When do you need a 4 channel scope???

[nctnico]

(1) Oddly enough Tektronix made the opposite error in the distant past.  Their 2440 series of DSOs only have envelope detection listed as a feature but support peak detection as well.

On Japanese DSOs peak detect is typically called envelope mode so when it comes to envelope mode and peak detect you really need to read the manual to check what the oscilloscope manufacturer has implemented.

(2) Automatic peak-to-peak triggering is another feature almost all DSOs lack but the last generation of analog oscilloscopes also lacked it.

I never heard of this mode until now. I had to look it up and it may be handy for repair work. It automatically triggers on the largest peak of the input signal.

[David Hess]

(1) Oddly enough Tektronix made the opposite error in the distant past.  Their 2440 series of DSOs only have envelope detection listed as a feature but support peak detection as well.

On Japanese DSOs peak detect is typically called envelope mode so when it comes to envelope mode and peak detect you really need to read the manual to check what the oscilloscope manufacturer has implemented.

It makes sense but I wish they would be consistent and some manufacturers definitely take advantage of the confusion between the two.  Marketting is hardly going to advertise that their product lacks a significant feature if obfuscation produces sales.

So what do they call envelope mode?

(2) Automatic peak-to-peak triggering is another feature almost all DSOs lack but the last generation of analog oscilloscopes also lacked it.

I never heard of this mode until now. I had to look it up and it may be handy for repair work. It automatically triggers on the largest peak of the input signal.

Automatic peak-to-peak triggering sets the trigger level range to the negative and positive peaks of the signal in real time so the trigger level becomes a percentage of the peak-to-peak signal level, is continuously updated, and the trigger sensitivity tracks the signal level making triggering on low level signals easier.  (1) This often replaced AC trigger coupling as a default so some oscilloscopes which had it dropped AC trigger coupling.  Switching between AC and DC coupling has no effect if automatic peak-to-peak triggering is used.

The problem with auto level is that it fixes the trigger level, sometimes at the wrong level, and then there is annoyingly significant delay if it has to reset the trigger level again which if you are rapidly probing signals is all the time.  Auto level effectively just presses the 50% Set Level control for you.

I was so used to automatic peak-to-peak triggering that I took it for granted until it vanished as a feature.  When I refurbished my Tektronix 7603, I missed automatic peak-to-peak triggering which is not supported with the standard 7B53A dual timebase so I bought a 7B92A dual timebase to replace it ... not realizing that the 7B92A and none of the 7000 series dual timebases support automatic peak-to-peak triggering.  The more recent 7000 series single timebases do support it however so often I use one of them in my 7603.

(1) I suspect Jim Williams liked this feature although I think his favorite but older oscilloscopes did not support it.  He called it out once as the basis of a triggering circuit that he designed which duplicated it.

[tautech]

A simple to use DSO Zoom is all you need, much more user friendly that the old dual delayed timebase on a CRO.

I actually find a dual delayed timebase to be easier to use because the horizontal controls scale with the timebase setting sort of like how when automatic peak-to-peak triggering is used, (2) the trigger level control scales with the waveform levels.  It is easier to design a poor user interface for pan and zoom.

Pan and zoom is very useful but not always a replacement for delayed acquisition with a second timebase.  The sample rate is limited by the main time/div and record length (3) no matter what the zoom is set to.  This came up here a couple months ago here with a Rigol DS1000Z where a new user wanted to delay 1 second to capture the following pulse from the 1 PPS output of a GPS receiver and found that the sample rate was then too low to capture the edge accurately even with the Rigol's long record length.

Older and higher end modern DSOs have delayed acquisition and triggering as well as pan and zoom which can get confusing when both are used at the same time.  More than once I have used automatic setup or restored the default state on the DSO I was using because that was faster than figuring out how to undo the combined pan, zoom, unlocked horizontal traces, and delayed sweep features.  Or in the case of recent DSOs, combining delayed acquisition with pan and zoom caused them to crash.

(1) Oddly enough Tektronix made the opposite error in the distant past.  Their 2440 series of DSOs only have envelope detection listed as a feature but support peak detection as well.

(2) Automatic peak-to-peak triggering is another feature almost all DSOs lack but the last generation of analog oscilloscopes also lacked it.  (4) Auto level is a poor substitute and is a major reason I often use an old analog oscilloscope or DSO which support automatic peak-to-peak triggering for rapid diagnostics.

(3) This is a good justification for large record lengths if dual timebase capability is not present and explains why dual timebase capability was common even in low end DSOs up until recently.  Single timebases with pan and zoom replaced dual timebases in low end DSOs when large acquisition memories became inexpensive enough.

(4) Automatic level triggering is cheaper to implement than automatic peak-to-peak triggering if the trigger level is set digitally which explains why the last generation of analog oscilloscopes and almost all DSOs with an analog trigger lack automatic peak-to-peak triggering.  Automatic peak-to-peak triggering should be free on DSOs which have digital triggering but inexplicably they lack it.

I've read, re-read and thought about this all day. 
Sure it's been a little while since I've had to use a CRO and I used the Dual timebase heaps BUT with the Siglent's I have I can't see a use case or any CRO feature you've listed unsurmountable with the UI in a SDS1kX.   

The Zoom gives me 2 independent timebases that with one push of the timebase control lets me toggle between each, the holdoff and Horizontal position (pan) give me any delay I might need and I can pan in either timebase if needs be.
The Trigger level is common to the main and Zoom windows and in the case with Siglent's can be set to 50% of the waveform amplitude with a push of the Trigger level control.
All trigger, pan and level controls can be returned to their default positions with a single push of each.

A screenshot, switching PSU charging a SLA battery.
Some Holdoff, Delay, Pan, Normal Trigger, Peak detect and AC input coupling.

[Howardlong]

Dual timebase isn't the same as a dual independent trigger.

The example I gave, that of a PWM fan controller with tacho feedback, is probably the most obvious example of where this might be convenient.

The Tek 2465, for example, has a VERT trigger mode where, when in ALT mode (as opposed to CHOP) and _not_ in AUTO LVL mode, "each displayed channel in turn provides the triggering signal".

[guenthert]

Naiively perhaps I would have thought I would get away with a large memory and properly selected sampling rate.  In the PWM fan example, assuming PWM frequency is 400kHz, we might want to have at least 10 samples per period, hence I'd chose a sampling rate of 4MHz (or thereabout).  If I have a memory depth of 12MS, a full buffer gives me 3s (well, 1.5s for two channels).  This clearly doesn't work for all conceivable use cases, but should be more than enough to clearly see the fan's tacho output (a few dozen to a few hundred Hz), no?

[nctnico]

No. Only with alternate triggering you can get the two signals stable on the screen.

[bson]

Naiively perhaps I would have thought I would get away with a large memory and properly selected sampling rate.  In the PWM fan example, assuming PWM frequency is 400kHz, we might want to have at least 10 samples per period, hence I'd chose a sampling rate of 4MHz (or thereabout).  If I have a memory depth of 12MS, a full buffer gives me 3s (well, 1.5s for two channels).  This clearly doesn't work for all conceivable use cases, but should be more than enough to clearly see the fan's tacho output (a few dozen to a few hundred Hz), no?

Use the scope to inspect the waveforms to make sure they're clean and at the proper levels, then feed them into a dual channel counter.  It will tell you all about the two signals in terms of frequency, duty cycle, etc.

[guenthert]

No. Only with alternate triggering you can get the two signals stable on the screen.

Well, single shot of course.  Not sure what 'stable on screen' would mean for two unrelated signals or why I would want it.

[David Hess]

Sure it's been a little while since I've had to use a CRO and I used the Dual timebase heaps BUT with the Siglent's I have I can't see a use case or any CRO feature you've listed unsurmountable with the UI in a SDS1kX.   

The Zoom gives me 2 independent timebases that with one push of the timebase control lets me toggle between each, the holdoff and Horizontal position (pan) give me any delay I might need and I can pan in either timebase if needs be.

Zoom only yields 2 independent timebases if the zoomed area is within the acquisition record of the first timebase.  If the delay is outside of the acquisition record like in the example I gave (1), then the sampling rate has to be lowered.  So the timebases are *not* independent and there is in fact only one timebase and one acquisition record with two different views.  Large record lengths make dual timebases unnecessary so long as the delay stays within the acquisition record.

So why can't the delay be longer than the acquisition record when digital triggering is used?  The problem is that with only one acquisition record, if the delay extends beyond the end, then the sampled data at the trigger point is lost and the trigger cannot be reconstructed.  The digitizer logic qualifies the trigger but the processor determines the actual trigger point later after reconstruction is done.

The Trigger level is common to the main and Zoom windows and in the case with Siglent's can be set to 50% of the waveform amplitude with a push of the Trigger level control.

And with automatic peak-to-peak triggering, I only have to set the trigger level once (2) and never remove my hands from what I am doing to adjust the trigger level.  So for instance when set to 50%, it *always* triggers at 50% of the waveform peak-to-peak amplitude.  It is great for rapid probing or if the signal is constantly changing.

Automatic level triggering works almost as well but is slow with a significant delay when the probe is moved or the signal changes.  (3) It just pushes the 50% trigger level control for you when triggering is lost.  Under some conditions automatic level triggering will select the wrong level.

I looked through the Siglent SDS1000x manual and found no equivalent to either; the trigger level either has to be adjusted manually or the 50% trigger button has to be used.  Does it matter?  Not really.  It is just a usability issue but I certainly prefer oscilloscopes that have these features over ones that do not because they are faster and easier to use.

(1) The Rigol would have needed a 1 gigasample record length to sustain its maximum sample rate over the 1 second delay using just pan and zoom.  If the DSO can trigger and then delay before acquisition which is standard with dual timebases and analog triggering but rare with digital triggering for the reason described above, then the record length can be tiny.

(2) And usually never because it is difficult to adjust the trigger level such that it will not work.  In practice it just gets left somewhere around 50%.

(3) With an analog trigger, the delay is because of how the signal levels are measured.  The trigger level is iteratively adjusted until the signal peaks are found.  This has the advantage of operating over the entire trigger bandwidth without requiring wide bandwidth peak detectors.  A DSO with digital triggering should not have this problem but it may still pick the wrong level and stick with it.

[Howardlong]

No. Only with alternate triggering you can get the two signals stable on the screen.

Well, single shot of course.  Not sure what 'stable on screen' would mean for two unrelated signals or why I would want it.

As I've stated more than once in this thread before, an easy example is a fan PWM controller with tacho feedback. While such situations aren't typical, they're not uncommon either.

[MrW0lf]

Use the scope to inspect the waveforms to make sure they're clean and at the proper levels, then feed them into a dual channel counter.  It will tell you all about the two signals in terms of frequency, duty cycle, etc.

What is typical real-world accuracy need you have come across? How many digits, at what frequencies? Application? I'm doing some experimentation on this currently.

[guenthert]

No. Only with alternate triggering you can get the two signals stable on the screen.

Well, single shot of course.  Not sure what 'stable on screen' would mean for two unrelated signals or why I would want it.

Poor choice of words on my side.  I rather meant that 'stable on screen' for two unrelated signals would be actively misleading.  There is after all only one timeline (ignoring relativistic effects here).  Take your data, analyze it afterwards, see the relationship between the signals (or the lack thereof).  I rather want my instruments to lie to me as little as possible.

I think the two different time bases were a (clever) crutch invented for analog oscilloscopes, which just couldn't store sufficient amount of data (not more than one screen).  I don't think it has a place in the realm of digital scopes.

[nctnico]

No. Only with alternate triggering you can get the two signals stable on the screen.

Well, single shot of course.  Not sure what 'stable on screen' would mean for two unrelated signals or why I would want it.

Poor choice of words on my side.  I rather meant that 'stable on screen' for two unrelated signals would be actively misleading.  There is after all only one timeline (ignoring relativistic effects here).  Take your data, analyze it afterwards, see the relationship between the signals (or the lack thereof).  I rather want my instruments to lie to me as little as possible.

I think the two different time bases were a (clever) crutch invented for analog oscilloscopes, which just couldn't store sufficient amount of data (not more than one screen).  I don't think it has a place in the realm of digital scopes.

It does. As Howard Long already wrote there are good uses to show two unrelated signals stable on screen with the scope in run mode. For example one signal which shows the sample points and the signal to be sampled in order to determine there are enough sample points per period.

[rf-loop]

This is one example about ALT trigger.  XDS3102A
Two separate asyncronous signals also with very different frequency.
Both channels have individual trigger. Left bottom see trigger counting freq and right side trigger level markers. 


Nice feature when example compare two separate unlocked signal generators risetime or signal shape.
Also example in cases where signals have same freequency but need watch sender shape and reflected signal shape and there is high jitter due to random travel time.  etc



But one "negative point", both channels have same time base.


This is why I still like Tek 7844  real dual beam scope.

[RissViss]

Thanks for all the reply's.
Just wondering about the posts talking about triggers, is this still on topic?
Do not get me wrong it is a very interesting discussion, and I do not mind it continuing, just want to know if its on topic?

Have a day

[irakandjii]

For me, the best reason for a 4 channel scope is: Decoding high speed SPI busses.

[David Hess]

This is why I still like Tek 7844  real dual beam scope.

I do not have the 7844 with its dual beam capability but I do have several other 4-bay 7000 mainframes which can do the same thing in a more limited way by either alternating or chopping the horizontal timebases.  I never realized how useful this capability was until I got the first one.

Jim Williams gave some examples where independent triggering was used in Linear Technology application note 65.  The switching power supply driven Royer converters he worked with operate asynchronously.

He makes a comment about figure 42, "Most oscilloscopes, whether analog or digital, will have trouble reproducing this display" which I have never understood since it seems like a 2 channel DSO could replicate it easily.  A single beam analog oscilloscope would almost certainly have a problem with it for the reasons he identifies.

[123rolle]

I'm looking for a first scope. First I thought getting Siglent sds1104x-e because of its possibilities for waveform and MSO with extra equipment. But would 2 channel scope + logic analyzer separate be better choice? In which case do I need 3-4 channels if I also get logic analyzer?

[nctnico]

If you develop circuits, then a 2 channel oscilloscope runs out of channels quickly. You'll often want to see more than 2 signals (looking at various points in a circuit) so a 4 (or more  ) channel scope is the best choice. Having digital channels is handy but you'll need to verify whether the signal levels / edges are correct first using an analog channel. I'd go for an MSO rather than an external logic analyser because you see the analog and digital signals in the same time domain using an MSO.

[Howardlong]

I'm looking for a first scope. First I thought getting Siglent sds1104x-e because of its possibilities for waveform and MSO with extra equipment. But would 2 channel scope + logic analyzer separate be better choice? In which case do I need 3-4 channels if I also get logic analyzer?

I have experience of the SDS1104X-E, and as a 4-ch scope it's fine, but the MSO SLA1016 option is TERRIBLE!!!! It makes the scope very slow and there are many limitations. For example, the is not even a zoom function with the digital channels enabled. Other limitations are described here: https://www.eevblog.com/forum/testgear/siglent-sds1204x-e-released-for-domestic-markets-in-china/msg2007983/#msg2007983

I much prefer using the venerable Rigol DS1000Z Plus series and its RPL1116 LA probes. The limitations with this are the decode being limited to what's on the screen, and that it's only 4+0ch, 3+8ch or 2+16ch analogue/digital split. Bear in mind that the screen's small so it's a tight fit getting all that on anyway. I can't remember finding the analogue/digital split being a limitation for practical use cases, more important is the flexibility. It's not unusual to forego unnecessary analogue channels for digital channels on a scope's screen real estate anyway,

Overall despite what some might say, I'm much more a fan of an MSO compared to using a separate scope & LA, but then most of what I do is mixed signal work. The only time I've used a 2ch scope w/o MSO in recent times is out in the field with a handheld unit: even then I'd probably consider myself better off with an Analog Discovery.

[pdenisowski]

Do any (non-USB/PC) oscilloscopes come with HDMI and direct output to a 1080p (or greater) monitor? Having such large screen real-estate and being able to see so many channels sure would be a nice selling point!

The R&S RTO6 has a really big (15.6") capacitive touch screen and supports both HDMI 2.0 and Display Port+++

[Martin72]

In which case do I need 3-4 channels

Simply when you want to see more than 2 channels...
Actual private example:
Next weekend I want to monitoring a protection circuit from an audio amplifier.
So I want to monitor both inputs, the output and the muting pin at the same time...

[2N3055]

Do any (non-USB/PC) oscilloscopes come with HDMI and direct output to a 1080p (or greater) monitor? Having such large screen real-estate and being able to see so many channels sure would be a nice selling point!

The R&S RTO6 has a really big (15.6") capacitive touch screen and supports both HDMI 2.0 and Display Port+++

Wow, R&S RTO6 can be had for less than 600 USD ?? I will buy one immediately !!! No, wait, make that 2 !!



Seriously, person is discussing sub 600 USD scopes.
Statement that R&S RTO6 is a good scope, in this context seems,  I don't know, just a bit too enthusiastic, maybe?
Even saying that RTB2000 is good scope is not relevant in the price range discussed.

[tggzzz]

In which case do I need 3-4 channels

Simply when you want to see more than 2 channels...

... and you can't satisfy your objectives without it. (See the aphorism in my .sig!)

You need a scope to look at analogue waveforms, in digital systems to ensure signal integrity. That only needs two channels, to see the time relationships.

You need a digitising scope to capture one-off events that cannot be made repetitive.

To look at signals that are known to be valid digital signals, you can use a logic analyser or protocol analyser or printf() statements. Those can be particularly effective if you use them to filter out the irrelevant crap so you can concentrate on the fault events; that can be difficult with scopes.

Depending on circumstances, a scope with more than two channels may or may not be a cost-effective solution.

[pdenisowski]

Do any (non-USB/PC) oscilloscopes come with HDMI and direct output to a 1080p (or greater) monitor? Having such large screen real-estate and being able to see so many channels sure would be a nice selling point!

The R&S RTO6 has a really big (15.6") capacitive touch screen and supports both HDMI 2.0 and Display Port+++

Seriously, person is discussing sub 600 USD scopes.
Statement that R&S RTO6 is a good scope, in this context seems,  I don't know, just a bit too enthusiastic, maybe?
Even saying that RTB2000 is good scope is not relevant in the price range discussed.

Sorry, I thought it was a general question ("Do any (non-USB/PC) oscilloscopes... "), not specific to sub 600 USD scopes. 

There are, of course, other ways of putting a scope display on a large monitor, e.g. remote access, which is supported even by many modern lower-end scopes, but if you want an actual HDMI output, you're obviously talking about a completely different class of oscilloscopes.