Oscilloscope Zoom Out Quirk

[EEVblog]

I've tried to imagine a reason it could be better than using a zoom window, I can't see it. Having to manipulate the view is more button presses compared to having all that visible on the screen at the same time. I don't deny its useful in some circumstances

Glad you agree.

[jemangedeslolos]

Ultimately we all know what will start the third world war.

[nctnico]

While it is true many lost cost DSO's with Zoom mode only provide a 50/50 split of the display do indeed limit usability while upper entry level DSOs provide a much better split ratio where buyers of which are more accustomed to not needing so much visual reference of the primary timebase when in fact they want a larger zoomed timebase which is what they primarily work with.

You don't need to use the zoom function (split window view) at all if you don't want to.

Of course not however it gives the fullest pre and post trigger capture which in the case of a deep memory DSO could be 100+ Mpts before and after the trigger.

Maybe that is the case on Siglent scopes, but on InfiniVision scopes it is nothing more than a different view of the same data, after capture. You can always set the trigger delay to whatever you need, before or after the trigger event. The actual trigger event doesn't even have to be included in the capture at all, if that is what you want.

On a DSO data is acquired around the point-of-origin (usually center) position on the screen. The trigger point is just a time offset from the point-of-origin.

[2N3055]

While it is true many lost cost DSO's with Zoom mode only provide a 50/50 split of the display do indeed limit usability while upper entry level DSOs provide a much better split ratio where buyers of which are more accustomed to not needing so much visual reference of the primary timebase when in fact they want a larger zoomed timebase which is what they primarily work with.

You don't need to use the zoom function (split window view) at all if you don't want to.

Of course not however it gives the fullest pre and post trigger capture which in the case of a deep memory DSO could be 100+ Mpts before and after the trigger.

Maybe that is the case on Siglent scopes, but on InfiniVision scopes it is nothing more than a different view of the same data, after capture. You can always set the trigger delay to whatever you need, before or after the trigger event. The actual trigger event doesn't even have to be included in the capture at all, if that is what you want.

On a DSO data is acquired around the point-of-origin (usually center) position on the screen. The trigger point is just a time offset from the point-of-origin.

Absolutely incorrect in every way. Trigger point is THE reference point. Acquisition delay is specified from trigger point.  All else is display presentation. Acquisition engine will decide how much pre-trigger and post-trigger data will get in part based on relation of trigger point position in relation to display reference position, and it's position on screen..

[Wuerstchenhund]

I don't care about ntcnico's claims, he can defend them himself. But if you cannot see how having capture data outside of the display area could be potentially useful, then I don't know what else to say, apart from "surely the potential advantage is obvious to any experienced engineer." There is no need to give specific examples, use your imagination.

I don't think anyone argues that data outside a window which shows a narrow view of a small part of a signal is not important, and I haven't seen any comment suggesting that.

In fact, the idea that surrounding data is important is the main idea behind deep memory scopes and the standard method of "capture long and zoom in". Because knowing that things outside the original point of interest may turn out very relevant, the reason why we have large memory in mopdern scopes is so we can capture the whole sequence right from the start, and then zoom in to the areas of interest (which there might be many).

So the question is not if data outside the (narrow) window is not important, but why is the scope set in a way that important data is outside the window in the first place.

The idea that one *needs* this "zoom out" capability one can make-do, under limited conditions, on some scopes, simply because the scope was set to show a very narrow part of the actual signal when one spontaneously decides to that the rest of the signal should be examined, too, suggests to me a lack of preparation. Because especially when you're bug hunting it shouldn't really come to a surprise that parts outside that initial narrow piece of interest will become relevant. Capturing long means having every part of the signal inside view and thereby quickly available for examination in a coherent manner.

And even if you got yourself in a situation where you found your original capture was too short, instead of make-do with "zoom out" it might make a lot more sense to finally set that timebase to capture long so you avoid finding yourself in the same predicament again a few minutes later.

So I fully agree with you, data outside that narrow screen is useful, more often than not. Which raises the question why is important data outside the screen area in the first place. This off-screen data being important is even more reason to make sure right from the start to capture everything in a long acquisition.

[Wuerstchenhund]

Imagine a camera with a zoom lens.  An experienced photographer does not start with the lens zoomed out to the max, and then only zooms in from that point.  Instead, they start with the lens zoomed somewhere around where their experience tell them is the best zoom factor for the kind of picture they are taking, then they trim the zoom both in and out a little bit from that point while watching the model in the viewfinder, until he/she looks "just right".  It is a fast and fluid way of working with a camera.

Very bad analogy. Because optical zoom on a camera changes the resolution per scene area, so capturing pictures fully zoomed out will result in a loss of information when creating closer-ups from that 'acquisition'.

On deep memory scope however, capturing long and zooming in does not result in a loss of information, because neither the vertical resolution nor the sample rate changes over capturing short.

I can tell you that, if the loss of information (detail) wasn't a problem with zoom then many photographers would gladly capture fully zoomed out, because by capturing excess scenery they can later easily adjust the actual image size and position without having to go back to the place of scene and re-capture.

[Wuerstchenhund]

Gentlemen, you are ALL talking the same language!

Nico starts with zoom in and likes to be able to zoom out. The others start with zoom out and like to be able to zoom in.

It's all the same, but in reverse order.

All the rest are just details (and personal preferences or lifelong habits)!

Unfortunately, it's not the same.

"Zooming in" as in the standard method is done (you guessed it!) through the scope's zoom function. It provides references as for the position on the captured signal, can be displayed together with the full signal and works for single and continuous acquisitions.

As to "zooming out", most scopes don't allow zooming outside the scope area.

However, some scopes allow post-acquisition change of the timebase.

Nico's method uses this as a means to "zoom out". However, it's not really a zoom, it's a change of timebase post-acq. And because it's a change of timebase post-acq, it *only* works if the scope is halted (i.e. no new triggers) and you can't display it together with the "zoomed in" view (well, you could probably use waveform memory, but you'd lprobably ose the timing information).

So while the end result of both methods are the same in the cases where Nico's method works, they are *not* achieved the same way.

[SilverSolder]

[...] However, instead of using 'spare' memory for off screen capture, I rather like the Siglent technique of keeping a history of previous acquisitions you can go back and look at [...]

Is that what "save to disk" or "save to usb" etc does?

[Wuerstchenhund]

To go back to the transmission analogy, I learned to drive on a manual and hated automatics as I always knew (thought) I could do better, even when I had an Audi A6, IMHO I did a better job than the transmission.

That's more down to Audi's habit of selling automatic vehicles in the U.S. (and some other markets) with a 'comfort oriented' (i.e. slow) transmission mapping.

I remember many years ago when I was still in Germany, someone did some test between automatic (using torque converter) and manual cars. That was probably 20 years ago, and even back then the auto beat the average (German) driver by a long shot. The bottom line was that, unless you're a professional racing driver, the auto is doing the better job 

My first cars were manual but I never really liked it, and I find it particularly annoying in stop-and-go traffic. I'd probably take a manual over a car with one of these horrible CVTs, though.
So I'll stick with (torque converter) automatic until Ford comes up with the F-150 Electric , thank you very much!

[2N3055]

[...] However, instead of using 'spare' memory for off screen capture, I rather like the Siglent technique of keeping a history of previous acquisitions you can go back and look at [...]

Is that what "save to disk" or "save to usb" etc does?

No, scope memory is used in seamless segmented mode, meaning scope keeps hundreds of previous "screens" (trigger events) at all times, transparently. With no performance impact to normal use.

[SilverSolder]

Imagine a camera with a zoom lens.  An experienced photographer does not start with the lens zoomed out to the max, and then only zooms in from that point.  Instead, they start with the lens zoomed somewhere around where their experience tell them is the best zoom factor for the kind of picture they are taking, then they trim the zoom both in and out a little bit from that point while watching the model in the viewfinder, until he/she looks "just right".  It is a fast and fluid way of working with a camera.

Very bad analogy. Because optical zoom on a camera changes the resolution per scene area, so capturing pictures fully zoomed out will result in a loss of information when creating closer-ups from that 'acquisition'.

On deep memory scope however, capturing long and zooming in does not result in a loss of information, because neither the vertical resolution nor the sample rate changes over capturing short.

I can tell you that, if the loss of information (detail) wasn't a problem with zoom then many photographers would gladly capture fully zoomed out, because by capturing excess scenery they can later easily adjust the actual image size and position without having to go back to the place of scene and re-capture.

If you have enough RAM in the scope, of course I agree.   But if you capture a "zoomed out" picture of a waveform on your scope of say, 10 seconds, just to pick an extreme,  and your scope does not have enough RAM to capture that at full resolution, there is indeed a loss of quality compared to capturing a smaller time slice at a higher rate.   That's my situation...

[SilverSolder]

[...] However, instead of using 'spare' memory for off screen capture, I rather like the Siglent technique of keeping a history of previous acquisitions you can go back and look at [...]

Is that what "save to disk" or "save to usb" etc does?

No, scope memory is used in seamless segmented mode, meaning scope keeps hundreds of previous "screens" (trigger events) at all times, transparently. With no performance impact to normal use.

So you can scroll back and view previous events? -  That does sound kind of cool. 

[2N3055]

Imagine a camera with a zoom lens.  An experienced photographer does not start with the lens zoomed out to the max, and then only zooms in from that point.  Instead, they start with the lens zoomed somewhere around where their experience tell them is the best zoom factor for the kind of picture they are taking, then they trim the zoom both in and out a little bit from that point while watching the model in the viewfinder, until he/she looks "just right".  It is a fast and fluid way of working with a camera.

Very bad analogy. Because optical zoom on a camera changes the resolution per scene area, so capturing pictures fully zoomed out will result in a loss of information when creating closer-ups from that 'acquisition'.

On deep memory scope however, capturing long and zooming in does not result in a loss of information, because neither the vertical resolution nor the sample rate changes over capturing short.

I can tell you that, if the loss of information (detail) wasn't a problem with zoom then many photographers would gladly capture fully zoomed out, because by capturing excess scenery they can later easily adjust the actual image size and position without having to go back to the place of scene and re-capture.

If you have enough RAM in the scope, of course I agree.   But if you capture a "zoomed out" picture of a waveform on your scope of say, 10 seconds, just to pick an extreme,  and your scope does not have enough RAM to capture that at full resolution, there is indeed a loss of quality compared to capturing a smaller time slice at a higher rate.   That's my situation...

If you have short memory scope there won't be any data "outside the screen" either..

[SilverSolder]

Gentlemen, you are ALL talking the same language!

Nico starts with zoom in and likes to be able to zoom out. The others start with zoom out and like to be able to zoom in.

It's all the same, but in reverse order.

All the rest are just details (and personal preferences or lifelong habits)!

Unfortunately, it's not the same.

"Zooming in" as in the standard method is done (you guessed it!) through the scope's zoom function. It provides references as for the position on the captured signal, can be displayed together with the full signal and works for single and continuous acquisitions.

As to "zooming out", most scopes don't allow zooming outside the scope area.

However, some scopes allow post-acquisition change of the timebase.

Nico's method uses this as a means to "zoom out". However, it's not really a zoom, it's a change of timebase post-acq. And because it's a change of timebase post-acq, it *only* works if the scope is halted (i.e. no new triggers) and you can't display it together with the "zoomed in" view (well, you could probably use waveform memory, but you'd lprobably ose the timing information).

So while the end result of both methods are the same in the cases where Nico's method works, they are *not* achieved the same way.

There is only so much real estate on the screen.  If you already have 2 analog channels plus a handful of digital ones, using the zoom window becomes a bit of a pain (uses too much space) compared to just altering the timebase "post facto".   Those multi-channel captures are also the kinds of scenarios where you most often end up looking for things that happened "off screen", in my experience at least.

[2N3055]

[...] However, instead of using 'spare' memory for off screen capture, I rather like the Siglent technique of keeping a history of previous acquisitions you can go back and look at [...]

Is that what "save to disk" or "save to usb" etc does?

No, scope memory is used in seamless segmented mode, meaning scope keeps hundreds of previous "screens" (trigger events) at all times, transparently. With no performance impact to normal use.

So you can scroll back and view previous events? -  That does sound kind of cool.

Not only scroll. Search, measure, make histograms, create display persistence (overlap them all), decode protocols...etc...

[SilverSolder]

Imagine a camera with a zoom lens.  An experienced photographer does not start with the lens zoomed out to the max, and then only zooms in from that point.  Instead, they start with the lens zoomed somewhere around where their experience tell them is the best zoom factor for the kind of picture they are taking, then they trim the zoom both in and out a little bit from that point while watching the model in the viewfinder, until he/she looks "just right".  It is a fast and fluid way of working with a camera.

Very bad analogy. Because optical zoom on a camera changes the resolution per scene area, so capturing pictures fully zoomed out will result in a loss of information when creating closer-ups from that 'acquisition'.

On deep memory scope however, capturing long and zooming in does not result in a loss of information, because neither the vertical resolution nor the sample rate changes over capturing short.

I can tell you that, if the loss of information (detail) wasn't a problem with zoom then many photographers would gladly capture fully zoomed out, because by capturing excess scenery they can later easily adjust the actual image size and position without having to go back to the place of scene and re-capture.

If you have enough RAM in the scope, of course I agree.   But if you capture a "zoomed out" picture of a waveform on your scope of say, 10 seconds, just to pick an extreme,  and your scope does not have enough RAM to capture that at full resolution, there is indeed a loss of quality compared to capturing a smaller time slice at a higher rate.   That's my situation...

If you have short memory scope there won't be any data "outside the screen" either..

There is off screen data as long as the timebase is running fast enough to use less RAM than what is required to display the image on the screen.

So basically, the more RAM the scope has, the slower the timebase setting that the Megazoom off-screen idea will be work with...   bottom line, more RAM benefits all scopes?

[2N3055]

Gentlemen, you are ALL talking the same language!

Nico starts with zoom in and likes to be able to zoom out. The others start with zoom out and like to be able to zoom in.

It's all the same, but in reverse order.

All the rest are just details (and personal preferences or lifelong habits)!

Unfortunately, it's not the same.

"Zooming in" as in the standard method is done (you guessed it!) through the scope's zoom function. It provides references as for the position on the captured signal, can be displayed together with the full signal and works for single and continuous acquisitions.

As to "zooming out", most scopes don't allow zooming outside the scope area.

However, some scopes allow post-acquisition change of the timebase.

Nico's method uses this as a means to "zoom out". However, it's not really a zoom, it's a change of timebase post-acq. And because it's a change of timebase post-acq, it *only* works if the scope is halted (i.e. no new triggers) and you can't display it together with the "zoomed in" view (well, you could probably use waveform memory, but you'd lprobably ose the timing information).

So while the end result of both methods are the same in the cases where Nico's method works, they are *not* achieved the same way.

There is only so much real estate on the screen.  If you already have 2 analog channels plus a handful of digital ones, using the zoom window becomes a bit of a pain (uses too much space) compared to just altering the timebase "post facto".   Those multi-channel captures are also the kinds of scenarios where you most often end up looking for things that happened "off screen", in my experience at least.

Well, on R&S RTM3000 you can actually resize it . On Pico, you create views and size them at will. On Lecroy you can move and resize....
My problem with zoom mode on many scopes is also that it takes too much space. But that could be solved easily if we complain to manufactures long and loud enough. 

[nfmax]

Once again, may I point out that the 'zoom window' display (the equivalent of a dual timebase on an analogue scope) is not needed to zoom & pan either in or out post acquisition: just use the ordinary horizontal scale & position controls!

[2N3055]

There is off screen data as long as the timebase is running fast enough to use less RAM than what is required to display the image on the screen.

So basically, the more RAM the scope has, the slower the timebase setting that the Megazoom off-screen idea will be work with...   bottom line, more RAM benefits all scopes?

I answered to your scenario: after you slow down enough you don't have enough memory for either one long timebase capture, or one short timebase capture with afterscreen data.

And yes, for these kinds of problems more  memory is more. But if you want to use it all the time, it makes scope slow.

[2N3055]

Once again, may I point out that the 'zoom window' display (the equivalent of a dual timebase on an analogue scope) is not needed to zoom & pan either in or out post acquisition: just use the ordinary horizontal scale & position controls!

To make sure, for 3000T and Rigol that is correct. I don't know about the others.
But I LIKE zoom mode. It gives me overview where in the buffer I am. Also on 3000T, I just use touch screen to move around... I just wish that zoom window could change size (to be little smaller if I want) when it get crowded.
But, that is mostly because screen on 3000T is not very big.

[rsjsouza]

Once again, may I point out that the 'zoom window' display (the equivalent of a dual timebase on an analogue scope) is not needed to zoom & pan either in or out post acquisition: just use the ordinary horizontal scale & position controls!

To make sure, for 3000T and Rigol that is correct. I don't know about the others.
But I LIKE zoom mode. It gives me overview where in the buffer I am. Also on 3000T, I just use touch screen to move around... I just wish that zoom window could change size (to be little smaller if I want) when it get crowded.
But, that is mostly because screen on 3000T is not very big.

Interesting; the DS4014 has a handy top "entire buffer" view that does the same thing and uses almost zero screen realstate. (Unless you mean something else, of course)

[Wuerstchenhund]

Imagine a camera with a zoom lens.  An experienced photographer does not start with the lens zoomed out to the max, and then only zooms in from that point.  Instead, they start with the lens zoomed somewhere around where their experience tell them is the best zoom factor for the kind of picture they are taking, then they trim the zoom both in and out a little bit from that point while watching the model in the viewfinder, until he/she looks "just right".  It is a fast and fluid way of working with a camera.

Very bad analogy. Because optical zoom on a camera changes the resolution per scene area, so capturing pictures fully zoomed out will result in a loss of information when creating closer-ups from that 'acquisition'.

On deep memory scope however, capturing long and zooming in does not result in a loss of information, because neither the vertical resolution nor the sample rate changes over capturing short.

I can tell you that, if the loss of information (detail) wasn't a problem with zoom then many photographers would gladly capture fully zoomed out, because by capturing excess scenery they can later easily adjust the actual image size and position without having to go back to the place of scene and re-capture.

If you have enough RAM in the scope, of course I agree.

I guess you mean sample memory, not RAM

But if you capture a "zoomed out" picture of a waveform on your scope of say, 10 seconds, just to pick an extreme,  and your scope does not have enough RAM to capture that at full resolution, there is indeed a loss of quality compared to capturing a smaller time slice at a higher rate.   That's my situation...

That is true of course, however if your scope has insufficient memory to capture a full sequence of whatever you're dealing with then you either have to capture a shorter sequence or, if the signal properties permit, capture long but at reduced sample rate.

In any ase, if the memory is insufficient to "capture long" then it is also insufficient for Nico's method.

But quite often, signals with long periods are also low BW signals, so you might well be able to capture the whole sequence at a reduced sample rate that is still sufficiently high.

[jemangedeslolos]

As for MCUs or like many others things, the best scope is the one that fit your needs of the one you know best.

I like zoom mode because I can see my data in context but I can see the same data with different methods, Im not a Nazi.
It depends on what I'm trying to measure, my progress in debugging or my mood of the day.

And I can do all of that because my scope is the best 

[Someone]

Once again, may I point out that the 'zoom window' display (the equivalent of a dual timebase on an analogue scope) is not needed to zoom & pan either in or out post acquisition: just use the ordinary horizontal scale & position controls!

But not all scopes will or can capture longer memory around the visible window.... the entire point of this discussion. So for those scopes zoom is a way to achieve the same thing. (loses some screen realestate, gains independent control of trigger/window/acquisition positioning)

[2N3055]

Once again, may I point out that the 'zoom window' display (the equivalent of a dual timebase on an analogue scope) is not needed to zoom & pan either in or out post acquisition: just use the ordinary horizontal scale & position controls!

To make sure, for 3000T and Rigol that is correct. I don't know about the others.
But I LIKE zoom mode. It gives me overview where in the buffer I am. Also on 3000T, I just use touch screen to move around... I just wish that zoom window could change size (to be little smaller if I want) when it get crowded.
But, that is mostly because screen on 3000T is not very big.

Interesting; the DS4014 has a handy top "entire buffer" view that does the same thing and uses almost zero screen realstate. (Unless you mean something else, of course)

Pretty much something like that. Maybe a bit bigger, just to get a glimpse of shapes....