Questions about superposition of two antennas

[ffjonas]

Hey all,

currently I am working on a project, which uses the principle of superposition of two antennas. I placed two antennas on the plane and make sure that they are oriented identically. The distance between these two antennas is around 51cm. I did an experiment, in which I firstly get the magnetic field strength of the first antenna at x, y and z direction, then the magnetic field strength of the second antenna. In the end, I activated both antennas and obtained a superposed magnetic field strength. According to the theory, the superposed magnetic field should be the sum or difference of that of two antennas separately at x, y and z direction.

What I have observed is very strange. At certain positions and certain direction, for example, x direction, the values match each other. But at other positions, the values differ widely.

Is there someone who know the reason behind it and how I could improve the experiment? :)

Thanks in advance!

Fan

 

[skeptic2]

What I have observed is very strange. At certain positions and certain direction, for example, x direction, the values match each other. But at other positions, the values differ widely.

Is there someone who know the reason behind it and how I could improve the experiment? :)

Thanks in advance!

Fan

I don't understand what you are comparing above. Are you saying that when powering one antenna at a time, in some directions the magnetic field of one antenna doesn't match the magnetic field of the other, or

are you saying that when you compare the combined magnetic fields of both antennas to the magnetic field of a single antenna, the values differ widely?

What frequency are you feeding to your antennas? How are you measuring the magnetic fields?

 

[ffjonas]

:)

I have a device which can detect magnetic field generated by the two antennas. I was trying to compare the combined magnetic fields of both antennas with the magnetic field of a single antenna. So the combined magnetic field at the third step is supposed to be the vector sum of the magnetic field generated by single antennas at first and second step. That's what I am comparing.

The frequency of antennas is 125KHz.

 

[skeptic2]

What are you using for an antenna? A ferrite rod? If so, how are they oriented? Are they all parallel to each other or not? Does the length dimension of the antennas line up with of of the measurement axies or oare they offset by some angle? How far is your measuring antenna from the other two when you are making a measurement? Is it 51 cm or less or significantly greater than 51 cm? A diagram would help here.

I suspect what you are seeing, without having seen a diagram, is not variations in the magnetic fields of the antennas but variations in the coupling coefficients between the antennas and your measurement antenna.

 

[ffjonas]

Hey,

The two transmitter antennas that I am using are LF antenna (125KHz). They are directional antenna: just wind turns of wire on the ferrite rod and then connect a capacitor to tune it to desired frequency. The diagram is shown in the attachment.

Steps:
1)Activate A1 (the 1st antenna). AR measures magnetic field generated by A1: (Mx1, My1, Mz1). The magnetic field strength is presented as a set of magnetic fields in x, y and z directions.

2)Ativate A2 (the 2nd antenna). AR measures magnetic field generated by A2: (Mx2, My2, Mz2). The magnetic field strength is presented as a set of magnetic fields in x, y and z directions.

3)Activate A1 (the 1st antenna) and A2 (the 2nd antenna). AR measures magnetic field generated by both antennas: (Mx, My, Mz).

In the end, check if
•Mx = Mx1 +Mx2 ?
•My = My1 +My2 ?
•Mz = Mz1 +Mz2 ?

 

[ffjonas]

I did the experiment again and the values still do not match. What if the mismatch is caused by coupling between two antennas, what should I do to alleviate this undesirable affect?

Thanks!

 

[sophiecentaur]

You could separate the transmit antennas more. They would couple a lot less and you could, then, just see the superposition. I think you have to approach this assuming that superposition is real and it is the measurement that's going wrong.
I wonder if you have tried this over a band of frequencies. There may be some resonances which could be increasing the effect of mutual impedance between the two antennae.
Another thing: when just one antenna is powered, do you terminate it or just pull the plug out?
I am clutching at straws here but leave no stone unturned.

 

[ffjonas]

Hey, thanks very much for your reply.

I just an experiment, in which I separate two transmit antennas further (90cm). At some positions, it does not work well. At these positions there is a quite big difference between the two values I want to compare.

As a matter of fact, I can not change the transmitting frequency. It is fixed. The receiver antenna is a 3D coils on a transponder. There is some mental on the transponder as well. I am not sure if this affects the mesurement of magnetic field strength at different directions (x, y and z).

To prove that superposition of two antennas works is very important for me to proceed in my experiment because it is one of the key criteria.

 

[ffjonas]

Plus, when just one antenna is powered, the other one is terminated. And I also checked in the scope, which shows that two frames do not overlay.

There is 1s break between every activation. :)

 

[sophiecentaur]

Can you monitor the volts on the feeders whilst the other transmitter is switched on and off? Are there, in fact, two transmitters or just one with a splitter? What do you measure at the feed point of one of the two transmit antennae when the other is on (phase and amplitude)? Have you tried an attenuator in each of the feeds to increase isolation between the two sources? There could easily be some interaction between the two sources and you'd need to reduce it as much as possible.

 

[ffjonas]

Hey, because the two antennas are connected to a PCB board and driven by the board, I can monitor the voltage of two output signals related to these two antennas. They are connected to the same data which needs to be transmitted. But they can be controlled via software to activate one or the other or both.

 

[sophiecentaur]

I had another thought about this. If the problem is to do with coupling then you could use polarisation to isolate the transmit antennae from each other. Have them mounted at right angles to each other and look, with your search antenna, at the 40 degree component of each. But you could tell all this with a scope, looking at the outputs from the board and waggling the transmit antennae around or turning them on and off.
Just had another thought. These are magnetic (coils) and the field pattern is not what you'd get from an isotropic radiator. The direction of the H field will change by almost 180 degrees as you go round the antenna (if the axis of the coil is in the plane of the paper in the diagram). Think of the pattern of a bar magnet. The fields will not be parallel at all points in space so they will not 'just add up'. It's a vector addition. Sorry if you've already taken this into account but it could be very relevant if you haven't. To sort it our, the axes of the coils would need to be at right angles to the plane of the paper (i.e. the coil turns should lie parallel to the plane of the paper with the rods vertical). That way the H field will be the same value all the way round and will be vertical to the plane of the paper.

Most of my antenna experience is with electric (dipole) radiators which are simpler to deal with in many ways.

 

[ffjonas]

Hey, :)

I just did what you suggested (Have them mounted at right angles to each other). The error became smaller at position, like at the 40 degree component of each. :) However, when I was trying some other positions, the results do not look good.

I am not sure if I have presented it clearly enough before. The measured magnetic field generated by activating both antennas is close to the sum of that by activating just one at some certain positions. But not at some other positions.

Assume superposition of two antennas works in reality, I can see such a match between these values at a random position.

I have taken the vector addition into account. But thanks for reminding me.

The receiver antenna is a 3D coils in X, Y and Z direction. X and Y direction are in parallel with the plane. Z direction is at right angles to the plane.

So far, I still have no clue about the cause, but thanks very much for your reply. It helps me a lot

 

[sophiecentaur]

So far so good.
Have you resolved the polarisation of the radiated fields? The coils need to be positioned with their axes vertical else the directions of the H fields will be all over the place.

 

[ffjonas]

Hey sophiecentaur,

the coils are oriented in x, y and z directions, which are at right angles to each other. I do not quite get what you mean by "else the directions of the H fields will be all over the place."

 

[sophiecentaur]

What I mean is that the fields you are transmitting must have their H fields vertical -i.e. the rods must be pointing vertically. That is the only orientation for which the radiation patterns are omnidirectional in a horizontal plane. The point of having it that way is that you should be able to see the two contributions to the H field adding up in a very simple way. This is because both contributions will always have parallel (magnetic) polarisations and be unaffected by the direction / bearing of the receive antenna from the transmit antenna. This is the simplest arrangement and will be a good check of the system. Later on, you could have all sorts of orientations of the transmitted fields and do a complete 3 axis analysis of the superposition.

I am not sure that you have taken on board the fact that the field pattern (radiation pattern) of the transmitting elements is very relevant to this exercise. Are my worries unfounded?

 

[skeptic2]

For some years I worked with ferrite rod antennas in the 200 kHz - 400 kHz range.

In the diagram we are assuming antenna A is transmitting and B, C, D and E are all receiving antennas. All antennas are in the plane of the page. The field around antenna A is a torus and where it intersects the plane it makes two circles. In order to get good coupling between the transmit antenna and the receive antenna, the H field must pass through the length dimension of the receive ferrite rod. This means that antennas B, C and D all will have good coupling with A. Antenna E, although parallel to A, will be at a null.

When you have two transmitting antennas with the length dimension in the plane in which measurements are being made, because of the separation between the two antennas, the receive antenna cannot simultaneously have optimum coupling with both antennas. If you want to see field superposition of two transmitting antennas, their length dimension, as well as the receive antenna, must be perpendicular to the plane of measurement.

 

[sophiecentaur]

@skeptic2
This is why I suggested using a different arrangement from your ABCDE setup. If you use the 'other' polarisation for all antennae (i.e. H field vertical) then all antennae will be orientated parallel with the wanted signal and you have an easier job showing superposition. If you get confirmatory results my way round then you can progress to a more complicated arrangement. If not, there are things to sort out with the setup before moving on.

 

[ffjonas]

Hey,

thanks for your reply! It was very helpful. :)

I changed one transmitting antenna to a more powerful one in order to create a stronger magnetic field. The other remains. The issue still remains.

I placed the two transmitting antennas in parallel to the board.

 

[sophiecentaur]

You have still not responded to my point about the H field polarisation. Do you not realize the significance of pointing the rods upwards and the effect on the field pattern in the plane of your measurements?

 

[ffjonas]

I just did the experiment with pointing two transmitting antennas upwards and tried to check the superposition of two antennas.

The results do get better. The difference gets smaller. :)

One question: If I place the two transmitting antennas this way, will they affect each other?

 

[sophiecentaur]

You want to have an idea of the coupling between the antennas. Look with a scope on one feed and see what happens when you short out the other or move it about. If you sync the scope ('Ext') to the RF signal at the oscillator then any phase change at the feed point should show up - and, of course, any amplitude change.
You could check that the radiation pattern of a single antenna is omnidirectional by moving the probe antenna in a circle around the transmitting antenna.

 

[skeptic2]

Which direction is your receive antenna pointing? Is the signal going to one transmit antenna in phase with the signal going to the other? Are all three antennas, 2 transmit and 1 receive, all tuned to resonance?

 

[sophiecentaur]

Resonance could be a hindrance and you might be better off with minimal current in the loops. With those small separations, there shouldn't be much trouble in detecting the signals.
But my suggested method would indicate how much coupling you are getting between antenna, in any case.

 

[skeptic2]

Don't forget there is both series and parallel resonance. With parallel resonance, current is at minimum.

My concern is that if the antennas are not at resonance, one antenna may appear as either an inductor or capacitor to the other and that may cause a phase shift which could affect the measurement. If both transmit antennas are at resonance, even series resonance, then at these distances the signal from one antenna will arrive at the other in phase and have minimal effect.

 

[sophiecentaur]

Don't forget there is both series and parallel resonance. With parallel resonance, current is at minimum.

You're right that the input current is a minimum but there will be a high current flowing in the series loop (sloshing from L to C and back again). This could, I think, cause a lot more coupling.

If the antenna are not matched to each other it is easy to calibrate this out - by measuring the outputs of the two antennae in the same place and comparing them. Actually, tuning two antenna to the exact same frequency would be difficult - pretty knife edge adjustment aamof. A flatter response could be easier to deal with.

It is normal to make measurements using a 'probe' with very little coupling, so as not to disturb the system.

 

[ffjonas]

The receiver antenna is placed on the horizontal plane. The ideal case is that it can be oriented randomly without suffering any negative effect.

I am still not clear about the reason why a non-optimal coupling of receiver antenna can affect superposition of two antennas (I mean, when pointing two antennas horizontally). I checked the results in the case of pointing two antennas horizontally again. The superposition depends on the positions of the receiver antenna. The results present a certain pattern on different positions. Could you please tell me why it is position-dependent?

Thanks

 

[Born2bwire]

Just a quick sanity check here though.

When you take your antenna measurements, you do have both antennas present in all measurements and both of them terminated the same way through all measurements, right? When you run both antennas in the presence of each other you get coupling but as long as you keep both antennas present during measurements and the termination behavior does not change between on/off, then I would expect you to be able to properly capture the coupling and allow superposition.

The next thing is that when you measure the fields, do you orient your probe in three orthogonal directions? The fields are polarized and you need to take measurements along three orthogonal directions (and you should of course use the same orientations for all three situations) so that you get the correct field vector regardless of orientation.

Finally, if you measured the field, you should also measure the magnitude and phase and take that into account when you do your superposition. It's my recollection though that phase can be tricky to measure but you can't measure magnitudes and expect them to add up. You need to account for direction and phase to get the proper interference behavior when you sum the two measurements.

Some of these issues can be avoided by setting up the experiment to work better for you as I believe sophiecentaur has been suggesting.

 

[sophiecentaur]

It would be better if you described the orientation of antennae in terms of either the polarisation or the plane of the antenna. It's never clear what you mean.

You are much more likely to get sensible results if you have all antennas parallel and with symmetrical patterns. For a start, at least, before you have ironed out all the problems.

The only reason that there is any interdependence must be that the antennas are 'seeing each other' due to some coupling. This is why I suggest having an inefficient system (really poor coupling) with untuned antennae and with attenuators in the feeds. It is not usual to 'sample' fields with anything other than a 'probe' which is electrically so small that it doesn't interact with the sources.
The spacing of these antennas is very small in terms of wavelength and, if the are resonant, you can expect significant coupling. I think your observations confirm this.

Think of a normal Yagi antenna. You have a single tuned dipole and nearby elements that are just off resonance. These parasitic elements completely change the pattern of the antenna because of the high coupling between them all. On the other hand, think of a so-called active array of non resonant antennae, With this sort of array, you can combine the outputs of the receiving antennae in appropriate phases (actively) and get 'any pattern you want' and they don't affect each other. This is, in effect, superposition working the other way around.
What you want is to set up two fields which are independent of each other, which means that you cannot afford to have any currents flowing in one antenna due to the currents in the other antenna. That means Low Coupling is required.

 

[ffjonas]

Thanks for the quick reply.

The two transmitting antennas are theoretically in phase. However, there is a very little difference in phase due to limitation of the hardware.

The receiver antenna does have 3D coils which are oriented orthogonally.

 

[Born2bwire]

Thanks for the quick reply.

The two transmitting antennas are theoretically in phase. However, there is a very little difference in phase due to limitation of the hardware.

The receiver antenna does have 3D coils which are oriented orthogonally.

I'm talking about the phase of the field, not the antennas. Do you have a setup that can actually measure phase or are you limited to magnitude? Because if you are only measuring field magnitude and direction you can't expect to superposition to work unless you can figure out the relative phase difference between the fields at that point. But I am at a loss on how you could do that because the coupling of the antennas throws things off.

EDIT: Well I guess a first-order approximation of the phase difference would be to measure the relative path lengths of your measuring point between the two antennas. You could then calculate the relative phase shift roughly by the path difference but this neglects the secondary effects of the coupling between antennas. This requires of course identical antennas and that they be excited with signals that have the same phase.

EDIT EDIT: The above would work best if you are doing far-field measurements. For near-field measurements I do not think that it would work. Heck, I'm not quite sure if it would work well in the far-field since the calculations we do for far-field radiation still take into account the phase shift over distance between observation and source.

 

[sophiecentaur]

B2bw
I'm glad someone else has picked up on the practicalities of this experiment. It needs some serious RF knowhow to get over some of the basic problems that seem to be arising.
The frequency he's operating at and the distances he's using would make phase / distance pretty small.
Also, if he doesn't operate with the transmit antennae having patterns that are symmetrical in the plane of the paper, he's got yet another imponderable.
There is also a question of the receive probe. Is it a real 'probe' or is it yet another (possible) resonant antenna?
I can't be sure how much of this that ffj is taking on board. I think it's harder than he realizes. Possibly his supervisor hasn't seen all the factors either. It wouldn't be the first time.