Why trigger behavior changes with vertical scale?

[bitbanger]

Hi folks - this has long been something that's really bothered me. Why is it when I change the vertical scaling even one division, I often run into a situation where I have to adjust my trigger to reliably trigger again?

Happened today, using a DSOX4034A for an inrush measurement. I was in trigger normal mode (so I didn't have to repeatedly arm), powering on and off the DUT to capture the current probe signal. As I'm honing in on the capture settings, I'm increasing vertical scale (higher V/div ie 1V/div to 5V/div), powering on, adjusting, powering on, adjusting timebase, delay, holdoff, etc, rinse and repeat. Well after I reach a certain point when I increase the vertical scale, I stop triggering. Drop back and I can trigger again. Say I'm measuring a peak inrush of 40A, my trigger was set to rising at 2A level. Now, just for example, I have to raise my trigger level to 10V to trigger again. I didn't honestly remember if I was increasing V/div or decreasing but you probably get the idea.

My suspicion is what I'm seeing in screen is essentially used to "software" trigger? And so with more on-screen vertical dynamic range with increasing V/division, I need to pull my trigger level out of the weeds so to speak?

Is there absolutely no analog trigger circuitry in modern storage oscilloscopes? Why is trigger ability dependant on vertical scaling?

I guess maybe I'm hung up on how an analog scope trigger worked, you set a level and that was that.

This has been something I've struggled to explain eloquently so would love to hear your thoughts.

[nctnico]

The reason is simple: the trigger input is taken from the vertical signal. So if the vertical signal gets larger / smaller, the signal to the trigger changes amplitude as well. And thus the triggering is affected.

[bitbanger]

Thanks for the reply. I suppose I was confused because I thought the signal used to compare for the trigger would not change - it is absolute. As does the trigger level setting. Perhaps I need to find a good block diagram of a modern front end (or check my assumptions about what is no different than a traditional setup)

[eTobey]

You have probably set the trigger quite high with falling edge or vise versa? There is something called hysteresis.

[rf-loop]

Thanks for the reply. I suppose I was confused because I thought the signal used to compare for the trigger would not change - it is absolute. As does the trigger level setting. Perhaps I need to find a good block diagram of a modern front end (or check my assumptions about what is no different than a traditional setup)

In digital oscilloscopes what use analog trig, typically signal is splitted  to trigger pathway just  before ADC
So level and offset changes in frontend before this split affect.

Lot of different if compare with old analog oscilloscopes where trigger system see whole signal. (when digital trigger see only this vertical band what is ADC range) 

If look 10V signal and just top of it using 100mV/div then ADC range is around 1V or bit more... and all what trigger system can know is inside this band)

If trigger system is fully digital. Full digital trigger engine is reading digital data from ADC only. What is not inside this data, trigger engine can not know.
Very extremely different if compare with old analog scopes or these older digital oscilloscopes what ahe analog sidepahway trigger.


Old simple image from my some "teaching" material when Siglent just started produce full digital trigger. With full igital trigger all happen after ADC using data from ADC.



Old a rough overview of what it consists of.

[bitbanger]

Thank you for the detailed reply - so there is some basis for my confusion haha.

In a modern, fully "digitized" trigger system: regardless if it is using attenuated, offset, amplified signal for compare, the trigger level *setting*, as shown on screen is presented with respect to the un-attenuated/original input signal to the BNC. So regardless of what happens in the back end, it is confusing because this trigger level setting relationship to the input signal does not change, again at least as presented to the operator.

[nctnico]

Thank you for the detailed reply - so there is some basis for my confusion haha.

In a modern, fully "digitized" trigger system: regardless if it is using attenuated, offset, amplified signal for compare, the trigger level *setting*, as shown on screen is presented with respect to the un-attenuated/original input signal to the BNC. So regardless of what happens in the back end, it is confusing because this trigger level setting relationship to the input signal does not change, again at least as presented to the operator.

In digital oscilloscopes, the level shown as a number on screen is scaled according to the input attenuation and offset. However, the amplitude matters as well as the trigger detector needs a bit of hysteresis. Usually the trigger sensitivity / hysteresis is specified in verticial divisions.

[David Hess]

My suspicion is what I'm seeing in screen is essentially used to "software" trigger? And so with more on-screen vertical dynamic range with increasing V/division, I need to pull my trigger level out of the weeds so to speak?

That is what is going on, but it would be the same with an analog trigger.

Is there absolutely no analog trigger circuitry in modern storage oscilloscopes? Why is trigger ability dependant on vertical scaling?

Modern DSOs have no analog trigger path.  The signal is digitized and the trigger recognizer operates in in digital hardware, often an FPGA, while the acquisition record is being filled.

An analog or digital trigger is dependent on the vertical scaling because that happens before the trigger examines the signal.  Vertical offset also affects the trigger, but vertical position may or may not depending on whether it is implemented before or after the trigger.

I guess maybe I'm hung up on how an analog scope trigger worked, you set a level and that was that.

I am not sure why you are having problems with the Keysight DSO's trigger.  You might try setting the trigger coupling to AC.

[David Hess]

In a modern, fully "digitized" trigger system: regardless if it is using attenuated, offset, amplified signal for compare, the trigger level *setting*, as shown on screen is presented with respect to the un-attenuated/original input signal to the BNC. So regardless of what happens in the back end, it is confusing because this trigger level setting relationship to the input signal does not change, again at least as presented to the operator.

That is implementation dependent, and largely a user interface decision.  Some (1) analog triggers work that way also, although it may not be apparent unless the trigger level is part of the readout.

(1) Which is to say, almost all Tektronix analog triggers going back to at least the 1980s if not earlier.  Tektronix peak-to-peak triggering does exactly what you describe but as a side effect of how it operates.  This might explain why these old oscilloscopes are known for excellent and easy to use triggering.  I think what they had is *better* then what I have seen on modern DSOs, even Tektronix ones, because it responds so much more quickly.  Tektronix replaced this with "auto level triggering" which sort of produces the same result but is much slower and more annoying, but has the advantage of requiring less dedicated hardware.