Impedance question matching/bridging

When I have a signal source with a Z_s=50Ω output impedance connected to a Z_c=50Ω coax cable connected to the Z_l=50Ω input impedance of my scope, In this case I understand that the power transferred to the load is 1/2 as per below illustration.



1. Does this Z_s=50Ω/Z_c=50Ω/Z_l=50Ω configuration mean that the voltage signal displayed on the scope is 1/2 of the source output?

2. If I wanted to use this Z_s=50Ω/Z_c=50Ω/Z_l=1MΩ configuration, would I need to terminate with a 50Ω resistor to ground at the load (as seen in the illustration lower right)? Would this be considered 'bridging' so that the displayed signal on the scope is a 1:1 representation of the source output?

3. If I wanted 10x attenuation of the source output signal, could I put a 10MΩ in series with the coax at the load? And should I put a 1.8pF capacitor across the 10MΩ to match the 1MΩ/18pF of the load?

Thank you.

1 + 3. Yes
2. That's no useful configuration. The displayed signal would be still half of the unloaded source voltage (at least if the cable length is short compared to the wavelength). But you have mismatch at both cable ends and are causing reflections.

The better configuration is 50 ohm source connected to 50 ohm cable and high impedance at the oscilloscope. This "source side termination" scheme works up to 100 or 150 MHz. Incident waves are reflected at the oscilloscope and absorbed in the load series termination. Due to non-perfect matching, you'll see small reflections in the low percent range after two times cable delay, e.g. after 10 ns for 1 m cable. For perfect signal quality, you'll prefer double side termination with 50 ohm oscilloscope input.

FvM said:

1 + 3. Yes

Click to expand...

Thank you, duly noted. In #3 where/how do I mount the 10MΩ/1.8pF? In a small metal enclosure with two BNC's and then connect F-BNC directly to the scopes input without any cable in-between?

FvM said:

2. That's no useful configuration. The displayed signal would be still half of the unloaded source voltage (at least if the cable length is short compared to the wavelength). But you have mismatch at both cable ends and are causing reflections.

The better configuration is 50 ohm source connected to 50 ohm cable and high impedance at the oscilloscope. This "source side termination" scheme works up to 100 or 150 MHz. Incident waves are reflected at the oscilloscope and absorbed in the load series termination. Due to non-perfect matching, you'll see small reflections in the low percent range after two times cable delay, e.g. after 10 ns for 1 m cable.

Click to expand...

I thought I already proposed the "better configuration" by saying "Zs=50Ω/Zc=50Ω/Zl=1MΩ" where Zl is the Hi-Z 1MΩ/18pF input of the scope? Or maybe I don't understand it quite right.

In the "better configuration" is displayed signal on the scope a 1:1-ish representation of the source output?

FvM said:

For perfect signal quality, you'll prefer double side termination with 50 ohm oscilloscope input.

Click to expand...

Yes, but the problem is that the scope is max 5Vp-p with 50Ω while the signal can reach up to 500Vp-p, so I need some sort of attenuation.

Tried an attenuator on the source end?

"Back when I had a real job" department, it was common to make a test port, on the 70 MHz IF strips we used. Everything had to be matched, impedance-wise, but you didn't want to load down the circuit when the test port was used. A large enough series-R, connected to a shunt-R, across the test port, was employed to get 20 (or 40) dB of isolation.

The higher levels of attenuation also had the advantage of minimizing the effect of connecting anything to the test port, as the Z would change less than with 20 dB of attenuation.

So, if you have a 500 V source, try either a 100:1 or 1000:1 attenuator. Use the one that gives the best results. You still have matching, at both ends, as I assume this is necessary to get good results.

I thought I already proposed the "better configuration" by saying "Zs=50Ω/Zc=50Ω/Zl=1MΩ" where Zl is the Hi-Z 1MΩ/18pF input of the scope? Or maybe I don't understand it quite right.

Click to expand...

You proposed to add 50 ohm load, I disagreed.

Yes, but the problem is that the scope is max 5Vp-p with 50Ω while the signal can reach up to 500Vp-p, so I need some sort of attenuation.

Click to expand...

O.k., a new topic. With 500V you'll also think about power dissipation and most likely don't want 50 ohm matching.