EEVblog #685 - What Is Oscilloscope AC Trigger Coupling?

[EEVblog]

In this tutorial Dave describes what AC trigger coupling is on an oscilloscope and why it can be useful. Not only on old analog CRT oscilloscopes, but modern digital scopes as well.
How and why is it different to AC channel input coupling?
Also, use of the 50% trigger control, and how the venerable DS1052E is still more usable than the new DS1054Z.

Previous video on Rigol DS1054Z and DS2000 Jitter Problem

[eV1Te]

Great video, as always 

I already knew about the AC triggering but I found the explanation and also the comparison of the Agilent to the Rigol very interesting.

Maybe you could do a oscilloscope shootout like you did with millimeters a while back, comparing the responsiveness of the buttons for each brand and some of the functions, like the 50 % trigger level? 

The only thing I am not satisfied with on my DS1000z is the responsiveness of the interface, easily worth a few $100 extra just for that if I had known before hand.

[nixfu]

Ok now I understand.    I guess I just too it for granted on my analogs because it is just the way things were by default. 

At least it's mearly a nuscience that it does not work correctly at the moment(you have to play with the trigger level and DC triggering), but at least I know what I am missing as far as functionality.   


I am also more glad I kept an analog scope around as well.   I still prefer analog for doing stuff like signal tracing and unknown circuit, and watching signals with its "real-time" unprocessed waveform display.

There is a place and time for both analog and digital scopes, especially if your doing random repair, diags, etc. 

[EEVblog]

The only thing I am not satisfied with on my DS1000z is the responsiveness of the interface, easily worth a few $100 extra just for that if I had known before hand.

I think responsiveness is ok for the price.
I show this in my review video supporters have seen, even with everything turned on, it's still not bad.

[rs20]

Dave, you briefly pointed out that the digital scopes perform triggering on the digital data stream, rather than the pre-input-coupling signal at the input. However, I disagree with your claim that the scope should behave identically nevertheless. It's more nuanced than that.

In particular, the digital stream already has AC input coupling applied (if selected), so if you select DC (or what could be more accurately termed "whatever input coupling is" or "what you see on the screen") trigger coupling and AC input coupling, the Rigol oscilloscope would trigger just fine on those four waveforms you presented to the scopes. So far, this is exactly right and good because:

• Seeing the trigger level visually on the screen is a very good thing, and only meaningful for DC trigger coupling (as you rightly point out). I can't stress this enough, a DSO that defaulted to AC trigger coupling and therefore wouldn't display the trigger level on-screen would be an awful step backwards. I'm all for consistency, but this is why digital scopes really should behave differently to analog scopes.
• It makes little sense to use true DC triggering with AC input coupling, so it seems fine that AC input coupling "auto-enables" "AC trigger coupling" in this sense. Implementing true DC triggering with an AC-coupled digital waveform would require some hacky measurement of the DC offset to compensate the digital triggering circuitry FPGA gates, or expensive analog trigger circuitry that would be prohibitively expensive at this price point.

The problem is, you often want to see the DC component, yet trigger off the AC-coupled signal. You did mention this, but not really in the context of how this is the only time you want AC trigger coupling in the Rigol design.

I expect the other big DSO players (Agilent, Tek) have made the same design decisions for the same reasons? Nevertheless, I think it's reasonable that Rigol's AC input triggering implicitly overrules the selected trigger coupling. I'm writing this to preempt the "but I use AC input coupling with DC-heavy signals all the time and always had success with the default 'DC' trigger coupling, this just doesn't mesh with what you've drawn on the whiteboard..." confusion that people might be thinking.

I agree that Rigol seemed to express some doubt that AC triggering was necessary, and it's important to recognise that it's less necessary with scopes that draw the trigger signal off after input coupling. It's more of a niche thing than with analog scopes. Only with that context in place can you fully emphasise that in real life (in particular, mixed voltage digital logic and analog signals living on different biases), it's still absolutely necessary. Fortunately, we've got news from Rigol that they're fixing the bugs, so everything's just perfect now! Apart from the lack of a 50% auto trigger level feature!

EDIT: Btw, sorry for double-posting across forums, Dave, but I do think this is worthy of wider discussion, being the pedantic nutter that I am!

[EEVblog]

In particular, the digital stream already has AC input coupling applied (if selected), so if you select DC (or what could be more accurately termed "whatever input coupling is" or "what you see on the screen") trigger coupling and AC input coupling, the Rigol oscilloscope would trigger just fine on those four waveforms you presented to the scopes.

Of course those 4 signals would have triggered fine with AC input coupling, you missed the entire point of the argument. It's for the situation when you want DC coupling on the input, not AC input coupling.

The problem is, you often want to see the DC component, yet trigger off the AC-coupled signal. You did mention this, but not really in the context of how this is the only time you want AC trigger coupling in the Rigol design.

I did more than just mention it, it was the entire example point in my video!

[Bud]

This is what the 50% control thing looks like on a Tek 2467B, was occasionally wondering what it does, now know, thank you.

[rs20]

In particular, the digital stream already has AC input coupling applied (if selected), so if you select DC (or what could be more accurately termed "whatever input coupling is" or "what you see on the screen") trigger coupling and AC input coupling, the Rigol oscilloscope would trigger just fine on those four waveforms you presented to the scopes.

Of course those 4 signals would have triggered fine with AC input coupling, you missed the entire point of the argument. It's for the situation when you want DC coupling on the input, not AC input coupling.

The problem is, you often want to see the DC component, yet trigger off the AC-coupled signal. You did mention this, but not really in the context of how this is the only time you want AC trigger coupling in the Rigol design.

I did more than just mention it, it was the entire example point in my video!

I'm sorry, you're right about this, my message was far to long and rambly. I should have focussed on my main point that expecting analog and digital scopes to behave identically is not right. May I just ask what your opinion is on this point:

Seeing the trigger level visually on the screen is a very good thing, and only meaningful for DC trigger coupling (as you rightly point out). I can't stress this enough, a DSO that defaulted to AC trigger coupling and therefore wouldn't display the trigger level on-screen would be an awful step backwards. I'm all for consistency, but this is why digital scopes really should behave differently to analog scopes.

I think it's important to highlight that "DC trigger coupling" is much more versatile in the Rigol scope than in the analog scopes (in that choosing AC input coupling always automagically forces faux-AC-trigger coupling), hence why it's the default. There's a really tricky compromise with these digital scopes: either choose DC trigger coupling and get to see you trigger level drawn on-screen, or choose AC trigger coupling and get to not worry about manipulating your trigger level continuously. With analog scopes, you never get to see the trigger level, so it's a non-issue.

[EEVblog]

With analog scopes, you never get to see the trigger level, so it's a non-issue.

Of course you do, it's the start level of the trace on the screen!

[rs20]

Of course you do, it's the start level of the trace on the screen!

I... I... but... "not for digital waveforms with such fast rise times that the scope doesn't show the rising edge"... no, that's just getting way off-track and pedantic... sigh... *walks off with tail between legs*...

I feel so sure I had a point worth sharing, but evidently I don't. You win this round! 

[AG6QR]

Interesting.  I had never noticed the AC coupling trigger menu item before the "jitter bug" (Sounds like a 1930s dance craze) video.  On that video, I became aware of it, and the use of it is now obvious.

But playing with it a bit on my DS1052E, I believe I've found a subtle bug in this area of that venerable scope.

I set up a sine wave at 300mV p-p, with a DC offset of 250mV.  So the entire waveform is always at a positive voltage, never crossing 0V.  Using DC coupling on the display, and AC trigger coupling, it couples beautifully at a trigger level of 0mV, and I can play with the trigger level and watch the waveform go back and forth a bit just as you showed in your video.

But the bug is this:  If I press the DS1052E's dedicated "50%" button that you showed, it sets the trigger level to 250 mV!  That would be the correct 50% level if I were using DC trigger coupling -- it's the midpoint between the top and bottom of the waveform on the DC coupled display.  But I'm using AC trigger coupling, so a value of 250 mV is outside the range of the waveform, and it fails to trigger.

When the trigger is AC coupled, I think the "50%" button ought to select the midpoint of the AC coupled waveform, which would usually be pretty close to 0 (not necessarily exactly zero -- consider a square wave with a very high or very low duty cycle).  It's useless to select the midpoint of the DC coupled waveform in this scenario, but that's what the scope does.

This isn't a serious bug, because it's easy enough to work around the issue by setting the trigger manually or pushing the trigger level knob to quickly set it to 0.  I don't hold out much hope that Rigol would fix such a minor problem for a product so near its end of life.

FWIW, I have firmware 00.04.01.

[TheNewLab]

OK, May I get back to the basics?

I get it, setting the AC Trigger coupling allows me to quickly find a signal on the screen. I get it, that using the AC Probe coupling eliminates the DC part of a signal so I may focus on the actual waveform.
I think I understand that to switch from AC Probe coupling to DC Probe coupling allows me to find any DC components of a signal riding on the overall signal.
However, where or when would I use the DC trigger coupling? Why not just keep the trigger coupling on AC all the time? Also with DSO type scopes, does this even matter if I use the AUTO function to find the signal in the first place?

 -Please excuse my ignorance in electronics- I started just to repair a stereo receiver, and got hooked on the colorful objects and all those cool trashcans! Oscar (the Grouch) would have a field day running around on circuit boards.

[rs20]

OK, May I get back to the basics?

I get it, setting the AC Trigger coupling allows me to quickly find a signal on the screen. I get it, that using the AC Probe coupling eliminates the DC part of a signal so I may focus on the actual waveform.
I think I understand that to switch from AC Probe coupling to DC Probe coupling allows me to find any DC components of a signal riding on the overall signal.
However, where or when would I use the DC trigger coupling? Why not just keep the trigger coupling on AC all the time? Also with DSO type scopes, does this even matter if I use the AUTO function to find the signal in the first place?

EDIT: What Dave says below; single-shot is a leading usecase for true DC trigger coupling.

AUTO is terrible because it messes with your horizontal timebase and vertical scaling, takes a second or two, and requires you to dedicate a hand to pressing the auto button. As Dave points out, this is completely awful if you're probing around different points on a circuit rapidly. In the same situation, AC trigger coupling works wonderfully (and/or AC input coupling with digital scopes).

On a digital scope, you might also want to use DC trigger coupling because it shows the trigger level visually on the screen as a horizontal line as you adjust it (this display doesn't appear with AC trigger coupling because it makes no sense). Seems silly, and in a sense it really is, but the average digital scope user demands that they see the trigger level visually on the screen as it's adjusted, so that's why DC trigger coupling is the default on digital scopes. Combine that with the fact that digital scopes effectively enable AC trigger coupling when AC input coupling is selected (even if the trigger coupling is left on "DC"), you find that "DC" trigger coupling is OK-ish most of the time, hence why most people didn't even know AC trigger coupling existed. But of course, Dave demonstrates in this video why you might want to switch explicitly to AC trigger coupling, and maybe even leave it permanently on AC trigger coupling (I'm strongly considering this, I think I can handle an invisible trigger level) even on a digital scope.

[EEVblog]

However, where or when would I use the DC trigger coupling? Why not just keep the trigger coupling on AC all the time?

You need DC trigger coupling in order to trigger off the exact point you want if you have a DC or very low frequency input signal.
e.g. AC coupling would be the wrong thing to use if you had a slow changing single shot waveform, like say the start-up ramp of a power supply you were trying to capture.
DC trigger coupling is an absolutely essential feature.

[tautech]

Interesting.  I had never noticed the AC coupling trigger menu item before the "jitter bug" (Sounds like a 1930s dance craze) video.  On that video, I became aware of it, and the use of it is now obvious.

But playing with it a bit on my DS1052E, I believe I've found a subtle bug in this area of that venerable scope.

I set up a sine wave at 300mV p-p, with a DC offset of 250mV.  So the entire waveform is always at a positive voltage, never crossing 0V.  Using DC coupling on the display, and AC trigger coupling, it couples beautifully at a trigger level of 0mV, and I can play with the trigger level and watch the waveform go back and forth a bit just as you showed in your video.

But the bug is this:  If I press the DS1052E's dedicated "50%" button that you showed, it sets the trigger level to 250 mV!  That would be the correct 50% level if I were using DC trigger coupling -- it's the midpoint between the top and bottom of the waveform on the DC coupled display.  But I'm using AC trigger coupling, so a value of 250 mV is outside the range of the waveform, and it fails to trigger.

When the trigger is AC coupled, I think the "50%" button ought to select the midpoint of the AC coupled waveform, which would usually be pretty close to 0 (not necessarily exactly zero -- consider a square wave with a very high or very low duty cycle).  It's useless to select the midpoint of the DC coupled waveform in this scenario, but that's what the scope does.

This isn't a serious bug, because it's easy enough to work around the issue by setting the trigger manually or pushing the trigger level knob to quickly set it to 0.  I don't hold out much hope that Rigol would fix such a minor problem for a product so near its end of life.

FWIW, I have firmware 00.04.01.

Yes you would expect the 50% button to set the trigger to ~mid-point of a waveform. AC or DC.
Fair enough that a Channel Position button would set the trace to mid-screen (0) but what good is that for a trigger level setting.
Yes call it a Bug.

[peufeu]

I... I... but... "not for digital waveforms with such fast rise times that the scope doesn't show the rising edge"... no, that's just getting way off-track and pedantic... sigh... *walks off with tail between legs*...

Actually it does too, because a good analog scope has a big reel of cable inside as a delay line, so it can show you what happened just before the trigger point... 

[megajocke]

Input coupling selection is usually not done in the position shown in Dave's diagram. It is usually the very first thing the signal passes through, even before it goes to the attenuator and input amplifier, and this goes for both digital and analog scopes. The trigger pickoff point is therefore almost always after the input coupling selection.

For that reason, when using AC input coupling, the trigger is also looking at the AC coupled signal even when in DC trigger coupling mode. If AC trigger coupling mode is selected, the signal will be high pass filtered twice before reaching the triggering circuitry/code.

[Mark Hennessy]

Input coupling selection is usually not done in the position shown in Dave's diagram. It is usually the very first thing the signal passes through, even before it goes to the attenuator and input amplifier, and this goes for both digital and analog scopes. The trigger pickoff point is therefore almost always after the input coupling selection.

For that reason, when using AC input coupling, the trigger is also looking at the AC coupled signal even when in DC trigger coupling mode. If AC trigger coupling mode is selected, the signal will be high pass filtered twice before reaching the triggering circuitry/code.

I made the same point in a post I composed last night, but I decided not to bother posting in the end.

Yes, the trigger pick-off point is after the input stage - it's usually just before the beam switching (and another take-off point is usually provided just after the beam switch for ALT triggering - which lets you view 2 traces at once that are not related to each other in the time domain - occasionally very useful).

I figured that Dave drew the diagram that way for the sake of getting the message across - I'm very aware from my job that sometimes in education, it's necessary to tell little white lies - although TBH it probably wouldn't have made much difference to have drawn it correctly.

One thing I am sure about: lots of engineers, at every level of education and experience, are often confused by AC and DC matters. The ability to think about each separately is an important skill - a very basic application of the superposition theorem - and time spent thinking about it is time well spent. I see that as a positive outcome from the discovery of this issue

Another thing I see evidence of almost daily: many people - again, at every level - are confused by oscilloscope triggering. Ever since peak-to-peak trigger setting became the norm (early '80s?), you hardly need to think about triggering when viewing simple waveforms. When teaching, I make a big deal about being able to see the start of the sweep - I line it up on the left-hand edge of the graticule (it was against the LH edge of the CRT on the video - possibly slightly off to the left at times) and make people adjust the trigger level (and other settings) while watching it. Surprisingly few seem to realise that it's just a comparator feeding a D-type flip-flop!

I had a baptism of fire though. My first 'scope had a faulty trigger IC that was an unobtainium custom type. Nothing that a new PCB with 5 or 6 ICs couldn't fix though! I learnt some hard lessons doing that

That 'scope has analogue storage, and is great for capturing PSU start-up behaviour and similar things. That's when single-shot with DC trig coupling is essential (as was said earlier).

I show this in my review video supporters have seen

Dave, will this review be released generally one day? I'm awaiting your detailed opinion of this one with interest.

[(In)Sanity]

It's funny,  my Tek 2465BDM has DC trigger coupling on by default.  It also has a 50% auto trigger button.   Not bad for an analog scope.   I have a DS2000A series as well. 

Rigol should allow the user to select what they want the "Zero" button to do.
Rigol should address the external trigger being DC only on the DS2000 series and either fix it or remove the menu options for AC.
Rigol should ......

Overall they are good scopes "for the money".    Rigol needs to do everything in their power to get people like me to stop using that term.   

Just my two cents,

Jeff

[open loop]

Thanks for the video this has explained/cleared up things for me, I have now a much better understanding on the subject! 

I was that person you mimicked "Well I have never used AC coupling ... don't care about the bug..."  will play around with it when I get a chance.

Still would go and buy the Rigol 1000Z as I cant afford the Agilent MSO-x 3000 scope, what are my needs? Intensity graded display, serial decoding and high re-fresh rate.

[ornea]

I too was beginning to wonder why you would ever use DC coupled triggering after watching your discussion on AC coupled triggering.  Thanks for clearing this up. 

[Salas]

FORCE: in Normal and single trigger modes, press this key to generate a trigger
signal forcefully.

Says the DS2000 manual.  Alright, can you mention practical example cases?

[rs20]

FORCE: in Normal and single trigger modes, press this key to generate a trigger
signal forcefully.

Says the DS2000 manual.  Alright, can you mention practical example cases?

Here's one: in "Normal" mode, the oscilloscope will just go dead if the trigger level is not intersecting the waveform. If you want to just get a rough snapshot of where the signal is sitting, what it looks like roughly, you can just whack the "Force" button and you can see what's going on.

Another regards single-shot mode. If you're tracking some interesting event at a very slow time-base (100ms/div or whatever), then you might want to trigger when you see some particular interesting thing happen. It might be something optical, not electrical, or maybe it's just hard to set the trigger up just right and you just want to eyeball it. It's the most sophisticated triggering system in the world: your eyes & brain! But the "jitter" of the human triggering system is pretty terrible .

[Salas]

In single mode I can make an off-trigger waveform appear by pressing force. In run mode it just keeps running horizontally.

[NiHaoMike]

When the triggering is done digitally on some new scopes, what I would like to see is a smart trigger mode based on autocorrelation.