EEVblog #1320 - Premature Oscilloscope Triggering

[EEVblog]

Don't be caught out by embarrassing premature triggering!
Dave demonstrates how and why your oscilloscope may trigger before it's supposed to, and how to solve it.

[jbaribeault]

So I've got the Siglent SDS1104X-E that Dave showed doesn't have the HF Reject; what are my options when I see this? Some kind of filtering on the probe?

[2N3055]

There are two problems here: fact that you can have pulses that are so fast that you cannot see them at your current timebase, and fact that scope will trigger on them nevertheless.

That is exactly why Nnctnico and myself are advocating use of peak detect mode all the time when you're "scoping" around completely unknown signal. If you use peak detect mode, you will see the very short pulses and it will be obvious on what is scope triggering. It is not a solution but removes confusion.

Second thing is fact that many scopes today have advanced triggering modes in addition to simple edge trigger. Things like slope, window, timeout, pulse width trigger are there to help with these situations.
Or simple edge trigger with HF reject if that does the job.

But first thing is that you have to see all the details. In this video Dave nicely illustrated the problem...

[StillTrying]

Some of us think the SDS1104X-E's trigger is all digital so it can only interpolate between the captured 1ns ADC samples.
So when storing at 10MSa/s or 100MSa/s it's quite possible the trigger level will see some 1ns ADC values that count as a trigger level, but don't make it into the memory or screen trace.
- probably.
https://youtu.be/GZHnrGIK9V8?t=261

Edit:
What 2N3055 said, peak-detect should always show the level that caused the trigger.

[SilverSolder]

So I've got the Siglent SDS1104X-E that Dave showed doesn't have the HF Reject; what are my options when I see this? Some kind of filtering on the probe?

You could try a different trigger type than edge detect - for example, pulse width, if your scope supports that.

[Ed.Kloonk]

We've all suffered from premature triggering at some point. No need to be ashamed about it.

[David Hess]

Some of us think the SDS1104X-E's trigger is all digital so it can only interpolate between the captured 1ns ADC samples.
So when storing at 10MSa/s or 100MSa/s it's quite possible the trigger level will see some 1ns ADC values that count as a trigger level, but don't make it into the memory or screen trace.

The limited bandwidth before the digitizer spreads and lowers any pulse so *something* should be visible, but peak detection is the easy solution to not miss it.

The trigger has to operate during decimation but that is usually not the whole story.  Usually asecond "trigger qualifier" executes after the acquisition to see if the trigger conditions were actually met, and if not, the acquisition is discarded.  This is also the point where interpolation of the trigger point aligns the incoming acquisition with the display record.

[tautech]

So I've got the Siglent SDS1104X-E that Dave showed doesn't have the HF Reject; what are my options when I see this? Some kind of filtering on the probe?

Welcome to the forum.

HF and LF are in the each channels Trigger Coupling menu.

There are several lessons to be learnt from Daves video other than the obvious need for Peak select to be ON if you need to see what was actually triggered on.
Obviously mem depth that the acquisition system has to work with is different than what the trigger works with so what we see on the display may have hidden parts if mem depth settings are low.
Dave as this video shows was sidetracked into investigating the misalignment of the waveform of interest WRT the actual trigger point which at this point plainly indicates there's some other factors triggering the scope when in fact all he was needing to see was the PSU output rail rising edge.
Lifting the trigger level closer to the PSU output level may have it totally miss the glitches and trigger correctly on the waveform of primary interest.
The 20 MHz BW limit would also help control triggering of parts of the waveform that weren't of interest.

Lots of tools in these little X-E scopes to get the info you need once you understand clearly what that may be.

[David Hess]

What's actually going on here is that the trigger system inside the oscilloscope is a separate analog system to the analog-to-digital converter and what's displayed on the screen.

Older DSOs worked that way but newer ones with some exceptions going back at least a decade implement triggering in the digital domain after the analog-to-digital converter but before decimation which has the effect you described where it is possible to trigger on an event which is not displayed, as you showed in your video.  The solution is to use a longer record length, faster time/div setting, a trigger qualifier, or peak detection so that every sample point contributes to the display.

So there is no separate analog trigger path and this is easy to verify; for DSOs which share a digitizer between channels, set the trigger source to an unused vertical input which would normally share the digitizer, and watch to see if the sample rate for the displayed channel is reduced.

High frequency reject is a good solution but comes with a problem, at least with oscilloscopes that actually do implement analog triggering.  The low pass filter which implements high frequency reject introduces group delay so the trigger point no longer aligns with the unfiltered signal, which did *not* happen in your example.  At slower sweep speeds this would never be noticed but at higher sweep speeds, it can move the trigger point a significant distance or completely off of the display and introduce jitter.

This issue also comes up when the oscilloscope's bandwidth filter is not applied to the trigger signal path.  The early Tektronix 7A13 vertical amplifier had this problem but Tektronix fixed it in the late model by adding a separate bandwidth limiter to the trigger circuit path.  Usually the bandwidth filter could be implemented before the trigger pickoff but not in the case of the 7A13 because it would have interfered with the fast overload recovery feedback circuit.  For simplicity many analog oscilloscopes implemented the bandwidth filter after the channel switch but this means it does not affect triggering.

[TK]

So I've got the Siglent SDS1104X-E that Dave showed doesn't have the HF Reject; what are my options when I see this? Some kind of filtering on the probe?

Just raise the trigger level

[2N3055]

So I've got the Siglent SDS1104X-E that Dave showed doesn't have the HF Reject; what are my options when I see this? Some kind of filtering on the probe?

Just raise the trigger level

It DOES have HF reject... Just not where Keysight has it. It's not in the same menu..
Manuals are available, shame nobody reads them..

[StillTrying]

What happens to the sample rate on a 2 channel 1GSa/s scope when you use the EXT Trig, or even AC LINE as the trigger source?

Manuals are available, shame nobody reads them..

Of the 'affordable' scopes I've read them all over the years.

HF and LF are in the each channels Coupling menu.

The trigger's HF & LF Reject is on the Trigger menu.

[SilverSolder]

So I've got the Siglent SDS1104X-E that Dave showed doesn't have the HF Reject; what are my options when I see this? Some kind of filtering on the probe?

Just raise the trigger level

It DOES have HF reject... Just not where Keysight has it. It's not in the same menu..
Manuals are available, shame nobody reads them..

Is HF Reject for the displayed signal the same as HF Reject for the trigger, though?  (Perhaps if the scope gets the trigger from the displayed signal...)

[2N3055]

So I've got the Siglent SDS1104X-E that Dave showed doesn't have the HF Reject; what are my options when I see this? Some kind of filtering on the probe?

Just raise the trigger level

It DOES have HF reject... Just not where Keysight has it. It's not in the same menu..
Manuals are available, shame nobody reads them..

Is HF Reject for the displayed signal the same as HF Reject for the trigger, though?  (Perhaps if the scope gets the trigger from the displayed signal...)

You have selectable "BW Limit" (a 20 MHz low pass filter) for input channel and in Trigger Coupling you can choose:

 DC: allow DC and AC components into the trigger path.
 AC: block all the DC components and attenuate signals lower than 8 Hz. Use AC
coupling to get a stable edge trigger when your waveform has a large DC offset.
 LF Reject: block the DC components and reject the low frequency components lower
than 2 MHz. Low frequency reject removes any unwanted low frequency components
from a trigger waveform, such as power line frequencies, etc. that can interfere with
proper triggering. Use LF Reject coupling to get a stable edge trigger when your
waveform has low frequency noise.
 HF Reject: reject the high frequency components higher 1.2 MHz)

You would use  "BW Limit"  to intentionally make your scope 20 MHz scope and filter out higher frequencies because you don't even want to see them on the screen. You wan't to ignore them completely. You would use that, if, for instance, you have a known interference source, you know it's not your problem and you just want to concentrate on your signal. 

You would Trigger coupling filtering (both HF and LF reject) to keep looking at full signal with all components and full frequency range, but want to trigger on a specific part of it.

As I said many times before, even the cheapest scopes have many triggering options that only 10 years ago you had to buy mid range scope to get it.  Cheap Rigols, Siglents, MicSigs, Gw Insteks are good enough for large percent of usual scope work. Their more advanced models are even much more than that..

Biggest problem I see with people using B brands is expectation they will work exactly the same as their favorite A-brand ( i.e. they expect Siglent or Rigol will have exactly the same naming and have exactly the same menu layout, which they won't) so instead of learning new ways, they proclaim cheap instruments "stupid" and insist that, for instance, "Keysight has best layout" while it is only familiarity of using their equipment for 20 years... Or simply make false statements like "this instrument doesn't have XYZ" because it's not where they expect it and don't even bother to check if that is true. And then, when corrected, they attack back with " if it weren't stupid it would be here where it should be..", and back to square one.

I've seen this with musical instruments for years, where musicians would wax poetic about all kinds of keyboards, guitars, effects, amplifiers.. Instead of practicing playing and making music, which might make them better musicians, they waste time talking.
Take Chick Corea and give him cheapest, tuned, functional upright piano, and take mediocre player and give him the most expensive Yamaha Grand Piano.. Guess which one will make more beautiful music...

[tautech]

HF and LF are in the each channels Coupling menu.

The trigger's HF & LF Reject is on the Trigger menu.

Yes silly me, brain fart !   

You're quite right as these screenshots from a SDS1104X-E show the channel menu has only the 20 MHz BW limit.

[Kleinstein]

...
Is HF Reject for the displayed signal the same as HF Reject for the trigger, though?  (Perhaps if the scope gets the trigger from the displayed signal...)

The whole point of the video is that the trigger is not allways from the displayed signal, especially if the sampling rate is lower than the maximum.

[SilverSolder]

...
Is HF Reject for the displayed signal the same as HF Reject for the trigger, though?  (Perhaps if the scope gets the trigger from the displayed signal...)

The whole point of the video is that the trigger is not allways from the displayed signal, especially if the sampling rate is lower than the maximum.

Even if the scope doesn't have a separate analog trigger.  Got it -

[TK]

 HF Reject: reject the high frequency components higher 1.2 MHz)

Does it mean that scope is BW limiting to 1.2MHz?  Will it trigger on a 50MHz low to high edge signal?

[jbaribeault]

Turns out I hadn't updated the firmware on the scope; options are there Loving this scope - so many features for the cost; simply incredible.

[2N3055]

 HF Reject: reject the high frequency components higher 1.2 MHz)

Does it mean that scope is BW limiting to 1.2MHz?  Will it trigger on a 50MHz low to high edge signal?

I don't know how sharp the filter is. But I guess that 50MHz should be attenuated enough to achieve stable trigger on slower parts of waveform..
1.2 MHz is interesting choice.. Keysight 3000T has it at 50 kHz (both LF and HF)

[StillTrying]

HF Reject: reject the high frequency components higher 1.2 MHz)

Does it mean that scope is BW limiting to 1.2MHz?

The trigger's LF Reject only filters the HF noise and spikes off the signal used by the trigger, the BW of the scope doesn't change, and the screen still shows the same original signal + the HF noise. A channel's 20MHz BW limit would also effect the trigger if that channel is used for the trigger.

If the trigger's HF Reject is on you'd notice the trigger start to not work so well on >2MHz sine waves. It's very easy to forget that you've been using the trigger's HF Reject or Noise Reject, and then find that the scope's trigger barely works at high frequencies or small amplitudes.

[David Hess]

What happens to the sample rate on a 2 channel 1GSa/s scope when you use the EXT Trig, or even AC LINE as the trigger source?

Test it to find out; it depends on how the external trigger was implemented.  Some DSOs have a separate and simplified analog trigger only for the external in; you might be able to tell if only level triggering is supported.  And some include a separate digitizer but no or very limited sample memory.

Is HF Reject for the displayed signal the same as HF Reject for the trigger, though?  (Perhaps if the scope gets the trigger from the displayed signal...)

You have selectable "BW Limit" (a 20 MHz low pass filter) for input channel and in Trigger Coupling you can choose:

Exactly, nominally the bandwidth limit applies to the displayed signal and the trigger although I gave some examples were that was not true in older instruments.

The vertical input AC coupling (high pass, 6 to 120 Hz) and bandwidth limit (low pass, 20MHz and higher) are optimized for different conditions than the trigger's LF and HF reject which are typically around 50 kHz.

 HF Reject: reject the high frequency components higher 1.2 MHz)

Does it mean that scope is BW limiting to 1.2MHz?  Will it trigger on a 50MHz low to high edge signal?

It could completely miss it, just like LF reject could completely miss triggering on a line frequency waveform.

[StillTrying]

What happens to the sample rate on a 2 channel 1GSa/s scope when you use the EXT Trig, or even AC LINE as the trigger source?

Test it to find out;

I would if I could but I've only got the CML+ here, and that's different, it can certainly keep 1GSa/s RT (or even 10s of GSa/s ETS) on one channel while triggering off something else, - such as the channel that's switched off.

[David Hess]

What happens to the sample rate on a 2 channel 1GSa/s scope when you use the EXT Trig, or even AC LINE as the trigger source?

Test it to find out;

I would if I could but I've only got the CML+ here, and that's different, it can certainly keep 1GSa/s RT (or even 10s of GSa/s ETS) on one channel while triggering off something else, - such as the channel that's switched off.

ETS implies that it does support analog triggering but that is not absolute and there are some alternative triggering implementation which have not been discussed here.  Low and even medium cost oscilloscopes for more than a decade now have largely lacked analog triggering completely simply because of economics; with increasing digital integration, digital triggering becomes essentially free.

[TheNewLab]

So I've got the Siglent SDS1104X-E that Dave showed doesn't have the HF Reject; what are my options when I see this? Some kind of filtering on the probe?

Just raise the trigger level

It DOES have HF reject... Just not where Keysight has it. It's not in the same menu..
Manuals are available, shame nobody reads them..

Used to read them. Now, a Nikon DSLR manual is over 500pgs!  The New Rav 4 SUV has TWO manuals, 1st about 800 pgs the 2nd about 200pgs LOL

I fear in the near future reading the entire 3 volumes of Gibbon's work Decline and Fall of the Roman Empire will a quicker read than a product manual.

That said, the Siglent is kinda short