Accelerate Design Flow with Optenni Lab – Volume II

Posted by Olli Pekonen | 16 April 2024

Introduction

In this Optenni blog article, the themes introduced in the previous blog on Feb 12^th, 2024 will be expanded to multiport antennas, in particular to tunable antennas. We also rely on the teachings of an Optenni blog article on preassessments from Nov 28^th, 2023.

Antenna to Study

Let’s take a look into a schematic representation of an aperture tuned antenna in the figure below.

In the picture, there is a simplified view of a three-throw switch with two capacitors and an inductor as tuning component in a tuning port (let’s call this port 2), and the Antenna feed port (to be called port 1), to which the matching circuit is to be connected. Let’s further assume that this antenna is to be used at three different bands, calling them low band (700 – 800 MHz), mid band (1800 – 2000 MHz) and high band (2600 – 2900 MHz), as implied by the three throw port components.

At first blush, the design of this kind of an antenna seems simple in conventional RF circuit simulators, at least if we perform the optimization of the device in the return loss sense, and not in the efficiency sense. Just minimize the S₁₁, and we are done, right?

Well, how are you going to treat the tuning components and the states of the switch? How about just looping them all through, one optimization for one band, a total of three, with the target of minimizing S₁₁.

This rises three further questions:

Q1: How do you really know if the component in the tuning port is going to be an inductor or a capacitor, even if you get its value optimized?
Q2: With the three proposed optimization runs, you will get three different matching circuits for the feed port. But you only can have one, catering for all the three bands and the three tuning components (through the states of the switch) at the same time. How to cope with this?
Q3: And how would you synthesize the topology of even one matching circuit, and not resort to guessing and optimizing the values of the components of a fixed topology, which is also based on a guess?

The answers are as follows:

A1: In general, you do not. You have to try everything out, 2³ = 8 times for three inductors or capacitors.
A2: Frankly, the only answer we know is: Use Optenni Lab… namely, the current RF circuit simulators are not well suited for this kind of a problem.
A3: See A2.

Of course, with Optenni Lab, you would do the design to maximise the total efficiency, not just focus on S₁₁. Remember that you get excellent S₁₁with a lossy system, with a 50 Ohm resistor in the extreme. Obviously, this is not the path we should take.

Using Preassessment Results

It is good to remember that Optenni Lab has preassessment tools for the antenna data which can be used even before any matching circuit design is attempted. We covered this already in our blog article on 28^th November 2023 (which conveniently relates to the same antenna) .

With the teachings of the said blog, we know that for the high band, we can likely cope with a simple short circuit as the tuning component, for the mid band we are best off with a small inductor, and for the low frequency, Optenni Lab should choose. This kind of preassessment will save us a lot of time, as nothing is to be optimized for the high band, and no capacitors need to be considered for the mid band.

Posing the Design Task to Optenni Lab

The secret of Optenni Lab handling the various bands each having a different tuning component, and a same matching network so smoothly is called a “frequency configuration”. You simply make three frequency configurations in the Optenni Lab interface, in the Multiport Antenna Matching mode (for example), as show in the figure below. We call the frequency configurations set 1, 2 and 3 for low, mid and high bands, respectively.

Then you assign the targets, one for each of the frequency configurations. You may use classical efficiency targets, and boost the process with impedance targets. Below, we have examples for set 1 and 2 (set 3 settings are obvious and thus not repeated)

Finally, we need to set the switch to follow the frequency configurations. For example, for frequency configuration set 1, set the switch to state SP3T_100, for set 2 to state SPT3_010 etc.

The tuning port setup now looks like this:

Please remember that that the tuning components at switch ports RF2 and RF3 were deduced from the preassessment results. Different states of the switch (and their corresponding frequency configurations) are shown with colored dashed lines.

Finally, the overall synthesis task looks like this, with the somewhat condensed contents of circuit 2 shown above.

Note how trivially easy it is to ask Optenni Lab to keep the synthesis results the same in the feeding port 1. We simply place an automatic 0-6 component synthesis block to the port, and Optenni Lab takes care of the rest, using the same (but yet to be synthesized) topology for all three frequency configurations and states of the switch.

After some 60-90 seconds of number crunching, we get the following promising result.

For example, for the high frequency band, the result is as follows, nicely catering for the needs at that band, only -1.8 dB off from the perfect efficiency result of 0 dB.

We can also study e.g. the power balance diagram at the center of the high band, with the following outcome:

This clearly illustrates that we have achieved more or less perfect matching condition at this band, but the always present resistive losses and component losses make the result less than ideal.

The same observations can be naturally repeated for other bands.

Note how well Optenni Lab keeps the designer up-to-date of his/her design progress. More information means less guessing, less errors, better product quality and a faster time-to-market.

Conclusion

Handling multiple bands in antenna matching, each requiring both identical (static) matching portions, and band-dependent (dynamic) matching portions is very easy with Optenni Lab. All this can be done

relative to the total efficiency of the antenna which is the true figure of merit of an antenna system, and
with full synthesis capability for matching and tuning circuit topologies.

We encourage the antenna community to stop wasting time in guessing their matching and tuning networks. Instead, please use Optenni Lab. Hit that Start Free Trial button at Optenni homepage now!

Olli Pekonen
Sales Director
Email: olli.pekonen (at) optenni.com

Accelerate Design Flow with Optenni Lab – Volume I

Posted by Olli Pekonen | 12 February 2024

Introduction

In this Optenni blog article, we take a look into some of the ways many Optenni Lab accelerates the design flow of antenna designs, and antenna matching in particular. We first focus on a straightforward one-port antenna, and then in later blog articles expand the story e.g. to the challenging domain of multiport (aperture tuned) antennas.

Designers responsible for a high performance antenna of a small electronics device are facing hard times when the size of the radiator is mostly dictated by mechanical constraints of the device. At the same time, the antenna has to cater for many bands, and have a high-enough radiation efficiency.

It is like the designer is in the middle of a Bermuda triangle, with i) small antenna size, ii) wide bandwidth requirements and iii) good radiation efficiency as the corners.

In fact, the three factors in the corners counteract each other – for example, it is relatively easy to make a small antenna that radiates well if it has to operate only on a very narrow band. Remember that in the extreme case, following the teachings of basic radio engineering, any load can be matched with just two components, but at a single frequency only. Similarly, a small sized multiband antenna is quite easy to create if any efficiency is allowed. In the extreme, the “antenna” is a small 50 Ohm resistor, resulting in an excellent wideband match, and in a dismal radiation.

Traditional Antenna Matching Design Flow

With traditional circuit simulation based tools, the design flow of a one-port antenna matching is based on the following:

A. Obtain the S₁₁ of the antenna.

B. Neglect the efficiency and other radiation aspects of the antenna.

C. Guess a matching circuit topology for the antenna into a static topology (e.g. a ladder of inductors and capacitors) and connect it to the one-port representing the S₁₁ of the antenna.

D. Optimize the values of the statically placed inductors and capacitors to get a suitable notch (return loss) to the S₁₁ at the input of the matching circuit at the desired frequency bands.

This traditional process faces many challenges, however. Let’s go through them one by one:

Step A is fine – we will need the S₁₁ always anyhow
Step B is problematic: Neglecting efficiency of the antenna makes the solution suboptimal, as the total efficiency is the true figure of merit. This is because the total efficiency takes both matching and radiation into consideration.
Step C is fine if the designer happens to guess the right combination of inductors and capacitors. But what if the designer makes a bad guess? In fact, there is no way to know this, so to be sure, every combination must be manually entered and tried out. For the possibility of either inductor or capacitor in shunt or in series (4 possibilities / component), there are 4^N combinations to go through. So for 2 components, there are 16 topologies, for 3 components, 64 topologies etc. In fact, it is highly unlikely that the designer guesses the right topology, and to counter this, a lot of manual placement is needed.
Step D is OK as such, but suffers badly from the inadequacy of steps B and C. For example, what good are optimal element values if the topology is bad in the first place.

The Optenni Lab Way

Optenni solves the challenges of the method steps A – D in one sweeping stroke. Namely,

Optenni Lab is efficiency aware, and the main target for optimization is very often simply the total efficiency of the antenna system. By reading in the radiation patterns of the antenna (or a tabulated radiation efficiency file), efficiency is fully integrated into everything Optenni Lab does. This solves challenge of step B.
Instead of manual placement of components, Optenni Lab synthesizes circuits automatically. So there is no need to tinker with a schematic circuit diagram, interchanging components from inductors to capacitors or wise versa (it should be added that for the users wanting more control, manual placement is possible, however). With this, challenges of step C are solved.

Example

The designer is faced with the design of a dual-band antenna operating in frequencies 1695 MHz – 2020 MHz and 2570 MHz – 2620 MHz. Both S₁₁ (in Touchstone format) and efficiency files are available, from 0.1 MHz to over 5 GHz.

The designer simply needs to read in the S₁₁ and efficiency files, enter the number of components the matching circuit may have (for two bands, 2 x 2 = 4 is a good guess, two components making a “resonator”) and the two bands of interest. This takes about 30 seconds to specify in the user-friendly user interface of Optenni Lab. Compare this with entering every possible combination of inductors and capacitors manually – we would have 4⁴ = 256 options to go through. Assuming that it takes a minute to change the topology and note for any improvement (this is likely a bit optimistic), the task would take more than 4 hours.

The synthesis run takes less than a second in an average laptop. This is the best topology Optenni Lab gives:

And this is the frequency response result

So in all, a manual tinkering worth of more than 4 hours in a traditional RF circuit simulation and optimization tool was reduced to a design task taking less than 40 seconds in a Optenni Lab.

But wait – when we write “the best result”, exactly what do we mean?

To cut the long story short, the best result (in this case) is the result with the best worst-case performance of the efficiencies over the two frequency bands. So the S₁₁ of the matching circuit is just a starting point, and it is the efficiency that counts. Note, however, that the S₁₁ has the valleys / notches at the frequencies where the efficiency peaks. This is reflected also in the Smith chart with S₁₁ arches that loop around the center of the Smith chart as closely as possible at the bands of interest as the frequency varies. If the S₁₁ would go through the Smith chart center, the matching would likely be excellent, but only in that single spot frequency, and far off in the other frequencies.

Conclusion

This was maybe one of the most simple use cases of Optenni Lab but it goes a long way to illustrate one of the prime ways Optenni Lab accelerates the design flow by

focusing on the most important figure of merit in making well-radiating antennas, the optimization total efficiency, and
by synthesizing circuits automatically instead of forcing the user to manually tinker with horde of circuit topologies.

Of course Optenni Lab offers so much more even for a single-port matching, with built in support for library components, their statistical, sensitivity and tolerance analyses, versatile user interface for result management and post processing etc. And things get even more exciting with multiport matching. We will cover the ways Optenni Lab can accelerate your design work in later blog articles.

If this blog article aroused your interest and Optenni Lab is new to you, do not hesitate to hit that “Start Free Trial” button at our WWW site.

Olli Pekonen
Sales Director
Email: olli.pekonen (at) optenni.com

Exploring Antennas With Optenni Lab

Posted by Olli Pekonen | 28 November 2023

Introduction

With the blog article of September 2023, we touched the main design philosophy of Optenni Lab, where the user may Explore the antenna, Accelerate the design of antenna matching, and Maximise the impact of matching.

In this blog article, let me focus on the first aspect of this philosophy, Explore.

Before matching circuits are synthesized and time is spent to gain adequate performance, it is very beneficial to get an overview of the theoretical limits of the antenna’s radiator. If the radiator has very little chance of surviving the communication task at hand, it is better to go back to the drawing board and try to redesign the radiator. After all, if a failure is to happen in engineering, it always pays off to fail fast and in early stages of the design, and then move on.

Optenni Lab Preassessment tools

There are three main preassessment tools in Optenni Lab: Bandwidth potential, electromagnetic isolation and total scan pattern.

Electromagnetic isolation can reveal that the isolation between two antenna ports is, in fact, caused by poor matching. If we take away the effect of poor matching, it may become evident that the two antenna ports are quite strongly coupled. This can become a problem for antenna systems with many radiators.

Total scan pattern is an important feature in assessing the ability of array antennas to direct radiation over the solid angle enclosing the array.

Preassessment Workhorse – The Bandwidth Potential

Bandwidth potential is likely the most frequently used preassessment tool in Optenni Lab. It gives a quick overview on the following: What is the width of the frequency band at each of the frequencies available in the antenna impedance data (Touchstone file) if we wish to reach a certain matching level (using an ideal two-port matching circuit).

A typical bandwidth potential graph is given in Figure 1 below. It shows that the antenna in question will be difficult to match to -10 dB return loss level at frequencies below 1.3 GHz. At approximately 1.8 GHz a massive bandwidth of 700 MHz is easily achievable. Again, at around 3.6 GHz a broadband matching of more than 100 MHz (twice the indicated potential of 50 MHz for a two-port matching) will likely be hard. And finally, around 4.4 GHz, a 700 MHz bandwidth will again be a simple task.

It is very typical that the bandwidth potential tapers out at both edges of the frequency spectrum, as the number of frequency points runs out (the closer we get to the start or end of the frequency data, the more narrow the bandwidth potential becomes out of necessity). Also at low frequencies, antennas tend to be short in terms of wavelength, and thus the ability of the antenna to radiate is quite limited, dropping the bandwidth potential further.

With single port antennas, the usage of bandwidth potential computation is very straightforward. Just go to the Assessment menu and choose Bandwidth Potential, set the matching level, and off you go, as in Figure 1.

Figure 1: Bandwidth potential reveals the performance of the bare antenna in terms of bandwidths

The Bandwidth Potential of Multiport Antennas

When the antenna is a multiport antenna, for example an aperture tuned two-port structure, things get a lot more interesting!

Consider the basic building blocks of an aperture tuned antenna with a feeding port 1 and tuning port 2 as shown in Figure 2. Green portion marks the meal strips of the radiator, and orange portions are the conducting ground plane. Tuning port 2 has a switch element with various load impedances (inductors or capacitors), that can be dynamically connected to the second port 2.

Figure 2: Basic building blocks of an aperture tuned antenna

Let’s also assume that the antenna is to operate at three distinct frequency bands, and let’s call them

“low band” at 800 – 900 MHz,
“mid band” at 1700 – 1800 MHz, and
and “high band” at 2500 – 2700 MHz,

so that for each of the bands, the switch is at one of its throw positions.

Now the basic design challenge is: What kind of static matching circuit is the best for port 1, and what the three tuning components should be in port 2?

But before we give this task for Optenni Lab to be solved, we can look into the bandwidth potential of the antenna so that we try different tuning port loads at port 2, and see how things look like from port 1.

In version 6.0 of Optenni Lab, studying tuning port loads together with bandwidth potential is really quite simple. In the past, the user had to export one-port S-parameter file and open it in an “auxiliary project” to generate the bandwidth potential graph. Not anymore! As shown in Figure 3, simply first terminate the port 2 with a load you wish to study (say, a 1 nH inductor).

Figure 3: Aperture tuned antenna subjected to 1 nH tuning port load

Then, choose Additional plots… from the plot selection panel,

Figure 4: Bandwidth potential is now accessible through Additional Plots

Add Bandwidth potential and Radiation efficiency as plot types.

Figure 5: Bandwidth potential and efficiency can be turned on or off

Finally, set the bandwidth potential parameters (matching level and symmetry).

At the same time, it is also beneficial to study the radiation efficiency of the radiator. As the tuner port load changes the currents of the radiator, the tuner port changes also the radiation efficiency. This aspect is easily overlooked when designing tunable antennas.

Figures 6 and 7 show typical results for the resulting bandwidth potential and for radiation efficiency, respectively.

Figure 6: Bandwidth potential with a 1 nH load in the tuning port 2

Figure 7: Radiation efficiency with a 1 nH load in the tuning port 2

As an additional plot, bandwidth potential computation is a game changing feature in Optenni Lab 6.0! It makes the generation of bandwidth potential chart (where several bandwidth potential curves are stacked) almost trivial. Simply tune the load at port 2 to get a selection of bandwidth potentials related to various inductance loads. You may then go also through some capacitor values, and short and open circuit options in port 2. Then, push all these results to a User Plot, and you easily get the following kind of chart:

At the same time, it is very easy to generate the following radiation efficiency chart, almost as a by-product.

Figure 8: Bandwidth potential chart

Figure 9: Radiation efficiency chart

Interpreting Bandwidth Potential and Radiation Efficiency Charts

Now let’s see what these two charts reveal for us.

For high band at 2500 – 2700 MHz, the choice of tuning component is clear: The best bandwidth potential is achieved with a short, and as the radiation efficiency is more or less independent of the tuner loading (as the curves are stacked on top of each other around 2500 – 2700 MHz), we can safely assume that the switch should couple to a short when the antenna is operating in the high band.

In the middle band at 1700 – 1800 MHz, both the bandwidth potential and radiation efficiency are at their highest values when there is a small, 1 – 2 nH inductor in the tuning port. Thus, we can let Optenni Lab optimize the inductor value, but there appears to be no need to consider also a capacitor in the tuning port at the middle band.

For the low band at 800 – 900 MHz, the radiation efficiency is again quite good with small inductors, but the bandwidth potential is really poor (more or less zero) for these loads. However, there seems to be a compromise based on a large inductor (10 nH) value, which implies a large impedance. Such a large impedance can also be a small capacitance. Thus, it is likely best that this tuning component is left for Optenni Lab to choose, based on the versatile Generic Reactance component.

Figure 10: Tuner optimization setup (RF1 – low band throw port, RF2 – mid band throw port, RF3 – high band throw port)

Benefits of this approach are clear:

For the high band, nothing needs to be synthesized or even optimized as there is simply a short in this port,
For the middle band, we just need to optimize for inductors, not for capacitors at all.

Ability to omit these two synthesis steps saves time considerably.

Conclusion

Taking a critical look into the radiator before any matching is attempted is usually very beneficial, especially for tunable antennas where the synthesis times can be long. With the bandwidth potential and radiation efficiency plots, the secrets of a tunable radiator become a lot more visible even when the tuning port load is varied.

If both bandwidth potential and radiation efficiency are low for any tuning port load and at any one of the frequency bands, redesign of the radiator of the tunable antenna should be considered. On the other hand, if any of the tuning components seems to indicate a good performance in both bandwidth potential and radiation efficiency, we can likely choose that component as a starting point for synthesis, and sometimes even set its value (especially for short or open circuits) without doing any synthesis or optimization. This speeds up the synthesis task considerably and gives the antenna designer a firm handle on the potential performance of the tuned antenna radiator, even before any matching is attempted.

Olli Pekonen
Sales Director
Email: olli.pekonen (at) optenni.com

Author Archive: Olli Pekonen

Accelerate Design Flow with Optenni Lab – Volume II