Probe Capacitive Reactance

[nbritton]

So in the video below, is what Jack Ganssle said about high frequency probe ohm measurements really true? i.e. is my 10 Mohm probe really only a couple hundred of ohms at 500 MHz? I would love it if Dave could do a video on this topic. Anyone know what the capacitive reactance parameters are for Rigol's probes?

[Tim F]

he already has . It is true that a 10Meg probe is not 10Meg at anything other than low frequency. This is because they require capacitive compensation to counteract the capacitance at the scopes input i.e. the input impedance of the scope is not 1Meg at anything other than low frequency either!

[nbritton]

So how do you compute the resistance value of your probes with a square wave? An ideal square wave is an infinite series summation of all the odd harmonics.

[tggzzz]

So how do you compute the resistance value of your probes with a square wave? An ideal square wave is an infinite series summation of all the odd harmonics.

One harmonic at a time. The higher frequencies see a lower impedance and if the source can't drive it they are attenuated.

All tools and techniques have limitations, and you have to understand them.

In this case a "low impedance Z0 probe"  has a higher impedance.

[T3sl4co1l]

Remember the origin of impedance: it's an AC steady state parameter, and not meaningful when waveforms other than sine waves are present.

This is why you have to treat each frequency at a time, or treat the RLC circuit (not some presumed constant impedance -- something which doesn't exist at all frequencies) fully in the time domain.

The time domain solution, in general, is the solution of a differential equation (because the time domain equivalent of an inductor or capacitor contains a derivative).  For ideal passives only, it will have an analytical solution -- some combination of exponentials and ringing signals.  But if nonlinear elements are involved, you're better off doing it one small timestep at a time, e.g. SPICE.  If it goes at all, which is far from guaranteed.

Tim

[Berni]

Yup it is true probes start to really load down your circuits at high frequency. Due to this you can sometimes cause a circuit to stop working or worse make a non working circuit work. For example you might have some analog circuit that oscillates all over the place, but the extra 10pF or so you add by touching the probe to something can slow it down enough to make it stable again. Hence you see its clean as a whistle on the scope but when you stop looking at it then it goes straight back to oscillating.

Here is an example from the manual for the Agilent 1152A active 2.5GHz probe.



This probe has an amplifier inside the tip of the probe to buffer the signal and send it to the scope over a properly terminated 50Ohm line. Due to this it only has 100K of resistance at DC so not very good. But it comes in to its own at high frequency due to having only 0.6pF of capacitance. And keep in mind that 6pF it is being compared against is actually quite low capacitance for a scope probe.

Most scope probes have the DC resistance and capacitance written on them somewhere. For example on a lot of Agilent probes its here:

[tggzzz]

So in the video below, is what Jack Ganssle said about high frequency probe ohm measurements really true? i.e. is my 10 Mohm probe really only a couple hundred of ohms at 500 MHz? I would love it if Dave could do a video on this topic. Anyone know what the capacitive reactance parameters are for Rigol's probes?

It is a simplification in another way as well...

If you use a 6inch/150mm probe ground lead then, due to simple physics, that will have an inductance of ~120nH.

Now, what happens if you have an inductor connected to a capacitor? You have a resonant circuit, in this case with a frequency response peak around 100MHz.
Now, what happens if you put a voltage step, e.g. from a digital circuit, into a resonant circuit? It "rings" at ~100MHz.

For examples of that, and cures, see https://entertaininghacks.wordpress.com/2015/04/23/scope-probe-accessory-improves-signal-fidelity/

Note that "low-impedance Z0 probes" will have an input capacitance <1pF, short ground leads and a 500ohm damping resistance, all of which mean they are good to >1.5GHz. You can also make homebrew Z0 probes which are good to >1GHz; see https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/

[dom0]

There's a good reason why high-Z probes are sometimes-often called LF (low frequency) probes....