoomed in the bottom window.
The "dual time base" thing on analog scope that ntcnico is trying to explain.
If you choose very close timebase, you will see the problem.