EEVblog HVP70 Differential Probe

[Vestom]

* CT3688 200MHz X10 (really X5). $989 on Amazon US, but seen as low as $568. This is a Sapphire design (SI-200) and it is available under some other names like Testec TT-SI 200. Seems similar to the Keysight N2792A ($1917)
Note: The CT3688 is quoted as X10, but it's X10 "into 50 Ohms", so it's really a X5 probe for comparison.

It does not make much sense to make a 200MHz probe into 1MOhms, so it is really a X10 probe. Running it without 50 Ohms termination will likely ruin your signal integrity and frequency response.

[CDaniel]

That's not true at all , the input impedance is 2x500Kohm and the probe has 5x attenuation into 1Mohm input oscilloscope and 10x into 50ohm . Because it has a 50ohm output series resistor ... You can use it anyway without ruining anything

[Vestom]

That's not true at all , the input impedance is 2x500Kohm and the probe has 5x attenuation into 1Mohm input oscilloscope and 10x into 50ohm . Because it has a 50ohm output series resistor ... You can use it anyway without ruining anything

Hmmm. Let us see about that:

• The CT3688 probe is specified to output into scope 50R termination: https://www.caltestelectronics.com/images/attachments/CT3688_datasheet_20170525.pdf
• The probe will then presumably be calibrated when driving a 50R termination. When driving 1M you will get full "benefit" of any tolerance in the source impedance. That will likely add some % DC tolerance.
• When driving 1MOhm you get a ~100% reflection. If your source termination or cable impedance is less than perfect you WILL get standing waves - and nothing is perfect...
• Your 1MOhm scope input will likely have a 10-20pF capacitance in parallel presenting an impedance of 40-80 Ohm @200 MHz and a significant phase angle. The CT3688 does not have any adjustment to compensate for that.

Now, why on earth not just enable the 50R termination in the scope (or add an external one) and get the full precision of your expensive test equipment?? (And a nice 10X multiplier as specified...)

And of course, the proof is in the pudding, so I attached an example of running a pulse into 50R and 1M

[rvalente]

Other than because of BNC isolation, why the majority of differential probes has banana jacks instead of bnc?

Using the Banana to BNC adapter brings problems in frequency response?

[Serpent10i]

They use banana because they want to have two equal paths so that the signals are as similar as possible and they can reject as much common mode noise as possible.  Remember these probes are measuring the difference between the two measurment points, not just one point and ground. Therefor instead of a center conductor and outer ground you want to have two similar conductors, making these a twisted pair is even better.

[rvalente]

They use banana because they want to have two equal paths so that the signals are as similar as possible and they can reject as much common mode noise as possible.  Remember these probes are measuring the difference between the two measurment points, not just one point and ground. Therefor instead of a center conductor and outer ground you want to have two similar conductors, making these a twisted pair is even better.

So basically, using a regular 100Mhz probe with a BNC to banana adapter would pretty much mess all the differential balance/symmetry?

[Serpent10i]

I'm not sure I understand your question, do you mean attaching a normal scope probe to the tips of the differential probes?

That would be a bad idea for several reasons, first off the scope probe would add a lot of extra cable in front of your measurement and that alone will add noise and lose signal fidelity, but lets say that you had a very very short one.

The specific coax cable used in scope probes center conductor has some resistance even on 1x mode (intentionally lossy, see Dave's video on 1x scope probes for more info), so your signals will be different between the outer conductor (normally ground) with nearly 0 resistance and the center with normally somewhere between a bit and a lot of resistance. They're also matched to the input of your specific scope, not matched for the input of whatever differential probe inputs you're using.

But lets assume that you're just using normal Coax cable with very low resistance and not matched for any specific scope input, this still isn't a great idea... As you suggested the wires being not being separately equal will also cause some signal issues, the outer shell will likely pickup more/different noise than the inner conductor leading to difference between signals and ones that will be measured/amplified by the differential probe.

tl;dr yes it's a problem, you should just use the banana cables with whatever tips you like

* BNC is the connector type, coaxial is the wire type with a single inner conductor and then an outer shell.

[tautech]

They use banana because they want to have two equal paths so that the signals are as similar as possible and they can reject as much common mode noise as possible.  Remember these probes are measuring the difference between the two measurment points, not just one point and ground. Therefor instead of a center conductor and outer ground you want to have two similar conductors, making these a twisted pair is even better.

So basically, using a regular 100Mhz probe with a BNC to banana adapter would pretty much mess all the differential balance/symmetry?

Yes, exactly.
If you need some sort of grabber with a differential probe, most are supplied with HV rated parrot beak grabbers.

[David Hess]

I'm not sure I understand your question, do you mean attaching a normal scope probe to the tips of the differential probes?

Early high voltage differential probes did exactly that with normal oscilloscope probes connected to a differential input vertical amplifier.

[rvalente]

I'm not sure I understand your question, do you mean attaching a normal scope probe to the tips of the differential probes?

I've asked because feels funny to me to imagine a 100MHz signal in a alligator clip with banana plugs, even if the cables are twisted.

But of course, your answer was spot on, tks mate