0 Members and 1 Guest are viewing this topic.

electrolust

• Supporter
• Posts: 398
• Country:
passive probe capacitive load vs DUT
« on: March 17, 2017, 12:16:31 am »
Say I want to probe a mosfet gate, to evaluate ringing, that has Ciss = 10pF.
My passive probe has an input capacitance of 10pF and the scope input itself has 10pF.

AIUI the total probe capacitance are the 2 caps in parallel, IOW 5pF in this case.

Now when I probe the circuit, instead of 10pF I have 15pF.  And that's assuming perfect probing right at the terminals.  If I want to find Rg = (L/C)^0.5 to suppress ringing, doesn't the 1.5x capacitance exaggerate the Rg value by (1.5)^0.5 = 22.5%?

Ignoring the fact that my scope is lower bandwidth and can't actually take advantage of high bandwidth probes, if I wanted much better probing of this circuit wouldn't I want an active probe with (say) 1pF or lower input capacitance?  But even with 1pF (active, so the scope's capacitance doesn't come into play) I'm looking at (1.1)^.5 = 5% error.

Correct me if I'm wrong, but since I'd have to obtain the active probe at considerable cost, given my objective I think it's safe to just reason about the actual unprobed circuit behavior and correct for it.  And actually in this specific case I don't mind if the end result is overdamped.  (Underdamped would be bad though.)  But anyway I did want to know if I correctly understand the effects of the probe itself on this circuit.

tggzzz

• Super Contributor
• Posts: 10227
• Country:
Re: passive probe capacitive load vs DUT
« Reply #1 on: March 17, 2017, 12:36:05 am »
My passive probe has an input capacitance of 10pF and the scope input itself has 10pF.

AIUI the total probe capacitance are the 2 caps in parallel, IOW 5pF in this case.

1: if you have 10pF in parallel with 10pF, then you have 20pF.
2: that's not how scopes and probes work anyway. The load on the DUT is the quoted probe tip capacitance in parallel with the quoted probe tip resistance.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less

• Super Contributor
• Posts: 1875
• Country:
• Reactor Operator SSN-583, Retired EE
Re: passive probe capacitive load vs DUT
« Reply #2 on: March 17, 2017, 01:24:13 am »
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer

electrolust

• Supporter
• Posts: 398
• Country:
Re: passive probe capacitive load vs DUT
« Reply #3 on: March 17, 2017, 07:06:49 am »
My passive probe has an input capacitance of 10pF and the scope input itself has 10pF.

AIUI the total probe capacitance are the 2 caps in parallel, IOW 5pF in this case.

1: if you have 10pF in parallel with 10pF, then you have 20pF.

oops, of course.  I meant that the total capacitance seen by the DUT is the 2 caps in series.

Quote
2: that's not how scopes and probes work anyway. The load on the DUT is the quoted probe tip capacitance in parallel with the quoted probe tip resistance.

Is http://electronics.stackexchange.com/a/208561 not the correct diagram for a typical scope probe?  Ignoring Cc, doesn't the DUT see Cp+Cs?  (plus Rp+Rs of course)  For sure, I know that the resistive load is not the probe tip load, say 9M for a 10:1 probe -- to get 10:1 you take the 9M in series with the scopes 1M.  Why wouldn't the capacitive load be done the same way, incorporating the scope's input capacitance?

I'm looking at a probe spec sheet that says 10M, but with a footnote that says: in combination with the scope's 1M input.  There is no such footnote for the capacitive load spec.
« Last Edit: March 17, 2017, 07:08:48 am by electrolust »

electrolust

• Supporter
• Posts: 398
• Country:
Re: passive probe capacitive load vs DUT
« Reply #4 on: March 17, 2017, 07:12:45 am »
Make your own fet probe for pennies on the dollar -

Sure, I was going to get there ... wanted to make sure I understood the need to even do so though.

Thanks for the all the links!  I already knew of the existence of DIY FET probes, although not sure which would be reasonable.  It's great to have a curated list to follow.  But, first things first ...

electrolust

• Supporter
• Posts: 398
• Country:
Re: passive probe capacitive load vs DUT
« Reply #5 on: March 17, 2017, 07:21:43 am »
Now when I probe the circuit, instead of 10pF I have 15pF.  And that's assuming perfect probing right at the terminals.  If I want to find Rg = (L/C)^0.5 to suppress ringing, doesn't the 1.5x capacitance exaggerate the Rg value by (1.5)^0.5 = 22.5%?

Sorry, another error.  I guess that would be: understate by 1-(1/1.5)^0.5 = 18%.  So if that's the case, then it's a little more important that I get it right since just ignoring the error means the circuit will be underdamped, which is not desirable for me.  Otherwise, the same point applies, can't I just get away with a correction factor?

electrolust

• Supporter
• Posts: 398
• Country:
Re: passive probe capacitive load vs DUT
« Reply #6 on: March 17, 2017, 08:14:48 am »
2: that's not how scopes and probes work anyway. The load on the DUT is the quoted probe tip capacitance in parallel with the quoted probe tip resistance.

I found http://www.syscompdesign.com/assets/images/AppNotes/probes.pdf which explains it in pretty easy to understand terms.  IIUC from section 5, the probe capacitance can pretty much be disregarded, because the effective probe capacitance (fixed cap + adj cap in series) "must" be 1/9 (for 10x probe) of the scope's input capacitance in order for the probe to be compensated for flat response.  So we can calculate capacitance seen at DUT solely in terms of the scope input capacitance = Cs/10.

That seems pretty obviously clear from that doc, but then why do you say it's just the probes quoted capacitance, since clearly (from the doc) that value is irrelevant.  But if that value is meaningless, why does every probe even list this spec?  EDIT: so that you can make sure it will be compatible with your scope?
« Last Edit: March 17, 2017, 09:01:26 am by electrolust »

The following users thanked this post: nugglix

tggzzz

• Super Contributor
• Posts: 10227
• Country:
Re: passive probe capacitive load vs DUT
« Reply #7 on: March 17, 2017, 08:54:48 am »
My passive probe has an input capacitance of 10pF and the scope input itself has 10pF.

AIUI the total probe capacitance are the 2 caps in parallel, IOW 5pF in this case.

1: if you have 10pF in parallel with 10pF, then you have 20pF.

oops, of course.  I meant that the total capacitance seen by the DUT is the 2 caps in series.

Quote
2: that's not how scopes and probes work anyway. The load on the DUT is the quoted probe tip capacitance in parallel with the quoted probe tip resistance.

Is http://electronics.stackexchange.com/a/208561 not the correct diagram for a typical scope probe?  Ignoring Cc, doesn't the DUT see Cp+Cs?  (plus Rp+Rs of course)  For sure, I know that the resistive load is not the probe tip load, say 9M for a 10:1 probe -- to get 10:1 you take the 9M in series with the scopes 1M.  Why wouldn't the capacitive load be done the same way, incorporating the scope's input capacitance?

I'm looking at a probe spec sheet that says 10M, but with a footnote that says: in combination with the scope's 1M input.  There is no such footnote for the capacitive load spec.

As a rule of thumb, ignore stackexchange for any question which requires deeper understanding and discussion than "which button do I press to get X to go Y".

That is the traditional simplified model of a *10 scope probe that is suitable for considering how and why the compensation capacitor works. It doesn't mention the critical point that the probe cable is a very lossy transmission line. If you measure the resistance of the centre conductor, you will find it is a couple of hundred ohms.

There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less

electrolust

• Supporter
• Posts: 398
• Country:
Re: passive probe capacitive load vs DUT
« Reply #8 on: March 17, 2017, 09:06:20 am »
great refs, I probably have to wait until next week to get through them.

alm

• Super Contributor
• Posts: 1257
• Country:
Re: passive probe capacitive load vs DUT
« Reply #9 on: March 17, 2017, 09:21:02 am »
I found http://www.syscompdesign.com/assets/images/AppNotes/probes.pdf which explains it in pretty easy to understand terms.  IIUC from section 5, the probe capacitance can pretty much be disregarded, because the effective probe capacitance (fixed cap + adj cap in series) "must" be 1/9 (for 10x probe) of the scope's input capacitance in order for the probe to be compensated for flat response.  So we can calculate capacitance seen at DUT solely in terms of the scope input capacitance = Cs/10.

That seems pretty obviously clear from that doc, but then why do you say it's just the probes quoted capacitance, since clearly (from the doc) that value is irrelevant.  But if that value is meaningless, why does every probe even list this spec?  EDIT: so that you can make sure it will be compatible with your scope?
The schematic in that appnote is oversimplified, it ignores the cable. The cable adds another ~50-100 pF, which is in parallel with the scope input, so the effective input as seen from the 9 MOhm resistor becomes 1 MOhm plus a few hundred Ohm for the lossy cable, in parallel with 10 pF (scope) + lets say 70 pF (cable). This would require about 8 pF capacitance in parallel with the attenuation resistor. The series capacitance of the attenuator (8 pF) and the scope input (80 pF) would be about 9 pF. Add to this a few pF for the probe tip and to add range to compensate for different scope input impedances, and you end up at the 9-15 pF or so that are common for 1:10 high-Z probes.

The end result is the input capacitance figure quoted by the probe manufacturer. That should be a fairly accurate representation of the total input impedance of the probe + matching scope. Of course as you add additional leads and hooks to the probe, these themselves may affect the impedance. A short ground lead will add a couple of dozens nH of inductance, for example.

The Tektronix Probe Circuits concepts book, linked from the page tggzzz posted, explains all this in great detail.

The following users thanked this post: nugglix, electrolust

David Hess

• Super Contributor
• Posts: 10012
• Country:
• DavidH
Re: passive probe capacitive load vs DUT
« Reply #10 on: March 17, 2017, 04:27:01 pm »
You guys covered the probe tip capacitance; it is specified independently of the oscilloscope input capacitance.  There is some variation when the compensation is adjusted depending on the probe design but it is usually not significant.  If you want to independently find the probe tip capacitance, measure its rise time from a medium impedance source and then add enough parallel capacitance at the tip to double it; now the added capacitance is equal to the probe tip capacitance.

As far as the MOSFET gate with Ciss = 10pF, either make some other measurement or:

1. Build or use an active probe with a significantly lower input capacitance.
2. Build a quick and dirty medium impedance resistive divider to feed the oscilloscope probe.  This will divide the probe's roughly 10pF of input capacitance down to a reasonable level at the cost of lowering the sensitivity.

Below is an example of a medium speed 0.29pF (or 0.37pF) active probe from Troubleshooting Analog Circuits by Bob Pease.

The following users thanked this post: electrolust

electrolust

• Supporter
• Posts: 398
• Country:
Re: passive probe capacitive load vs DUT
« Reply #11 on: March 17, 2017, 10:01:18 pm »
I'm looking at a probe spec sheet that says 10M, but with a footnote that says: in combination with the scope's 1M input.  There is no such footnote for the capacitive load spec.

I think I get it now at a surface level, but I will work my way through all the links to get a deeper understanding.  So the reason there's no footnote is that, that number is the final number.  The 10M number has a footnote because it requires the 1M and not 50r input.

So that's even worse for my measurement case then.  The error is 29% understated.  I better build a DIY active probe, and it will be interesting to compare.

• Super Contributor
• Posts: 1875
• Country:
• Reactor Operator SSN-583, Retired EE
Re: passive probe capacitive load vs DUT
« Reply #12 on: March 18, 2017, 12:34:36 am »
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer

StillTrying

• Super Contributor
• Posts: 2268
• Country:
• Country: Broken Britain
Re: passive probe capacitive load vs DUT
« Reply #13 on: March 19, 2017, 02:49:05 am »
Perhaps you could prob the gate through a 1.5pF cap.

Here's a quick LT simulation of the active probe circuit 2 posts above if NE1 wants to play with it.
It looks OK, BW 100M+, gain very nearly 1, input cap ~1.5pF, and DC levels all over the place.
CML+  That took much longer than I thought it would.

electrolust

• Supporter
• Posts: 398
• Country:
Re: passive probe capacitive load vs DUT
« Reply #14 on: March 20, 2017, 10:06:25 pm »
The schematic in that appnote is oversimplified, it ignores the cable.

OK, the schematic may have been, but the discussion seems to include it.  As does the stackexchange discussion.  But maybe not in a useful way.

The Tek ABC document referenced a few times in this short thread does NOT mention cable capacitance.  That document clearly assumes that the probe tip capacitance spec INCLUDES the effect of the total probe circuit, through to the scope input capacitance.  It consistently discusses the probe's 10M resistance, when this is literally 9M + 1M at the scope, in the same breath that it discusses the probe's input capacitance as it affects the DUT.  When again, it's the entire probe+scope capacitance that affects the DUT.  So while that doc does explain a lot of things in good depth, I think it's still lacking in explaining my question.

One of the things I really disliked about the ABC doc is the heavy push on "buy official licensed OEM probes only".  Geez, get over it.

Agilent 5990-9175EN is an interesting doc, that explicitly considers the cable capacitance.  I think that doc is quite more approachable than the Tek doc.  It doesn't have the coverage that the Tek doc does, but it does get [quickly and directly] to the point that I specifically care about.

I think my error was that I thought the 1:9 capacitance factor (and only that factor) was seen at the DUT.  I have to sit down with pencil but I think it's the same as the 9:1 factor for the resistance.  Yeah the tip+cable has to have the 1:9 factor so that at the voltage divider point (this is well described in the appnote I linked) there is a balance, for flat response, but the overall capacitance seen by DUT is the entire capacitance.  Which is captured in the tip capacitance spec.

Now, the fun part: building a DIY active probe.
« Last Edit: March 20, 2017, 10:09:07 pm by electrolust »

edpalmer42

• Super Contributor
• Posts: 1646
• Country:
Re: passive probe capacitive load vs DUT
« Reply #15 on: March 20, 2017, 10:28:36 pm »
If all you want to do is look for ringing, there's a much easier solution than building an active probe.  Use the probe hook to hold onto a 1K resistor and probe the circuit with the open end of the resistor.

At 100 MHz, your 10pf probe has an impedance of ~160R.  With 1K in series with it, it's effect on the circuit becomes negligible.  Obviously, the level is affected, but you can factor that out.

At the very least, it's a quick and easy way to see if you need to build an active probe.

Ed

alm

• Super Contributor
• Posts: 1257
• Country:
Re: passive probe capacitive load vs DUT
« Reply #16 on: March 20, 2017, 10:49:12 pm »
The Tek ABC document referenced a few times in this short thread does NOT mention cable capacitance.  That document clearly assumes that the probe tip capacitance spec INCLUDES the effect of the total probe circuit, through to the scope input capacitance.  It consistently discusses the probe's 10M resistance, when this is literally 9M + 1M at the scope, in the same breath that it discusses the probe's input capacitance as it affects the DUT.  When again, it's the entire probe+scope capacitance that affects the DUT.  So while that doc does explain a lot of things in good depth, I think it's still lacking in explaining my question.
Yes, the ABC document is only about using (and selling ) probes, not designing them. For the user the only thing important is the input impedance spec of the entire system (i.e. 1:10 probe connected to a compatible 1 MOhm scope). The important thing for the user is that the probe has an impedance of 10 MOhm // ~10 pF and that it attenuates the signal by 10x (usually corrected automatically by modern scopes). If you want more details about the internals, I highly recommend the Tektronix Probe Circuits Concepts book.

If all you want to do is look for ringing, there's a much easier solution than building an active probe.  Use the probe hook to hold onto a 1K resistor and probe the circuit with the open end of the resistor.
A 10 MOhm // 10 pF probe in series with a 1 kOhm resistor with zero parasitic capacitance would have a -3 dB bandwidth of only ~8 MHz, falling off with -20 dB/decade. So any signal with a bandwidth beyond a few MHz will be severely distorted. Granted, real world performance will be better due to parasitic capacitance, but the bandwidth of any uncompensated voltage divider will be terrible regardless.

edpalmer42

• Super Contributor
• Posts: 1646
• Country:
Re: passive probe capacitive load vs DUT
« Reply #17 on: March 20, 2017, 11:37:51 pm »
If all you want to do is look for ringing, there's a much easier solution than building an active probe.  Use the probe hook to hold onto a 1K resistor and probe the circuit with the open end of the resistor.
A 10 MOhm // 10 pF probe in series with a 1 kOhm resistor with zero parasitic capacitance would have a -3 dB bandwidth of only ~8 MHz, falling off with -20 dB/decade. So any signal with a bandwidth beyond a few MHz will be severely distorted. Granted, real world performance will be better due to parasitic capacitance, but the bandwidth of any uncompensated voltage divider will be terrible regardless.

Yeah, I phrased that poorly.     100 MHz is probably too high for that trick to work very well.  If the ringing is only 10 MHz, it should work.

Another idea would be to take the 'resistor probe' idea to the logical conclusion and make a resistive divider probe.  There have been one or two threads about those probes.  The frequency response can reach 1 GHz with only an hour or two spent building the probe.

Ed

alm

• Super Contributor
• Posts: 1257
• Country:
Re: passive probe capacitive load vs DUT
« Reply #18 on: March 20, 2017, 11:54:05 pm »
If the low(ish) DC resistance is acceptable, then I agree that a 'Low-Z' probe is an excellent way for a good low capacitance probe that, if well constructed, can easily have a bandwidth of a few hundred MHz (or more). But even a 20x probe is only 1 kOhm at DC. If you have voltage to burn, then even higher attenuation is possible. A 200x probe would be 10 kOhm, but you would be down to 10 mV/div or so for a 10 V signal. The advantage is that the capacitance can be tiny (commercial probes using the same principles get < 1 pF), so the impedance at 100 MHz is much higher than the 160 Ohm a so-called high-impedance probe might have.

A low-Z probe is much simpler and easier to make than an active probe. And loading at high frequencies can actually be pretty similar, since that is mostly determined by capacitance.

Link to an article by Howard Johnson, who certainly did not invent it (it is already described in the Tektronix Concepts book from the sixties), but made the DIY approach popular:
http://www.sigcon.com/Pubs/straight/probes.htm

tggzzz

• Super Contributor
• Posts: 10227
• Country:
Re: passive probe capacitive load vs DUT
« Reply #19 on: March 21, 2017, 12:06:23 am »
If the low(ish) DC resistance is acceptable, then I agree that a 'Low-Z' probe is an excellent way for a good low capacitance probe that, if well constructed, can easily have a bandwidth of a few hundred MHz (or more). But even a 20x probe is only 1 kOhm at DC.

And a standard *10 "high impedance" probe has a significantly lower impedance even at 100MHz, let along 1GHz! (15pf at 100MHz is 100ohms).

Moral: friends don't let friends use *10 "high" impedance probes at 50MHz and above
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less

alm

• Super Contributor
• Posts: 1257
• Country:
Re: passive probe capacitive load vs DUT
« Reply #20 on: March 21, 2017, 12:15:29 am »
And a standard *10 "high impedance" probe has a significantly lower impedance even at 100MHz, let along 1GHz! (15pf at 100MHz is 100ohms).
Funny, that sounds quite similar to the last sentence of the paragraph you quoted .

Moral: friends don't let friends use *10 "high" impedance probes at 50MHz and above
Better friends teach their friends about reactance and source impedance, so they can make an intelligent trade-off instead of relying blindly on rules of thumb .

edpalmer42

• Super Contributor
• Posts: 1646
• Country:
Re: passive probe capacitive load vs DUT
« Reply #21 on: March 21, 2017, 02:06:46 am »

electrolust

• Supporter
• Posts: 398
• Country:
Re: passive probe capacitive load vs DUT
« Reply #22 on: March 21, 2017, 09:08:51 am »
I get it now.  (I think  )

I just watched w2aew (Alan) 's #9 video on YT, basics of probes.  It doesn't introduce anything I hadn't read already, here or in the links everyone provided, but Alan's great way of explaining things, watching it instead of reading it, his clever use of "animation", and something as simple as highlighting the capacitive elements in his probe circuit drawing, all of these combined now makes me smack my head as to how simple it is.  Or could be that I now misunderstand enough to be in bliss territory.

To summarize, the large cable capacitance is in parallel with the scope's small input capacitance.  Combined, this is essentially the cable capacitance.  Then, the small probe tip capacitance (compensation cap) is in series with that.  Combined, this is essentially the probe tip capacitance.

alm said as much, but it was the way w2aew presented it that got through to me.  Now to binge watch Alan's videos ...

The following users thanked this post: w2aew

tggzzz

• Super Contributor
• Posts: 10227
• Country:
Re: passive probe capacitive load vs DUT
« Reply #23 on: March 21, 2017, 10:19:39 am »
To summarize, the large cable capacitance is in parallel with the scope's small input capacitance.  Combined, this is essentially the cable capacitance.  Then, the small probe tip capacitance (compensation cap) is in series with that.  Combined, this is essentially the probe tip capacitance.

At high frequencies the cable's capacitance is irrelevant; it acts as a transmission line. In order to minimise the effect of reflections when the signal reaches the 1Mohm//20pF scope input, the scope's cable is a lossy transmission line.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less

macboy

• Super Contributor
• Posts: 1978
• Country:
Re: passive probe capacitive load vs DUT
« Reply #24 on: March 21, 2017, 01:18:05 pm »
2: that's not how scopes and probes work anyway. The load on the DUT is the quoted probe tip capacitance in parallel with the quoted probe tip resistance.

I found http://www.syscompdesign.com/assets/images/AppNotes/probes.pdf which explains it in pretty easy to understand terms.  IIUC from section 5, the probe capacitance can pretty much be disregarded, because the effective probe capacitance (fixed cap + adj cap in series) "must" be 1/9 (for 10x probe) of the scope's input capacitance in order for the probe to be compensated for flat response.  So we can calculate capacitance seen at DUT solely in terms of the scope input capacitance = Cs/10.

That seems pretty obviously clear from that doc, but then why do you say it's just the probes quoted capacitance, since clearly (from the doc) that value is irrelevant.  But if that value is meaningless, why does every probe even list this spec?  EDIT: so that you can make sure it will be compatible with your scope?
The above linked paper is not authoritative. What are that guy's credentials? Just because you read it on the internet, doesn't mean it's true. With that, read on... (tongue in cheek)
The paper you linked implies that the load will see only the series combination of the small probe tip capacitor and the scope input capacitance. Well, that's not true in the real world. In the real world, there are stray/parasistic capacitances to deal with. These add significantly to the overall scope probe load, and account for most of the capacitance seen by the circuit under test.

If the probe specifications say that it presents a load of 10 Mohm in parallel with 10 pF, then that is exactly what the probe tip load is. No guesswork involved. The oscilloscope capactiance has literally no impact on this spec. A scope with a smaller input capacitance and one with a larger one will result in the same probe tip capacitive loading. There is a really good reason for this. For most 10x probe designs, at the BNC connector end of the probe there is a "compensation box" which has, among other things, a small trimmer capacitor in it. This is placed in parallel with the scope's input, so that this capacitor is effectively in parallel with the scope input's capacitance. When you adjust the probe's compensation against the scope's cal output, this is the trimmer you are adjusting**. Since the capacitance at the probe tip end is fixed, this one needs to be variable, so that the ratio of the (fixed) probe tip capacitance to the scope input capacitance yields the exact same 10x ratio that the resistive components have (10 Mohm : 1 Mohm). This ensures that the low frequency components, which flow mostly through the resistors, and the high frequency components, which flow mostly through the capacitors, are all attenuated by 1/10. This adjustable trimmer cap, when adjusted correctly, ensures that all scope inputs effectively have the same capacitance. Otherwise, the probe couldn't work.

If the 10 pF probe load is too high, you have options. Tek currently makes some passive 10x 10 Mohm probes with about 4 pF capactance, much better. You would need to determine for yourself if they are compatible with your scope (meaning is your scope's input capacitance within the range that can be trimmed by the compensation capacitor). Another option is a Low-Z passive probe, which presents a 1 kohm load to the DUT, but only 1 pF or less. So at very high frequencies (> 30 to 50 MHz), the effective total impedance is lower than so-called High-Z probes. These attenuate by 20x and require a 50 ohm scope input. Going even higher in price you can get into active FET probes.

Tektronix has published good papers on probe theory and design. You can believe what you read in those.

** In some probes, especially cheaper low frequency ones, the compensation trimmer is at the probe end, not the BNC end, but this is much less common in quality probes. In those probes there will be a fraction of a pF difference in probe tip load depending on the the trimmer setting.

Smf