When measuring high speed signal with distortion, how do we know its source

[matrixofdynamism]

If we find that a signal has distortions when it should be more clean, how do we first prove that the scope and/or scope probe itself is not main cause of the signal to be distorted when measuring and causing that distorition to then appear on the scope display?

 

[chuckey]

If you can see distortion on a scope then the distortion is greater then 10%. The amplifiers in scopes are designed to be linear over the whole screen height.
Frank

 

[FvM]

I guess the term "distortion" doesn't exactly fit the phenomena you are seeing. More likely you are talking about effects like ringing edges, over- and undershoot or signal reflections. In other words anything that can be caused by unsuitable signal probing.

 

[matrixofdynamism]

@FvM, yes precisely. So how do we know if the observed undesired effect is actually present on the signal or is being caused by the signal probe.

 

[chuckey]

If you connect your probe to the cal output of the scope and examine the rising edge of the square wave you can see what ringing you have with the probe. In a decent probe you can minimise them by adjusting the compensating capacitor.
Frank

 

[jiripolivka]

With high speed signals I experienced many oscilloscope users failed to terminate scope inputs with a proper line impedance. Then they wondered about distortion.
Limiting usually does not matter to signal distortion. Resonances (tuned circuits, also unwanted) contribute a lot, also variable jitter (signal sources, echoes by unwanted coupling) can be seen.

 

[kripacharya]

A more detailed description of your expected 'signal', the source generating this signal, and of your oscilloscope itself will be useful in providing a clearer response about possible causes of distortion.

For example - if you have a high-end scope with good pre-adjusted & calibrated probes, and you are checking a home-brew breadboard circuit, then chances are its your circuit itself which is causing the 'distortions'.

 

[matrixofdynamism]

@kripacharya. I asked this question out of curiousity. I have the answer now. First we need to be sure that we are using a high end scope which can work with the frequency signal we are measuring and it must be properly terminated.

But then a question remains, the scope probe will always load the signal being measured to some extent. Does this mean that there are some high speed signals that should not be measured, or measuring them would cause such distortion that the design will malfunction?

 

[FvM]

My short version of my answer is:

You can't measure high-speed signals with a passive high impedance probe. To probe high speed signals undistorted, you need an active probe (unfortunately quite expensive) or a resistive probe, e.g. 1k 20:1 probe.

 

[volker@muehlhaus]

You can't measure high-speed signals with a passive high impedance probe. To probe high speed signals undistorted, you need an active probe (unfortunately quite expensive) or a resistive probe, e.g. 1k 20:1 probe.

Typically, high speed circuits operate at some defined impedance level and you can use the oscilloscope as the load (with 50 Ohm termination), instead of probing (in parallel) at some circuit location.