Oscilloscope input impedance

[nuno]

So, what's the resistance a 1Vpp or so 20MHz signal (sine) sees on a "tipical" probe-oscilloscope system (let's say a 100MHz Rigol with some 150MHz probes or similar)?

[edavid]

http://www.tek.com/learning/probes-tutorial

[AG6QR]

Here's the info for the Rigol RP2200, the standard x1/x10 passive probe that goes up to 150MHz

Input resistance: 10 megaohms +/- 2%

Input capacitance 17pF +/- 5 pF

These numbers are for the x10 setting.  If you go to the x1 setting, it's 1Mohm and 100pF.  The bandwidth is also much more limited at x1; it's about 7MHz versus 150MHz at x10.

I got that off of Rigol's website, searching for the tech support info for the RP2200.  Nearly any x10 passive probe for any make of oscilloscope will have the same 10Mohm nominal input impedance.  The capacitance isn't quite as standardized.  It will be generally in the same neighborhood, but may vary somewhat.

[nuno]

I guess you meant 9M for the probe, the other 1M is on the scope. And that is for a DC to low frequency signal. How about a 20MHz signal?

[T3sl4co1l]

He said what it is for AC: ~17pF.

So for frequencies above F = 1 / (2*pi*10Mohm*17pF) = 936Hz, the probe is capacitive*.

Up at 20MHz, well, that's four zeros and twice, or instead of 10M it's going to be around 500 ohms.  Or calculating... 468 ohm.

The phase shift should be around 10% lossy, i.e., mostly capacitive but it will load the circuit besides the added capacitance.

*A corollary: capacitive coupling to the probe is therefore flat above this frequency.  Which is why you can wave your probe near something, and observe the AC waveform with good fidelity, though of course without low frequency content, DC offset or calibrated amplitude.  But that's good to know for qualitative purposes.

Tim

[nuno]

He said what it is for AC: ~17pF.

So for frequencies above F = 1 / (2*pi*10Mohm*17pF) = 936Hz, the probe is capacitive*.

Up at 20MHz, well, that's four zeros and twice, or instead of 10M it's going to be around 500 ohms.  Or calculating... 468 ohm.

No need to take into consideration the cable + scope input capacitance? Or the 17pF speced for the probe already takes that into consideration?

[T3sl4co1l]

Yes, the probe is just an open connection (well, at DC) until it's connected to the scope.  The numbers are for the complete system.

[nuno]

So, a scope isn't really a "high input impedance" instrument for frequencies above, maybe a MHz or so? Or how do we define "high impedance"? Why is there such a worring in having a vertical amplifier with an input impedance of many many Megs at and "near" DC? And finally, what kind of techniques are employed to make an amplifier (even if gain = 1) for a 20MHz, or 50MHz or even 100MHz or more signal with 1M or more input impedance (assuming it makes sense to have an amplifier with such input characteristics)? Using inductances to counteract capacitive parasitics?

[tggzzz]

So, a scope isn't really a "high input impedance" instrument for frequencies above, maybe a MHz or so?

Fundamentally correct, but for some purposes replace "MHz" with "kHz".

In particular notice that at a few hundred MHz "10:1 low impedance Z0" probes with an impedance of 500ohms are actually higher impedance than so-called high impedance probes.

Or how do we define "high impedance"?

Relative to the impedance of whatever is being measured. For many audio circuits, "high" would be >>600ohms. For RF, >>50ohms or >>75ohms.

[T3sl4co1l]

It would of course be impractical to attempt to guarantee a 1M input impedance over an extremely wide range.  Even the best bootstrapped JFET buffers I'm aware of aren't much under 0.1pF, which is less than the BNC connector anyway, so what's the point?

And that leaves no room to put in attenuators.  So, they specify it -- and tune it to -- something much more practical: instead of attempting to remove capacitance, they set it to something manageable.  Usually 10 to 30pF.

Each stage of the attenuator has to be compensated so it presents the same capacitance to the outside, and ahead to the next stage, which makes alignment tricky (check the pulse shape and the input capacitance for every attenuator setting..), but they put the trimmers in to do that.  Or these days, it's probably toleranced well enough by design so it doesn't need compensation; there usually only one or two stages, switched by relays, and the rest of the range is either handled digitally (with enough extra ADC bits) or with a PGA (programmable gain amp).

Tim

[edavid]

So, a scope isn't really a "high input impedance" instrument for frequencies above, maybe a MHz or so?

Right.

Why is there such a worring in having a vertical amplifier with an input impedance of many many Megs at and "near" DC?

1M is not "many many Megs", but it's in order to measure high impedance circuits.

And finally, what kind of techniques are employed to make an amplifier (even if gain = 1) for a 20MHz, or 50MHz or even 100MHz or more signal with 1M or more input impedance (assuming it makes sense to have an amplifier with such input characteristics)?

It's always a FET source follower used as an impedance converter.

Using inductances to counteract capacitive parasitics?

In the way you're talking about, that's not possible in a broadband circuit, but you can look up peaking coils and T-coils.

Here's a good book on the subject:  http://www.davmar.org/TE/TekConcepts/TekVertAmpCircuits.pdf

[The Electrician]

So, what's the resistance a 1Vpp or so 20MHz signal (sine) sees on a "tipical" probe-oscilloscope system (let's say a 100MHz Rigol with some 150MHz probes or similar)?

Have a look at this: https://www.eevblog.com/forum/reviews/cheapest-100mhz-oscilloscope-probes-hands-on-review/msg99277/#msg99277

[nuno]

Thanks for the answers guys, very enlightning.

Why is there such a worring in having a vertical amplifier with an input impedance of many many Megs at and "near" DC?

1M is not "many many Megs", but it's in order to measure high impedance circuits.[/quote]
I meant, forget the "forced" 1M, the JFET has many M input impedance (as far as I understood...).

Here's a good book on the subject:  http://www.davmar.org/TE/TekConcepts/TekVertAmpCircuits.pdf

Thanks, it's a nice book I've already been reading parts of. Not "enough" of a book, though. I also find John Kobbe's et all passive probe patent from Tektronix early days to be consice and interesting: http://www.google.com/patents/US2883619

Looks like an interesting discussion Electrician, but I'm gonna have to go learn about Smith charts.

I was reading about probes/vertical amplifier and making some calculations, and realized that the input impedance for such signal would be in the 500 Ohm, and initially "I couldn't believe it".

Needless to say, I'll never look at the scope the same way.

[tggzzz]

I was reading about probes/vertical amplifier and making some calculations, and realized that the input impedance for such signal would be in the 500 Ohm, and initially "I couldn't believe it".

Needless to say, I'll never look at the scope the same way.

Always understand your instrument's limitations - whatever the instrument and whatever the manufacturer. Doubly true for modelling and simulation, of course.