Accurately measure voltages across loop antennas

Hello antenna experts,

I am trying to measure the peak of ac voltages across loop antennas. The more accurate, the better. I wonder what equipment I should use? Will a digital oscilloscope suffice? The voltage could be lower than ~100mV. And also, I'd also like to observe the phase difference between two voltages.

Please advice some test equipment that can fulfill both purposes. Thanks in advance.

Hello,

What do you want to measure, the open loop voltage?

What is your frequency of interrest and what is the expected loop inductance and wire length?

I assume you want to measure the loop voltage to calculate magnetic field strength.

Starting from 10 - 20 MHz, you'll want a high impedance buffer (e.g. an active oscilloscope probe) to measure the open loop voltage without disturbing it by the cable capacitance load. Of course the loop size and number of turns matters.

Thank you guys.

Just to give some info in case you have anything to add. Operation frequency: 10M to 2 GHz. wire length: 8 cm (2cm x 2cm rectangle loop). Inductance: on the order of nH. And WimRFP, you're right. I am attempting to calculate the field strength, mutual inductance between two loops, etc. FvM: thank you for th advice. I think that makes lots of sense.

You need a tracking generator? tracking filter? Otherwise how much sensitivity is needed.

This is what Network Analyzers do best or Spectrum Analyzer with tracking generator.

I would measure using the usual instruments (network analyzer or signal generator + spectrum analyzer/power meter) with some high impedance (kOhm) voltage divider at the loop antenna. This gives you (relatively) high impedance to measure the voltage. Calculating the attenuation caused by the voltage divider to the 50 ohm input is easy.

A single turn loop gives the lowest inductance (but also relative low output voltage). Depending on thickness of wire or strip, inductance will be in the 50 nH range.

When you know the inductance and the load impedance, you can determine the EMF of the coil based on the loop's impedance and (50 Ohms) load. The –3 dB corner frequency will be in the 150 MHz range.

You may add a balun function between loop and coaxial cable to get better rejection from E-field components. Because of this balance, the wire length halves increasing the useful frequency range where transmission line effects can be omitted.

Instead of a separate balun function, you may construct a loop with natural balun (like the coaxial loop with cut in the shield opposite to the feed point). Regarding balun function, a 0.01 to 2 GHz passive balun with high input impedance is hard to impossible to design.

If you really need high input impedance, you can insert a very small (chip) unbalanced amplifier into a loop with natural balun function (and feed it via the coaxial cable). The natural balun function makes the amplifier to float. Depending on input capacitance, you may need resistive loading to avoid resonance. This route will introduce more inaccuracy with respect to a passive approach. Of course you can go for a passive attenuator to reduce the load to the coil, but this also requries you to know its attenuation.

Can you explain the two voltage issue?

Assuming a balanced setup, the wire length is 0.02m and for 2 GHz lambda = 0.15m. you will have some deviation from simple induced voltage math due to transmission line effect. If you have access to an EM simulator, you may simulate the loop with 50 Ohms load and driven with a plane wave field.

If you measure close to structures that may affect the inductance, (for example close to a ground plane or tuned circuit), loading the coil will affect accuracy.

If you are unsure whether the input VSWR of your measuring instrument (spectrum analyzer, measurement receiver, oscilloscope, etc) is low, you may insert 3..6 dB attenuation so that the loop "sees" 50 Ohms (or other known reference impedance).

If you are in the far field of your structure (for the higher part of the frequency range), you may use simple dipole or other antenna with known gain. From the output of the antenna you can calculate the plane wave power, hence the magnetic field component.

grit_fire said:

Thank you guys.

Just to give some info in case you have anything to add. Operation frequency: 10M to 2 GHz. wire length: 8 cm (2cm x 2cm rectangle loop). Inductance: on the order of nH. And WimRFP, you're right. I am attempting to calculate the field strength, mutual inductance between two loops, etc. FvM: thank you for th advice. I think that makes lots of sense.

Click to expand...

It is highly unlikely that you will be able to get a loop antenna to work over that wide range of frequencies. There are 2 problems: 1) how do you get elements of the antenna to resonate at all those frequencies without nulls, and 2) how do you couple into the antenna.

Here is an example of the complexity needed to make such an antenna:
https://www.ahsystems.com/catalog/SAS-521F-2.php

As far as measuring, AFTER you have chosen a method to couple to the antenna, you would use a spectrum analyzer to measure output power and convert that to voltage.

The "loop antenna" is as an electrical small magnectical antenna or "H-sensor" with dimensions << λ. It can work with a sufficient high load impedance, as suggested. If high sensitivity isn't required, the passive voltage divider suggested by Volker is the easiest way to get frequency independent response.

Another option is to make the loop considerably smaller and connect it to a 50 ohm load directly. That's the typical design of a EMI H-sniffer.

Yes, either very high load impedance, or very low load impedance, so it would be a non-resonant loop. I tried that once on a VHF project, but was not too happy about performance.

But to make it work at 2 GHz, the loop will be so small that it will be extremely inefficient at 10 MHz.

But to make it work at 2 GHz, the loop will be so small that it will be extremely inefficient at 10 MHz.

Click to expand...

Of course, a H-sensor is far from an impedance matched antenna, the same with the E-sensors used for field strength calibration in an EMI lab.

biff44 said:

so it would be a non-resonant loop.

Click to expand...

Looking at existing H field probes, there are different sizes to cover the range up to 3GHz.
**broken link removed**

The smallest H-probe there ("1 cm") is specified with 2.3GHz resonance frequency.

volker_muehlhaus said:

I would measure using the usual instruments (network analyzer or signal generator + spectrum analyzer/power meter) with some high impedance (kOhm) voltage divider at the loop antenna. This gives you (relatively) high impedance to measure the voltage. Calculating the attenuation caused by the voltage divider to the 50 ohm input is easy.

Click to expand...

Thank you. Do you assume the input impedance to the antenna is 50 Ohm? Or you meant the input to the matching network+antenna which will indeed be 50 Ohm, but in that case, the voltage measured would at the input of the matching network instead of the loop itself. Isn't it? Or did I misunderstand anything？

- - - Updated - - -

WimRFP said:

A single turn loop gives the lowest inductance (but also relative low output voltage). Depending on thickness of wire or strip, inductance will be in the 50 nH range.

When you know the inductance and the load impedance, you can determine the EMF of the coil based on the loop's impedance and (50 Ohms) load. The –3 dB corner frequency will be in the 150 MHz range.

You may add a balun function between loop and coaxial cable to get better rejection from E-field components. Because of this balance, the wire length halves increasing the useful frequency range where transmission line effects can be omitted.

Instead of a separate balun function, you may construct a loop with natural balun (like the coaxial loop with cut in the shield opposite to the feed point). Regarding balun function, a 0.01 to 2 GHz passive balun with high input impedance is hard to impossible to design.

If you really need high input impedance, you can insert a very small (chip) unbalanced amplifier into a loop with natural balun function (and feed it via the coaxial cable). The natural balun function makes the amplifier to float. Depending on input capacitance, you may need resistive loading to avoid resonance. This route will introduce more inaccuracy with respect to a passive approach. Of course you can go for a passive attenuator to reduce the load to the coil, but this also requries you to know its attenuation.

Can you explain the two voltage issue?

Assuming a balanced setup, the wire length is 0.02m and for 2 GHz lambda = 0.15m. you will have some deviation from simple induced voltage math due to transmission line effect. If you have access to an EM simulator, you may simulate the loop with 50 Ohms load and driven with a plane wave field.

If you measure close to structures that may affect the inductance, (for example close to a ground plane or tuned circuit), loading the coil will affect accuracy.

If you are unsure whether the input VSWR of your measuring instrument (spectrum analyzer, measurement receiver, oscilloscope, etc) is low, you may insert 3..6 dB attenuation so that the loop "sees" 50 Ohms (or other known reference impedance).

If you are in the far field of your structure (for the higher part of the frequency range), you may use simple dipole or other antenna with known gain. From the output of the antenna you can calculate the plane wave power, hence the magnetic field component.

Click to expand...

Thank you very much.

First, let me explain the two voltage issue you mentioned. I have 1 TX coil/antenna and 1 RX coil/antenna. In fact, I only want to measure the voltage at the TX coil (V1). sorry I wasn't clear about that before. Then I have V1=jw*L1*I1+jw*M*I2...(1). V2=Jw*M*I1+jw*L2*I2........(2) and V2=I2*Z2..........(3) The M, V2, I2 will be the unknown once I measure V1. After that, I can solve for M. Hope that clarify something.

I am not too sure about the bold part I highlighted. When you know the inductance and the load impedance, you can determine the EMF of the coil based on the loop's impedance and (50 Ohms) load. The –3 dB corner frequency will be in the 150 MHz range. --> did you refer to the RX loop side?

And why would I want very high input impedance? you meant the input impedance looking into which component? Sorry that I am not understanding well on this. Why is high input impedance desirable here?

Thank you!

- - - Updated - - -

biff44 said:

It is highly unlikely that you will be able to get a loop antenna to work over that wide range of frequencies. There are 2 problems: 1) how do you get elements of the antenna to resonate at all those frequencies without nulls, and 2) how do you couple into the antenna.

Here is an example of the complexity needed to make such an antenna:
https://www.ahsystems.com/catalog/SAS-521F-2.php

As far as measuring, AFTER you have chosen a method to couple to the antenna, you would use a spectrum analyzer to measure output power and convert that to voltage.

Click to expand...

Thanks a lot. In fact, it was misleading for me to say the frequency range. My apology. The antenna will ultimately work around 2 GHz. I am conducting a series of experiments with different PCB working under different frequencies, including 10 MHz, 100 MHz and so forth till 2 GHz. So I didnt mean that one antenna should work across a very wide band.

- - - Updated - - -

You guys' posts are really helpful and rich in knowledge. I appreciate that. Btw, I am trying to avoid the spectrum or network analyzers for my experiment. Only the voltage measurement will be more desirable.

ωAn electrical small loop has an almost inductive impedance, surely not 50 ohm, and it can't be operated impedance matched except for very small bandwidths. For a frequency independent H measurement, it must be terminated with Z >> ωLant.

The connected measurement equipment and cables have 50 ohm impedance.

I would consider a wide band splitter or return loss bridge to measure the S11 levels on the antenna (Input Return Loss) with a hot-carrier diode to detect the levels into a DMM. That's all I used on my 1st VHF antenna design. it was good enough for transmitting on one bldg roof top and measuring on another for S12 response.

Terminating the sum port with the antenna turns it into a return loss bridge with HC diode and small RF cap added on either port.

Using high impedance load to the coil (that is Zload >> ZL = 2*pi*f*L), the voltage measured equals the open loop voltage. This results in relative simple math. When loading with 50 Ohms, you need to correct the measured voltage because of the 50 Ohms load.

If you don't like to do the full match for inductors that are loaded with certain impedance, there is a lazy trick.

In pspice you can use "inductor coupling" where you can set the coupling factor "k" You may know inductance M and k are related. When you run the simulation and change k until the simulation is same as measurement, you can calculate M. Of course you need to know the inductance of both loops.

So to determine M, you can drive from a 50 Ohms source and load the other inductor with the 50 OHms input impedance from the analyzer.

Other way to determine M is by using Lshort/Lopen = 1-k^2 (this formula is known from transformers and related to the leakage inductance). Lshort is the inductance of the first coil when the second is short circuited. If you have a vector network analyzer, you can get reasonable results, even with relative low values of k. The problem with the short circuited coil measurement is that when k is very low, the change in inductance may be too small to measure.

There are more ways to determine the change in inductance. You may resonante the coil with a capacitor and determine the dip with a scalar measurement instrument (for example the diode probe as mentioned by SunnySkyGuy). Short circuit the second coil and measure the shift in resonant frequency. From there you can calculate the relative change in inductance, k and M.

Up to some 100 MHz you can ignore any transmission line effect when using coils with total 0.08m wire length.