Characteristic impedance of oscilloscope coaxial cable

[Zaphappy]

I am trying to determine some parameters for this cable. Capacitance is about 114pF (7.6pF/FT) and 277 ohm center conductor (18.5/FT).

My inductance bridge with 1 end shorted says about 0.45uH - So Z should be about 62.3 ohms.

If I use the cable delay to determine inductance it works out to be 4.6uH L=(Td x Td)/C) Td = time delay.

More like 200 ohm Z

I only have one sample so I can't destroy it just yet to measure diameters and I don't know the dielectric.

My question is would the high series resistance of the coax give measurement errors in either the bridge or the time delay ?

I tried modeling in SPICE. The simple 3 component model with the calculated 4.6uH seems to give me the same observed 8MHz peak resonance that I observed in actual sweep but when I model a 10-20 element lumped model it does not do the same thing...

I am confused - Treat this as a transmission line or low pass filter ?

 

[chuckey]

The reason for the high resistance is to stop the cable ringing when it is handling fast transitions because it is "terminated" in one megohm!!!
Frank

 

[FvM]

Some explanations are in this previous thread: https://www.edaboard.com/threads/181660/

Please notice the linked Tektronix publication Oscilloscope Probe Circuits. It has a detailed discussion of the resistive coaxial cable concept. Usual probe cables are around 200 ohm characteristic impedance by the way.

 

[Zaphappy]

Thank you FvM - My final calculation using the cable delay came out to 193 ohms so I think I'm in the ball park. I realized that my spice model may be flawed though. I used the model in this document http://www.hpl.hp.com/hpjournal/96apr/apr96a11.pdf and the results were mush closer.

I still don't understand how these scope probes have such high bandwidth when the RC roll off of the cable is below 10MHz. The spice model does not seem to
"respond" to a capacitively coupled termination impedance like the actual cable does. Doug Ford mentions this in his fine article on scope probes.

http://www.dfad.com.au/links/THE SECRET WORLD OF PROBES OCt09.pdf

 

[FvM]

I still don't understand how these scope probes have such high bandwidth when the RC roll off of the cable is below 10MHz.

The RC is distributed and doesn't count as 10 MHz low-pass. Further more the C is part of the 10:1 attenuator network, as a result you get 300 to 500 MHz Bandwidth depending on the design details.

 

[jiripolivka]

For the low-frequency use of oscilloscope cables (below ~1 MHz), the probes utilize ~200-Ohm cables and as oscilloscope input impedance is 1 MOhm with ~20-50 pF parallel, there is a compensation trimmer to adjust a good pulse shape.

For frequencies above 1 MHz, and generally above 10 MHz, oscilloscope input must be 50 Ohms and the probe must use a 50-Ohm cable. No compensation is necessary. The probe must be treated carefully to keep the "live" and mainly the ground connection short, and caution be taken not to contact points with a DC voltage (this often burns the 50-Ohm resistor at the scope input).

 

[FvM]

For frequencies above 1 MHz, and generally above 10 MHz, oscilloscope input must be 50 Ohms and the probe must use a 50-Ohm cable.

I would immediately agree if you say "for frequencies above 500 MHz". It' also right that above several 10 MHz, a 10:1 resistive probe involves less circuit loading than a 10:1 passive high impedance probe.

The passive 10:1 probes of my oscilloscopes have nevertheless a bandwidth of 500 MHz. The bandwidth that can be utilized in a particular measurement setup depends, of course. It's e.g. 265 MHz for 50 ohm source impedance.

 

[jiripolivka]

I would immediately agree if you say "for frequencies above 500 MHz". It' also right that above several 10 MHz, a 10:1 resistive probe involves less circuit loading than a 10:1 passive high impedance probe.

The passive 10:1 probes of my oscilloscopes have nevertheless a bandwidth of 500 MHz. The bandwidth that can be utilized in a particular measurement setup depends, of course. It's e.g. 265 MHz for 50 ohm source impedance.

You are right but everything depends on two principal points:

- tested-circuit loading : at below ~10 MHz, scope input impedance and the resistive 10:1 divider
is often optimal but still the mismatch forces the user to adjust the pulse response by the trimmer on probe tail. Above ~20-50 MHz, circuits cannot be contacted without loading them, probes use small capacitors (< 1 pF) or coils...

- Impedance matching - like above, the higher the frequency, the stronger effect of a mismatch.
Amplifier probes often do nothing more than impedance matching; but they also introduce some distortion...

 

[Zaphappy]

Well so far I can't make this "probe" flat past 8 or 9MHz.. Perhaps the series resistance (277 OHM) is too high. Critically damped ? I have not inserted any inductance yet as suggested by the excellent document FvM linked - Great stuff thanks.. I'm uneasy about a discreet inductor as I need VERY flat response and I feel the response may wander a bit..