How to use a Network Analyzer With Current Probes?

[SniperFox]

Hi everyone,

I'm wondering how to set up current probes to use it with a network analyzer. I don't know how to calibrate them properly for short, open, and 50ohm terminations (oh and through, too. SOLT). It's straight forward with BNC or SMA connections but what about current probes?

What I tried:
Short -> I made a small wire loop.
Open -> left it open.
Load -> I made a small wire loop with a 50ohm resistor.
Through -> I Put a small wire loop between both probes.

This didn't seem to work, as the trace has lots of spikes and is bouncing up and down all over the place... so I'm seeing nothing but noise when I try to take measurements.
I'm trying to measure S11 and S21 btw.


What I'm trying to do is use a network analyzer with current probes to measure the impedance of a LISN and SMPS power supply. This is  step towards systematically designing a conducted EMI filter.
I found this neat scholarly article explaining how to measure those impedances under operating conditions with clamp-on current probes. See link below.
https://www.researchgate.net/publication/260589392_EMI_filter_design_Part_II_Measurement_of_noise_source_impedances

If you scroll down a bit, you can view the article without downloading it.

They explain everything very well and derive the equations well too. The problem is that they don't explain how to calibrate and use the VNA with the current probes.
They also don't say what probes or VNA they used. Although I am 90% sure that they used CT1 or CT2 probes because a few of their references use those probes.

If anyone wants details on the instruments I'm using: I'm using E5061B VNA and two P6021 current probes.

(I tried searching the forums for this topic but could not find anything)

[PartialDischarge]

You don’t calibrate the current probe, the current probe is what you are trying to measure.
Calibrate the source of the instrument (thru calibration) terminating it with an external 50Ohm load, and connect it to  the receiving port terminated in 1MOhms.
After that you can add the current probe tied to the source, again with the source terminated in 50Ohms, and the measuring port to 1Mohm if you are using the p6021 termination


At 10dbm output you shold get 20mApk measured, or -46dbm in the 2mA/mV scale

[SniperFox]

You don’t calibrate the current probe, the current probe is what you are trying to measure.
Calibrate the source of the instrument (thru calibration) terminating it with an external 50Ohm load, and connect it to  the receiving port terminated in 1MOhms.
After that you can add the current probe tied to the source, again with the source terminated in 50Ohms, and the measuring port to 1Mohm if you are using the p6021 termination


At 10dbm output you shold get 20mApk measured, or -46dbm in the 2mA/mV scale

Thanks for the reply, MasterTech!
Sorry, but I'm not following you well... Could you clarify a few things I don't understand?

1. When you say I don't calibrate the probe, what do you mean the probe is what I'm trying to measure? I thought you're supposed to "calibrate out" the probe (and cable) so that it won't get in the way of the measurements whenever you use a VNA.

2. When you say calibrate the "source" of the instrument: You mean the injecting port?

3. "terminating it with an external 50Ohm load, and connect it to  the receiving port terminated in 1MOhms." - could you elaborate on this? I don't understand

4. "After that you can add the current probe tied to the source, again with the source terminated in 50Ohms, and the measuring port to 1Mohm if you are using the p6021 termination" - When you say source terminated in 50 ohms.. This is a setting on the VNA, correct? The input/output impedance of each port, right? If so, then I don't think the E5061 has a 1Mohm option. It only has 50Ohms...

Thanks a bunch!

[PartialDischarge]

1. When you say I don't calibrate the probe, what do you mean the probe is what I'm trying to measure? I thought you're supposed to "calibrate out" the probe (and cable) so that it won't get in the way of the measurements whenever you use a VNA.

Are you sure? So if you had a filter instead of a probe, would you calibrate out the filter so its response would show up as flat?
What you calibrate is the source of the current vs the input port, so that this response is flat. Since this application of yours is for less than 100MHz forget about the cable. But the response of the probe is what it is and don't let the VNA meddle with it.

2. When you say calibrate the "source" of the instrument: You mean the injecting port?

Yes, the port that generates energy

3. "terminating it with an external 50Ohm load, and connect it to  the receiving port terminated in 1MOhms." - could you elaborate on this? I don't understand

You want to measure current right? So how do you generate a current? By loading the source with a 50Ohm load (external load). Then you need to measure this current (the voltage at the output loaded node) at the input in the same conditions that you would if you had the probe attached, that is with a 1MOhm input.

4. "After that you can add the current probe tied to the source, again with the source terminated in 50Ohms, and the measuring port to 1Mohm if you are using the p6021 termination" - When you say source terminated in 50 ohms.. This is a setting on the VNA, correct? The input/output impedance of each port, right? If so, then I don't think the E5061 has a 1Mohm option. It only has 50Ohms...

The source is always 50Ohm but it needs additional and external 50Ohm termination (to generate a current). The source impedance is one thing, and the source termination is another thing.
The P6021 with termination box is to be used on 1MOhm terminated devices, you can't use it on a 50Ohm input device. Here you can do a few things if you VNA only has 50Ohm input port:

- Use a VNA that has 1MOhm input: HP 3589A, MS4630b
- Use a current probe that has 50Ohm output: AM503
- Use an amplifier between the P6021 and the VNA, that has 1MOhm input and 50Ohm output, this is easy to breadboard.

[SniperFox]

1. When you say I don't calibrate the probe, what do you mean the probe is what I'm trying to measure? I thought you're supposed to "calibrate out" the probe (and cable) so that it won't get in the way of the measurements whenever you use a VNA.

Are you sure? So if you had a filter instead of a probe, would you calibrate out the filter so its response would show up as flat?
What you calibrate is the source of the current vs the input port, so that this response is flat. Since this application of yours is for less than 100MHz forget about the cable. But the response of the probe is what it is and don't let the VNA meddle with it.

No, I'm not sure..
About 2 weeks ago I didn't even know what a network analyzer was, so I am very new to this equipment. I still don't fully understand the what the calibrations methods do. So if we had a filter, I have no idea...

3. "terminating it with an external 50Ohm load, and connect it to  the receiving port terminated in 1MOhms." - could you elaborate on this? I don't understand

You want to measure current right? So how do you generate a current? By loading the source with a 50Ohm load (external load). Then you need to measure this current (the voltage at the output loaded node) at the input in the same conditions that you would if you had the probe attached, that is with a 1MOhm input.

If I understand this correctly:
Let's say I used current probes that have a 50ohm output instead of 1Mohm. Is this how would I do this step? (Also, I'm going to call port 1 of the VNA the source, and port 2 of the VNA the input from now on)
1. Before connecting any current probes to either of the ports, I first attach a 50ohm load directly to port 1
2. Then I connect a voltage probe to port 2 of the VNA, and attach that probe across the 50ohm resistor and click thru calibration?

...What about open, short, load, and isolation? Do I have to do those?

[james_s]

It would help if you could say what exactly you are trying to measure here. What is the end goal? Do you want to characterize the current probe itself? VNAs are usually used to characterize things like antennas and filters.

[PartialDischarge]

No, I'm not sure..
About 2 weeks ago I didn't even know what a network analyzer was, so I am very new to this equipment. I still don't fully understand the what the calibrations methods do. So if we had a filter, I have no idea...

You need to learn more about RF, about how VNAs work, and play with them with something easier like a low pass passive RLC filter, play with the parameters and understand everything.

[PartialDischarge]

If I understand this correctly:
Let's say I used current probes that have a 50ohm output instead of 1Mohm. Is this how would I do this step? (Also, I'm going to call port 1 of the VNA the source, and port 2 of the VNA the input from now on)
1. Before connecting any current probes to either of the ports, I first attach a 50ohm load directly to port 1
2. Then I connect a voltage probe to port 2 of the VNA, and attach that probe across the 50ohm resistor and click thru calibration?

...What about open, short, load, and isolation? Do I have to do those?

Not a voltage probe, you connect directly port2 to the loaded port1, then press "cal" that is why it is a thru calibration. But this would be for a 1Mohm input port, since your port 2 is 50Ohm, you would connect directly port 1 to port2, the 50Ohm at port 2 would be the "load" that you would later add externally.
 I don't know your equipment, but in many VNA a thru is enough, although an open is as easy as leave the ports unconnected

[SniperFox]

If I understand this correctly:
Let's say I used current probes that have a 50ohm output instead of 1Mohm. Is this how would I do this step? (Also, I'm going to call port 1 of the VNA the source, and port 2 of the VNA the input from now on)
1. Before connecting any current probes to either of the ports, I first attach a 50ohm load directly to port 1
2. Then I connect a voltage probe to port 2 of the VNA, and attach that probe across the 50ohm resistor and click thru calibration?

...What about open, short, load, and isolation? Do I have to do those?

Not a voltage probe, you connect directly port2 to the loaded port1, then press "cal" that is why it is a thru calibration. But this would be for a 1Mohm input port, since your port 2 is 50Ohm, you would connect directly port 1 to port2, the 50Ohm at port 2 would be the "load" that you would later add externally.
 I don't know your equipment, but in many VNA a thru is enough, although an open is as easy as leave the ports unconnected

Okay, I think I get it. Thank you so much!

[SniperFox]

It would help if you could say what exactly you are trying to measure here. What is the end goal? Do you want to characterize the current probe itself? VNAs are usually used to characterize things like antennas and filters.

Hi James, I can't really summarize what exactly I'm trying to do here, so instead I posted the scholarly article in the OP if you have time to look at it. I'm trying to do what they are doing in the article.

[Someone]

The correct answers are already in the duplicate thread:
https://www.eevblog.com/forum/projects/how-to-use-a-network-analyzer-with-current-probes-230008/
A P6021 (and other typical current probes) are one directional (non-reciprocal) and break the assumptions of a full two port network calibration.

[chris_leyson]

You could make a current transformer with a suitable RF toroid. For example 10 turns would give you a 10:1 transformer. Terminate the 10 turn current sense winding with a 50 Ohm load and the through or single turn winding would see 50/N^2 or 0.5 Ohm.

[SniperFox]

The correct answers are already in the duplicate thread:
https://www.eevblog.com/forum/projects/how-to-use-a-network-analyzer-with-current-probes-230008/
A P6021 (and other typical current probes) are one directional (non-reciprocal) and break the assumptions of a full two port network calibration.

I created 2 threads because I wasn't sure which one this topic would belong to.

I think MasterTech is saying something slightly different here than graybeard is saying in the other thread.. I guess I'm going to try both approaches.

[james_s]

If you post something in the wrong area it can be moved. You should never create multiple threads on the same topic, that creates confusion and tends to piss people off, basic forum etiquette that applies almost anywhere.

[graybeard]

You don’t calibrate the current probe, the current probe is what you are trying to measure.
Calibrate the source of the instrument (thru calibration) terminating it with an external 50Ohm load, and connect it to  the receiving port terminated in 1MOhms.
After that you can add the current probe tied to the source, again with the source terminated in 50Ohms, and the measuring port to 1Mohm if you are using the p6021 termination


At 10dbm output you shold get 20mApk measured, or -46dbm in the 2mA/mV scale

The calibration method is implicitly described in the paper.  The secondary calibration proceedure incorporates the current probes characteristics.

You first need to establish your S parameter calibration planes at the current probes as shown in Figure 2.  If you are using the typical 50 Ohm Tek CT1 or CT2 probes you will need to make a set of  female SMB calibrations; short, open, load, and through to place the s parameter calibration planes as close as possible to the current probes.

You then need to make two secondary calibration measurements with the current probes; one with a resistance standard whose value is near the impedance you expect to measure, and one with a short.  You first get the ratio of the probe voltages at every frequency using equation 9 for both measurements; R_STD and short.  Substitute that result into two equations  10 & 11 and solve for the complex quantities K and Z_setup at each frequency point.  It is this secondary calibration that accounts for the current probes.

Since you now have K and Z_setup you can measure your unknown impedance and using equation 8 calculate the unknown impedance at each frequency point.

[PartialDischarge]

Hi graybeard, from his original post it gave me the impression that he was kind of lost in how to setup something simple in the VNA, so I just guided him to make an easy S21 current measurement of a device. It is at least important to understand matching impedances.