Difference between measuring radiated power and S11

To make both happy, S11 and Radiated Resistance (so the Radiated Power) I am using a small anechoic box (~1.8m length) and a radiated setup. This is ok only if you are using frequencies greater than about 800MHz, otherwise the dimensions of the antennas are too large.

The setup is: a VNA, the DUT antenna connected to P1, and a Reference Antenna connected to P2.

Tune the DUT antenna (including the matching network) until you get best compromise between S11 and S21. S21 in this situation is direct proportional to radiated power. The test setup can be calibrated replacing the DUT with a dipole antenna.

Thats the reason why mobile antennas are not directive and are omni in all principle planes. Theta 90 phi 0 and 180. With high gain, if the user is away from the BTS the radiated power is increased.

S11 is the parameter that tells you how much of available power that is transferred to the antenna. If that not results in best S21 is it due to limitations in the way S11/S21 is measured. Any else idea is misunderstanding of what S11 is. Antenna radiation resistance relative loss resistance gives antenna efficiency, but whatever these values are does still best S11 result in best S21. The sum of these vector values is what you measure as antenna impedance. Most common when best S11 not results in best S21 is due to that measurement cables are connect in a way that they affects the measurement. It is especially complicated when the actual antenna ground plane is small or lossy. A way to common problem is that NA calibration not is done including the actual measurement setup and the DUT. Another problem is when antenna is assumed matched against a transmitter when 50 Ohm is measured for for the antenna without checking what impedance that results in most emitted power from the actual transmitter or what impedance that results in best S/N for the receiver. These values are complex impedances so electrical delay in the measurement must also be accounted for when calibration is performed.

vfone,
Do I have to have a second antenna for measurement? Can I just have one antenna as TX and RX? If so, when does transmission and receiving happen? Is their a time delay between the two?

Thanks.

Say as Radar, it has just one antenna for TX and RX. The radar emits a pulse signal and then receive the reflected signal with various delay. The software algorithm is very complicated. So for simple test, you should use a second antenna.

vfone said:

To make both happy, S11 and Radiated Resistance (so the Radiated Power) I am using a small anechoic box (~1.8m length) and a radiated setup. This is ok only if you are using frequencies greater than about 800MHz, otherwise the dimensions of the antennas are too large.

The setup is: a VNA, the DUT antenna connected to P1, and a Reference Antenna connected to P2.

Tune the DUT antenna (including the matching network) until you get best compromise between S11 and S21. S21 in this situation is direct proportional to radiated power. The test setup can be calibrated replacing the DUT with a dipole antenna.

Click to expand...

Do you have a picture of the setup? I would be curious to see it.

kae_jolie said:

but in case the antenna is near other lossy objects like human body, the power radiated would be decreased due to losses (due to power absorbed by human body), but S11 should remain the same. Is this correct? Please confirm.

Click to expand...

S11 will change when the antenna is near other objects. Some (all?) cell phones have built in antenna matching that measures S11 and compensates the mismatch.

Antenna resonance frequency is lowered when surrounding dielectric is increased, such as when a human body more or less cover the phone.
When a cell phone antenna is designed must these effects be considered. Most common is to adjust the antenna to be a bit high in resonance in free space so it is optimized in function when Sam is holding the phone. He will of course reduce and affect radiation pattern but that can not be compensated for by adjusting S11. Dynamic adjusting of S11 in a phone is therefore of limited use, and is never used.
Not exactly adjusting of S11 but dynamic frequency switching antennas have been tested for phones. As far as I know have it never been implemented yet. Future LTE MIMO phones are however maybe needed to do something like that to fully use available bandwidth but there are also other types of solutions.

E Kafeman said:

Dynamic adjusting of S11 in a phone is therefore of limited use, and is never used.

Click to expand...

It is used:
**broken link removed**

Sorry, no. The technology is since long developed by several chip vendors, but it is not implemented.

volker_muehlhaus said:

S11 will change when the antenna is near other objects.

Click to expand...

volker, if statement above is true, how come HFSS power radiated measurement decrease consistently (linear) as the adjacent object increases in size, but S11 measurements fluctuate. It seems that power radiated measurement has a more accurate "make sense" reading. I am using the same dipole with the same impedance matched source for all different sizes of the adjacent object.

As the adjacent object increases in size, I notice the the power radiated and hence radiation efficiency decreases, however, S11 measurements are not linear. Why is that? I want to use S11 measurements in my paper, but because they are inconsistent, I can't use them, and I want to use power radiated measurements, instead.

kae_jolie, I do not know what method HFSS uses to calculate the "radiated power", and how reliable that is.

To see how S11 changes when the antenna is near other objects, just connect a network analyzer and do the experiment.

I noticed, that you mentioned HFSS before, but also real measurements. In my understanding most previous comments about the relation of S11 and transmitted power have been for real measurements. In HFSSS, you can do some things, that are difficult or even impossible in real measurements, e.g. determine the total amount of radiated power rather exactly. On the other hand, if your HFSS setup is incorrect, the results may be void. It's true, that you can achieve excellent results with this tool, but also complete garbage.

Without knowing the details of your setup, I guess nobody will be able to explain all observed effects. Assuming your HFSS setup is basically plausible, the previous statements about real antennas should also apply to your measurements. For the said adjacent objects, it would be interesting to know, if they are pure reflective, or absorptive? In addition, do they have dimensions that enable them to act as resonators?

S11 will change when the antenna is near other objects.

Click to expand...

This has been actually stated by all contributors, as far as I understand. But no particularly relation has been suggested, because there are many kinds of objects and many ways to place them.

volker, my experiment is trying to show that there is more power loss (or power absorbed by object) as the adjacent object size increases. Question is: Is it ok to use power radiated to show that? Is power radiated measurable in a lab like S11? If so, how?

Also, the HFSS directivity measurements are linear and I may use those instead of power radiated. Is directivity measurable in a lab, as well? If so, how?

Thanks.

---------- Post added at 17:43 ---------- Previous post was at 17:06 ----------

FvM said:

In HFSSS, you can do some things, that are difficult or even impossible in real measurements, e.g. determine the total amount of radiated power rather exactly. .

Click to expand...

The same goes for S11 measurements and any other simulated measurements. It's just a simulator that uses computational EM to simulate real measurements. It's understood the results are relative.

FvM said:

For the said adjacent objects, it would be interesting to know, if they are pure reflective, or absorptive? In addition, do they have dimensions that enable them to act as resonators?

Click to expand...

They consist of fat and muscle in a half-cylindrical shape.

They consist of fat and muscle in a half-cylindrical shape.

Click to expand...

In this case, S11 is basically unrelated to radiated power, because the losses are not negletable rather than intentionally high. It would be still meaningful to tune the antenna matching for minimal S11, and you can assume that you get the highest radiated power in this case. But the ratio of dissipated to radiated power will be unknown.

Is power radiated measurable in a lab like S11? If so, how?

Click to expand...

Is directivity measurable in a lab, as well? If so, how?

Click to expand...

As said, an accurate measurement would require an anechoic chamber, two-axis rotation of the sender and integration over the total sphere surface. I made a comment about the difference between HFSS and real measurements in this regard. If you don't have an anechoic chamber, outdoor measurements can give a rough estimation.

FvM said:

In this case, S11 is basically unrelated to radiated power, because the losses are not negletable rather than intentionally high. It would be still meaningful to tune the antenna matching for minimal S11, and you can assume that you get the highest radiated power in this case. But the ratio of dissipated to radiated power will be unknown.

Click to expand...

Fvm, thank you for shedding light on this. Could you elaborate on "the losses are not negletable rather than intentionally high" and "the ratio of dissipated to radiated power will be unknown" I am missing something to grasp the whole meaning. It's very important for me to know when to use S11 and when not to use it.

I am assuming that I can tell how much is absorbed by the fat/muscle body by radiated power because I see radiated power decrease as the fat/muscle body increase. In a real environment, if I can measure power radiated (assuming no other losses by other objects because only antenna and fat/muscle body will be present in small anechoic chamber), then I can comfortably assume most of the non-radiated power got absorbed by the fat/muscle body, hence, I have an indirect measurement of power absorbed in fat/muscle body.

Now I have not done any real measurements because I don't have the means to do it. All I have been doing is HFSS simulations which I later want to verify in FEKO.

Question: what's the boundary of anechoic chamber made out of? Is it lossless to prevent loss of EM energy outside the chamber? I am simulating this in HFSS by an air radiation box, is that ok?


Thanks.

kae_jolie said:

Question: what's the boundary of anechoic chamber made out of? Is it lossless to prevent loss of EM energy outside the chamber?

Click to expand...

That would be a shielded metal chamber. These do exist, for other types of measurements.

The anechoic chamber walls are almost perfectly absorbing lossy walls, so that there is no reflection.

As FvM has pointed out, you simulator can integrate over the entire sphere, to calculate the total radiated power. In measurement, that is very difficult.

At first, I must consider that I'm not using HFSS. I used some other tools like APLAC in the past and have a basic understanding of the methods for radiated power measurement. I previously guessed, that your HFSS results (relation of radiated power and S11) may indicate an incorrect usage, but now I'm quite sure, that they can be well explained by the nature of your experiment.

Previous contributions already discussed the effects that can be expected in your setup.
- The adjacent objects will change the antenna characteristic and thus affect S11. Assuming, the antenna matching has been optimally tuned before, then more power will be reflected and thus radiated power decrease, even with non-absorptive objects.
- In addition, the object can absorb part of the power.

So as a result, two effects are superimposing. In a typical mobile phone (or other portable transmitter equipment) antenna scenario, the disturbing objects (the user) are taken as granted. So the problem is to optimize antenna matching, as discussed by volker, possibly optimizing the antenna geometry or position for minimal disturbing effects, or simply knowing the amount of absorbed power. The problem is also related to health care.

In don't know if adjusting the antenna matching is a topic, too. Otherwise, you can either assume, that the S11 change can't be compensated, because there's no means for it. Or S11 could be minimzed in each case.

Human tissue absorption is usually modelled by saline solution conatainers (e.g. bags) in real meaurement. I know, that respective measurements and simulations have been performed in scientific and mobile phone manufacturer's projects. But I don't think that similar measurements are an easily achievable option for your project.

volker_muehlhaus said:

As FvM has pointed out, you simulator can integrate over the entire sphere, to calculate the total radiated power. In measurement, that is very difficult.

Click to expand...

volker,
As vfone pointed out above, S21 when using a reference antenna in real measurement, should be proportional to radiated power. I agree HFSS calculation of radiated power is different from real measurement, but it should be proportional to real measurement, meaning if simulated power radiated increases, so should measured power radiated. I am not concerned so much about the real measurement of power radiated, as long as I know that it can be measured in a lab somehow. I will use it in my paper instead of S11 measurements.

Vielen Dank.

FvM said:

Human tissue absorption is usually modelled by saline solution conatainers (e.g. bags) in real meaurement. I know, that respective measurements and simulations have been performed in scientific and mobile phone manufacturer's projects. But I don't think that similar measurements are an easily achievable option for your project.

Click to expand...

Mostly agree in what FvM writes but want to clarify that in most cases is human body impact of a mobile phone antenna in practical measurements according to very specific standards. SAM is a standard head and tissue fluids are standard recipes (mostly sugar or DGMB, almost no saline). Search for SAM phantom head and HSL-900 & HSL1800 for recipe and pictures. It do also exists standard hands, torso or full body. If we not followed these standards would it not be possible to compare results between different measurement ranges. Common measurements ranges when measuring different body parts impact of antenna are anechoic chambers or reverberation cambers but for low frequencies such as when embedded 100 MHz FM antenna is measured are mostly outdoor ranges used.
SAR for exampel, is very important that its measurement results can be verified/repeated in several independent laboratories so all equipment and uncertainty in measurement calibration must be included in the measurement report such as this one.