Current probes

Everything you need to start using them for EMC troubleshooting

by Ignacio de Mendizabal, 22/07/2024

​

Hello reader,

Glad to have you here with me, talking about electronics. Welcome.

If you are new with me, I am Ignacio, and I want to stop EMC being a problem. I want to make it so easy that people forget about it.

Look, I have never been a fan of the sentence “that is not my job”. I hate it. It is the opposite of the drive, resilience and hard work philosophy that I try to cultivate every single day.

As a Hardware designer, I need to perform EMC tests to start selling the products that my customers or I design. In the past, not knowing how the tests were performed drove me crazy. If there was something wrong with the measurements, I could not tell if the product being tested was failing the tests or if the person who was running that test for 28th time that week forgot something because of extremely boredom. When you pay 200 € per hour of test, you want to be effective.

I could say “It is not my job, I do not care”. But after seeing a product fail in two consecutive tests, in different laboratories, I could not leave it there. I decided to learn how the EMC tests were performed and why.

In today’s episode, you will learn about current probes, a critical device for pre-compliance and compliance conducted tests.

• What are current probes
• What to check when buying a new current probe
• How to double check the test-set up and be sure that you measure what you expect

What are current probes

Current probes are transformers, widely used in EMC to measure or inject current over cables

Why have at least one?

• They can be cheap
• The are small, portable and versatile
• They can provide accurate results, saving lab days (Remember the 200 € per hour)

They are used in two types of EMC tests:

• Conducted emissions: to measure what current circulates through the wires.
• Conducted immunity: to inject noise through the cables and see if the Device Under Testing will support it. Current probes are used in the substitution method, where two currents are used (one to inject noise, one to measure).

What to check when buying a new current probe

Not all the current probes are created equal and they should be aligned with your needs. Before buying the first probe that shows up in your browser, check the following

1. Bandwidth: upper limit but also lower limit. It is very disappointing when you unbox your new toy, you start playing with it, and you discover that you forgot to check some critical numbers (a friend of a friend told me once)
2. Transfer function: see how flat it is and see how you can compensate in your Spectrum Analyzer. If it is not possible, choose a different gear.
3. Calibration data: how was calibrated and when. What you read should be similar to what you get.
4. Size: thin cables for small applications or big cables for big power? Choose accordingly, the clips have a limit
5. Cost: how often will you use it? for who? Unless you are an accredited lab, don´t go crazy
6. Load capacity: To prevent errors due to saturation effects because of DC and AC currents (Thank you Marco A. Azpuru​a for this one)

How to double check the test-set up and be sure that you measure what you expect

When starting a new set of measurements, there are two situations: the ideal one, which we would love to have and the real one, which brings a roller-coaster of emotions.

Ideal

1. Install the test set-up
2. Power your Device Under Test (DUT) on
3. Take measurements and plot a graph. What you measure is what happens

Real situation

1. You miss some parts. Need to wait until they arrive.
2. The DUT is working half of the time. Is the Software running?
3. There are differences of 15 dB, bands where the measure drops.

Today, we focus on the last point. We take a deep breath and we go step by step. The first question is:

Is my current probe measuring what it is supposed to measure?

I have lost a lot of time due to messy setups or broken equipment. This made me become systematic every time I build a new test set-up.

Here is a trick: you can check the status of your current probe using S-parameters or Scattering parameters.

Why bother using S-parameters?

We can get a wideband analysis of the probe, knowing what happens in all the frequency range. Probes are not perfect in real-life, so you need to compensate for the differences from ideal behaviors.

They are non-intrusive: we do not need to break anything or change the system you are characterizing. This saves money and time in systems with very complex wire harnesses.

They are accessible. Today, we can access tools that were a dream some years ago. A lot of material is available worldwide, and this includes EMC tools. Here is a tip: look for demo equipment on sale. There are specific times of the year where they offer huge discounts because they buy new equipment and sell the old one.

S-parameters are a great tool, however, there are a few challenges with them:

They require Radio Frequency (RF) theory to know what exactly you are measuring. If you are not familiar with concepts such as impedance, Smith Charts or reflection coefficients, it will take you some time to understand what is going on. My hope is that my content helps you to soften that learning curve.

You get frequency domain measures but not time domain ones. While frequency measurements are the most crucial for EMC measurements, time domain ones are sometimes useful to measure transient phenomena.

Accessibility goes with the frequency range. The higher the frequency, the higher the price. Things stop to be cheap if you target a specific type of measures.

Characterizing a current probe

To see if our current probe is measuring what it should, we can build a small test set-up. We will use:

• A Vector Network Analyzer (VNA) or a Spectrum analyzer with the analyzer option enabled: Siglent SVA1015X
• A current probe: Tekbox TBCP1-150, with an operating range of [10 KHz, 250 MHz]
• Two coaxial cables
• An adaptor from coaxial to crocodile or two separate cables
• A 50 ohm resistor: 2 resistors of 100 in parallel.

The physical phenomena why we use current probes is:

1. An alternating current generates an alternating magnetic field
2. An alternating magnetic field over a magnetic material, generates an alternating current
3. An alternating current over a resistor, generates a voltage
4. We measure that voltage

Then, to characterize our current probe, we need to generate a controlled alternating current so we control what the current probe receives as an input.

How?

You got it. Generating an alternating voltage over a resistor

1. We connect the Tracking Generator (TG) output to a 50 ohm resistor. Since the connectors of an instrument like the analyzer are of a coaxial type, we need to “open” it. I built a small PCB to do it.

2. We install the current probe to measure the current circulating through the resistor. I did not have 50 ohm resistors available, so I used 2 resistors of 100 ohm.

Note: currents leave the source and go back to it, place the current probe around just one of the cables, not both of them.

3. Using the function of VNA, we measure the S21 parameter and we compare the results with the manufacturer data. The manufacturer provides a graph and a table with all the data, so we do not need to challenge our graphic gkills while trying to interpret the graphs. (Trick: chatGPT is a fan of recognizing text, give it the image and you will get a clean CSV file)

Then, here we have the two graphs: transfer impedance (manufacturer) and S21 (measured).

We need be careful with the difference of the frequency ranges. The manufacturer curve starts at 10 KHz, while the measurement with the VNA starts at 100 KHz due to the limitation of the Spectrum analyzer. It is not a problem because we do not need to measure that low-frequency currents and the current probe does not behave well at that low range. Then, I set the bottom limit for both curves to 100 KHz.

The good thing is that both curves follow a similar shape in a wide range, so the measurements are coherent with the calibration data. Sill, there are two open points:

1. The units shown by the spectrum analyzer are dB, while the units given by the manufacturer are dbOhm.
2. One magnitude is around 15, while the other one is around -15.

Let´s see how we can correlate and adjust both measurements:

1. A measurement in dBohm is a relative measurement of impedance. It tells you how far an impedance is from the nominal impedance. In RF measurements, this impedance is 50 ohm.

If we want to know the impedance in Ohms, we need to get Z

So now we know how to translate the manufacturer’s data in dBOhm to Ohms, we know the impedance of the current probe.

2. What we get from the Spectrum Analyzer are the S11 and S21 S-parameters, which help us to get valuable information

• The reflection coefficient (S11): it tells you how much energy is reflected back. Ideally, it should be 0, meaning that all the energy is transmitted and nothing is reflected.
• The transmission coefficient (S21): it tells you how much energy is transmitted through a device. Ideally, the transmitted energy should be 100% of the injected energy for the whole frequency range, so nothing is lost.

We are interested on the transmission of energy, so we focus on the S21 parameter.

First of all, the spectrum analyzer uses an internal attenuator of 15 dB to avoid an overload of the ADC and possible damages of the input frontend. It is not automatically compensated, so this means that the measurement is 15 dB below to its value.

Second, the S21 parameter provides the relationship between the output with the normalized impedance (50 Ω)

Applying both factors and plotting the two curves, we get:

​

Not bad, eh?

We can characterize our current probes!

I know what you are thinking: the curves are not the same

1. The data provided by the manufacturer is generic, not specific to the model I got. There can be differences from model to model.
2. In high frequencies, the test set-up is not the best, due to parasitic components of the resistors. We want to measure conducted emissions, happening in no more than some tenths of MHz. What happens that far is not a problem.

Summary

In this chapter, you learnt

• What the current probes are
• What to check when buying a new current probe
• How to verify that your current probe measures what is supposed to measure
• How to correlate measurements made with your Spectrum Analyzer with the manufacturer data

Current probes are fundamental for EMC troubleshooting.

I hope that you went one step further in the path for becoming an EMC hero.

Not all heroes wear capes, they bring current probes.

On the Shoulders of Giants

This article would not have existed without the inspiration of the work of Kenneth Wyatt and Bogdan Adamczyk about current probes.

​The HF Current Probe: Theory and Application - Kenneth Wyatt​

​Current Probe Measurements in EMC testing - Bogdan Adamczyk​

Share it or Subscribe

If you liked the content, share it!

In case you received it and you want to subscribe to future chapters

Subscribe to the Newsletter

If you have questions or inquiries, feel free to reply to this email.

I will wait you on the other side :)

You can find me on Linkedin and on the Website​

Best,

Ignacio