12
\$\begingroup\$

I just can't understand why this is completely necessary. Take a sine wave for example. It's repeating. Can't the oscilloscope really "choose" anywhere to start the data acquisition. Since it repeats you'll get the same thing, right?

\$\endgroup\$
9
  • 4
    \$\begingroup\$ Most oscilloscopes can trigger directly off of the input, which in that case will give you what you want. External trigger is for when you want to trigger off of signal A while recording signal B. \$\endgroup\$ Commented Mar 11, 2021 at 20:33
  • 2
    \$\begingroup\$ Well, you choose at what voltage level the scope triggers. This just gives you another option. \$\endgroup\$ Commented Mar 11, 2021 at 21:16
  • \$\begingroup\$ Imagine that you want to see what the beginning of an input signal looks like. Even assuming that the input signal is nominally expected to be a sine wave, that doesn't prove that the beginning of the signal is what you expect it to be. Setting the trigger allows you to have the oscilloscope capture the very beginning of the signal, not some random point in the future. \$\endgroup\$ Commented Mar 12, 2021 at 4:35
  • 13
    \$\begingroup\$ Can you clarify whether you don't see why ANY type of trigger is needed or you don't see why the external trigger is needed? \$\endgroup\$
    – user57037
    Commented Mar 12, 2021 at 7:57
  • 1
    \$\begingroup\$ I often use an oscilloscope to look at a signal that's repeating, and almost impossible to catch without a trigger. OK it's a different kind of repeating. The centreburst of an FTIR interferogram is over in a few ms, and repeats after about 1s. If I want to optimise that, I want it to show up every time, and at the same place, so I can watch the screen (or us the peak-peak measurement the scope offers). That's a niche use but not far from typical \$\endgroup\$
    – Chris H
    Commented Mar 12, 2021 at 11:34

7 Answers 7

35
\$\begingroup\$

You can demonstrate the problem very easily on most scopes by setting the scope trigger to "free run" and setting the trigger level above the peak of the test waveform. (You can usually use the scope's calibration point squarewave for this test.) The scope will run a trace, wait a while for a trigger signal, give up and run another trace starting at a random point. The result will be a blurred mess with the waveform barely visible.

enter image description here

Figure 1. Trace of a sinewave without triggering. Image source: Test and Measurement Tips.

Since it repeats you'll get the same thing, right?

Yes. You get the same thing over and over at different starting points. It becomes very difficult to make any measurements.

enter image description here

Figure 2. The same trace with correct trigger setting.


From the comments:

Yes I've certainly seen the first screenshot as a result of incorrectly triggering, but I don't see why that happens. It's showing different voltages at different times, when in reality there is only a single voltage at a given time.

There's only a single voltage at a given time but the scope re-uses each x-axis point over and over again because the display isn't infinitely wide.

On a digital 'scope the blurring happens due to persistence of vision. Your eyes and brain can't extract just one scan or screen update. On the other hand, if you take a photo of the screen you will get just one trace if the exposure time is short enough.

On an analog cathode ray tube 'scope you have the additional factor of the tube's phosphor coating decay time. This had the effect of making the most recent traces appear brighter as the older ones faded away.

\$\endgroup\$
8
  • 1
    \$\begingroup\$ Yes I've certainly seen the first screenshot as a result of incorrectly triggering, but I don't see why that happens. It's showing different voltages at different times, when in reality there is only a single voltage at a given time. \$\endgroup\$
    – user256639
    Commented Mar 11, 2021 at 23:36
  • 1
    \$\begingroup\$ @kefffin Without triggering its either that (overlapping) or show a non-overlapping signal in real time. I dont know about you but I struggle to see signals in real time above a few Hz, and even thats usually not ideal. \$\endgroup\$
    – Matt
    Commented Mar 12, 2021 at 3:48
  • 2
    \$\begingroup\$ @kefffin The idea of the oscilloscope is to show not voltage at a given time, but to draw voltage at different times in one point of the screen. If the voltages match, you'll get clear picture. To make voltage match, you need to sample sine wave at sin(0), sin(2\$\pi\$), sin(4\$\pi\$) and so on. An oscilloscope has no idea about the math and anyway there are other signals apart from sine waves. So it uses triggers to catch certain phase of a repeating signal. \$\endgroup\$
    – AlexVB
    Commented Mar 12, 2021 at 8:08
  • 4
    \$\begingroup\$ @kefffin Imagine sine waves flying past your window. You take a picture with your phone. You take another picture. And a third one just for giggles. You go to your computer, transfer the pictures from your phone and use the graphics program you trust the most to make a single picture out of these three: You take each picture, make it a bit transparent and put them on top of each other. Will the sine waves overlap? No, probably not, as the sine waves were moving. There will be an offset. Now imagine taking photos was triggered by a specific part of that sine wave flying past your window. \$\endgroup\$
    – Num Lock
    Commented Mar 12, 2021 at 8:27
  • 3
    \$\begingroup\$ @BenVoigt, if you set the scope to single acquisition without properly setting the trigger you won't see anything on most 'scopes as far as I'm aware. You'll get a blank display. \$\endgroup\$
    – Transistor
    Commented Mar 12, 2021 at 15:57
15
\$\begingroup\$

I think it would help you understand if we could slow down the display and let you actually see what's happening in slow motion.

So, here we go. Here's the very first part of the signal:

Oscilloscope with part of sine wave.


Okay, so far so good. Let's fast-forward a bit until the signal reaches the edge of the screen:

Oscilloscope with sine wave going all the way across.


Now, here's where things get interesting. Since there is no trigger, the second "frame" of the signal starts right where the first one left off*. As it's drawn, it erases the first frame. So, here's the start of the second frame (in a different color to help differentiate it from the first frame):

Oscilloscope with part of two sine waves with different phases and different colors.


And, again, let's fast-forward until the second frame is completed:

Oscilloscope with second sine wave going all the way across.


But, the thing to remember here is that this is all happens in a fraction of a second. Even though the first frame has been erased, the whole thing happens so fast that you can still see the first frame thanks to your persistence of vision. So, what you actually see is something like this:

Oscilloscope with both sine waves superimposed.


Oh, wait. The 'scope isn't actually going to change the color of the signal for each frame. So, this would be a little more accurate:

Oscilloscope with both sine waves superimposed in the same color.


Starting to look familiar?


*Well, okay, most oscilloscopes will have a slight delay in resetting the drawing position, but it's still not going to make the second frame line up with the first unless you're really, extremely lucky.

\$\endgroup\$
4
  • \$\begingroup\$ Looks like I didn't pay any attention to the resolution of the images, and made them just a tiny bit bigger than I intended. If anyone is having issues with them breaking the page layout or not loading or whatever, let me know, and I can go back and scale them down. \$\endgroup\$ Commented Mar 12, 2021 at 18:30
  • \$\begingroup\$ meta.stackexchange.com/questions/25051/… you don't need to manually scale them. \$\endgroup\$
    – Jason C
    Commented Mar 13, 2021 at 6:52
  • 1
    \$\begingroup\$ On an analog scope you could adjust the timebase (horizontal speed) manually to make the waves line up. On a digital scope, you usually only have discrete steps available. \$\endgroup\$ Commented Mar 13, 2021 at 11:05
  • 1
    \$\begingroup\$ TL:DR: the time between two horizontal scans won't be an exact multiple of the signal period. You've shown it very nicely with pictures, but that's how to put it into words. (Even if you get it close, it won't be perfect and the waveform shown on screen will drift left or right) \$\endgroup\$ Commented Mar 13, 2021 at 16:16
7
\$\begingroup\$

If you trigger properly on a repeating waveform the oscilloscope will overlay subsequent traces in a phase coherent manner, so you will see a nice waveform that is not constantly changing. If you don't use the trigger at all and just let the oscilloscope free-run, then subsequent waveforms will have an arbitrary or random phase shift compared to the previous one. So a bunch of "sine" waves with different phase shifts will be drawn one over the other. Keep in mind that when you are down in the 1ms per division and less, the oscilloscope may be rescanning very rapidly. So the visual effect is that subsequent scope traces are drawn over previous ones.

I am not trying to be critical. But triggering an oscilloscope is one of the key functions of an oscilloscope. I have used various trigger functionalities at various times for various reasons, and the triggering function is really indispensable. Only a very inexperienced person would ask why it is necessary.

\$\endgroup\$
5
\$\begingroup\$

I just can't understand why this is completely necessary. Take a sine wave for example. It's repeating. Can't the oscilloscope really "choose" anywhere to start the data acquisition. Since it repeats you'll get the same thing, right?

Back in the day analog scopes didn't have memory. One sweep displayed a waveform, then it would disappear. So if you wanted a stable waveform to display on screen, it had to be swept and displayed many times per second. And every sweep had to line up and be synchronized horizontally so they would all appear right on top of each other and look like a solid, stable image.

To do that a trigger is absolutely necessary since the same point of each waveform must end up at the same point of the screen every time.

With storage scopes, like all modern digital scopes, you can disable the trigger ("free run" mode) and push the "single sweep" button, so it does one sweep without trigger and displays it... and it doesn't disappear because the scope has RAM. But on analog scopes, if you wanted to do that, you'd need to point a film camera at the screen, synchronize the camera shutter with the "single sweep" button, develop the film, etc.

Can't the oscilloscope really "choose" anywhere to start the data acquisition. Since it repeats you'll get the same thing, right?

Without trigger, the waveform on screen will start whenever the scope starts its acquisition cycle, so that's a random point in the period. If you only need to view one screen's worth of a periodic signal, then that's not a problem.

Another use case for not using a trigger is when you're using a rather slow sweep speed like 100ms/div to observe some slow varying, non-periodic signal and you just look at it on screen.

It's a pretty rare occurence though, usually you want to observe stuff that requires to display many periods on top of each other, and that needs a trigger. For example, see if the frequency or amplitude is stable, see if there are glitches or noise on top of the signal, observe what happens on channel1 when something happens on channel2, etc.

\$\endgroup\$
6
  • \$\begingroup\$ On an analog scope, you can also adjust the timebase very precisely since it's analog. Adjust it to match the exact frequency of your wave and it will line up. (maybe only for low frequencies? not sure) \$\endgroup\$ Commented Mar 13, 2021 at 11:06
  • 2
    \$\begingroup\$ You can get it "relatively" stable on screen but that takes a lot more work than turning the trigger button ;) But if you need to display a slowly moving waveform for a movie set, it works great! \$\endgroup\$
    – bobflux
    Commented Mar 13, 2021 at 11:22
  • \$\begingroup\$ Nice answer. I'll add that storage oscilloscopes go back to analog scopes in the 1960s, such as the Tektronix Type 549. They used special cathode-ray tubes that would store a trace until you pressed an "erase" button. So you could take a single sweep and study it at leisure, no film required. \$\endgroup\$ Commented Mar 13, 2021 at 23:07
  • \$\begingroup\$ What's so hard for a digital oscilloscope to adjust the display width to match to full cycles of the signal automatlly. I still don't see the point of a trigger. \$\endgroup\$
    – ronenfe
    Commented Jul 23, 2021 at 21:04
  • \$\begingroup\$ @RonenFestinger How would it know where the signal starts without the trigger? (the AUTO button tends to do what you want btw) \$\endgroup\$
    – bobflux
    Commented Jul 24, 2021 at 16:35
2
\$\begingroup\$

The Oscilloscope does not know when to trigger, to show you the graph like a sine wave.

You can test what you are assuming here (that without trigger the signal would appear just fine). Just take a osciloscope and put the trigger mark above the sine wave maximum Vout (or below the minimum Vout). You will just see a very fast moving signal.

But I think you can only test this out on old ocilloscopes. New ones will just say "slow trigger" and may not display anything.

A more practical example you can check this video (EEVblog), since its harder to explain it via text: https://youtu.be/ta096oBzSac?t=840

\$\endgroup\$
1
  • 3
    \$\begingroup\$ Ah, I watched your recommended YouTube and found it very good. EEVblog #159​ - Oscilloscope Trigger Holdoff Tutorial (20 min) youtube.com/watch?v=ta096oBzSac&t=840s. I usually agree that (1) a picture is worth one thousand words, (2) a video is worth 10 pictures, (3) TIY (Try It Yourself) is worth 10 videos. Cheers. \$\endgroup\$
    – tlfong01
    Commented Mar 12, 2021 at 1:19
1
\$\begingroup\$

Triggering basically gives you a timing reference so you you tinker around with a bunch of other waveforms on the scope without everything shifting around in the time-axis making you lose track of where things are supposed to be.

Oscilloscopes also use the trigger as a locator for where to zoom in around. So if you want a close-up you need to select the appropriate trigger.

\$\endgroup\$
1
\$\begingroup\$

Yes, for a repeating pattern the trigger point is pointless, you can trigger anywhere or do a manual trigger yourself or let it auto trigger not based on level.

So long as the scope captures a period and can display that whole period (ideally a little more) then where you trigger does not matter, you can see the whole pattern.

Because this is not the normal use case for a scope, scopes are designed to be heavily based on triggers, AND you can certainly view a repeating signal using a trigger point, so you are likely to use a trigger even on repeating patterns. It is more of a case of how scopes work for normal use cases than in one case I do not need a signal based trigger.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ Overlaying is really only valuable for repeating signals, because it allows you to perceive the irregularities/variances (an eye diagram would be one of the possible cases) \$\endgroup\$
    – Ben Voigt
    Commented Mar 12, 2021 at 15:46
  • \$\begingroup\$ I didnt know I said anything about overlaying. the OP's question is do necessarily need a trigger to see a repeating signal on a scope. the answers all seem to be related to some other question as to why you need triggers. The answer is absolutely not you do not need to trigger on the signal to see a sample of the period of the repeating signal. If you want to do more fancy stuff then sure...I think it may be an interpretation of the question. As written I think that is the question and the answer is no signal based trigger is needed at all. \$\endgroup\$
    – old_timer
    Commented Mar 12, 2021 at 16:10
  • \$\begingroup\$ Whether you want a single shot capture or you want to view a repeating signal, you have to have some form of trigger. For a one shot you could manually trigger the sweep. For everything else it takes a trigger circuit. \$\endgroup\$
    – JRE
    Commented Mar 13, 2021 at 9:30
  • \$\begingroup\$ Yep, as stated. \$\endgroup\$
    – old_timer
    Commented Mar 13, 2021 at 14:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.