Coaxial Cavity Filter Design and Optimization Using SynMatrix

Coaxial Cavity Filter Design and Optimization Using SynMatrix Hello everyone, this is Aded from Ozen Engineering. In the studio, we will design a 5th-order 1GHz coaxial cavity filter using SynMatrix and optimize it using the AI Optimizer, also known as optiSLang.

Here, I'll start with the filter synthesis and I'll set the filter order to 5. Let's keep the return loss to 25 dB across the band, center frequency of 1 GHz, and the bandwidth of 50 MHz. Here, I'll click Calculate All to update the S-parameters.

We'll be using this topology, and we'll add some specifications. So, for the return loss, we want it better than minus 23 dB. I'll set the start frequency to 0.976 GHz and the stop frequency to 1.024 GHz. Then, I'll click here to refresh. We'll add a second specification for the isolation.

And I'll set the start frequency to 1.056 GHz. Then click here to refresh. Now we can move to the next step, which is 3D modeling. Click on cavity, and here I select coaxial cavity. I click confirm and start new design. Here, we'll be using this flat on top configuration.

I keep the default values and click calculate all. Here on the right side, we can see the results of the unloaded Q and frequency when we fix the diameter and vary the height of the resonator and also when we fix the height and vary the diameter of the resonator.

We can go to the next step, which is the signal cavity. Let's keep the default values for the cavity width and length. So, here I click apply next step. And also, I keep the resonator radius, click apply and next step.

Here, we will be running a parametric study for the tuning screw depth, and I'll vary it from 1cm to 1.7cm with a count of 4. So, I click construct model, let's save it somewhere, and then click run parametric simulation.

Once the simulation is complete, we can see our tuning screw depth versus frequency. Here, if I go to Single, Coupling Matrix, and change it from Normalized to Coupling Bandwidth, you can see here that our frequency is around 1 GHz. And this corresponds to this green line here.

And from this, we can estimate the tuning screw depth to 1. Now we can go to the next step, which is the coupling scheme. And here, we'll be using this through window configuration. And I'll set the iris width to 5mm. I'll click set as main coupling. Apply a next step.

Here, we'll be running a parametric study for the step height, and I'll vary it from 0.2 cm to 1 cm with a count of 4. I'll click construct model and run parametric simulation. Once the simulation is complete, we can click here, and here we have our target coupling coefficient.

So, for M12 and M45, it's around 0.97, and for M23 and M34, it's around 0. 68. Here in green, we have the required coupling range from the coupling matrix. So, for a coupling coefficient of 0.68, the step height is around 0.48 cm, and for a coefficient of 0.97, the step height is around 0.94 cm.

So, I'll go here and type in 0.94 for M12 and M45 and 0.48 for M23 and M 34. Here, I'll click export data to 3D model design, and we can go to the next step, which is input and output.

So, here we'll use the tab configuration, here I'm selecting input, I'll change the tuning screw depth to 1.589 centimeters, click set as input, and then I go to output and change the tuning screw depth to 1.589 centimeters as well. Click Set as output, Apply in next step.

Here, we'll be running a parametric study for the port height, and I'll vary it from 2.4 cm to 2.7 cm with count 4. Here, I'll click Construct Model, Run Parametric Simulation.

So, here we can see our target group delay, which is 10.138 nanoseconds, and we can estimate the port height to around 2.5 centimeters. Now, we can go to the last step, which is the full 3D modeling, and I'll select the input. Here, I change the port height to 2.5 centimeters.

Same thing for the output. Now, let's go and change the tuning screw depth for all resonators to 1.6 cm.

Now, let's go and check the coupling between the cavities, so the step height is 0.94 for M12, 0.48 for M23 and M34, and 0.94 for M 45. Now, we can go to modeling, since we don't have any cross-coupled structure, and click on model construction. Then run simulation.

Once the simulation is complete, we can go to HFSS. Here, we have our full 3D model, and we can plot the results. So, I go and plot the S11 and S 21. And here, we have our initial filter response that we will optimize using the optiSLang. So, here I go and copy this design and create a new project.

I will rename it coaxial cavity optimized. Let's paste the design here and save it. Now, let's go back to SynMatrix and go to intelligent optimization. We will run a custom optimization. Let's do some optimization first, so I'll select it here. Let's go to the simulation file and load our project.

Here, I'll keep the simulation settings as they are for now, and go to variable mapping, and map the physical variables to the elements of the coupling matrix. Here, we have all variables mapped correctly, so I'll go and click save, then I'll click save mapping file, and run.

Now, let's click extract matrix, and let's check the error levels. So, here we can see that all our resonators are tuned high in frequency. Resonator 1 and 5 by around 133 MHz, resonator 2 and 4 by around 38 MHz, and resonator 3 by around 20 MHz.

If we go back to the single cavity results, we can see here that increasing the tuning screw depth decreases the frequency. So, let's go back to our custom optimization and increase the tuning screw depth for all resonators.

So, for resonator 1 and 5, I'll change it to 1. 68. And for resonators 2 and 4, I'll change it to 1. 6. And for resonator 3 to 1.57 centimeters. Here, we have our updated results, so I'll click extract matrix.

And here, you can see a difference between the simulated response and the coupling matrix response due to the dispersion effects that are not reflected in the coupling matrix. To correct this, let's go back to single and apply dispersion. So, here I'll set it to 2. Here, I'll click extract matrix.

And now, the simulated response and the coupling matrix response match. Also, the return loss is better than minus 23 dB, so let's switch to the optiSLang optimizer. So, I'll go to Reset. Here, I'll switch to AI Optimization. Click Confirm. I'll change it to AI. Load the project.

Here, in the AI settings, I'll turn on the adaptive error correction. I'll keep the global trial modification to 0.01cm and change the global max changes to 0.03cm. I click OK. Now, let's go to the simulation settings. I will increase the maximum number.

Now, the simulation is complete, and we have our optimized SWR response. Here, the simulation stopped at iteration 29 because the filter is spec compliant, so I'll select iteration 29 and click Implement. Here, in HFSS, I plotted the results and added some limit lines.

This concludes this demo, thanks for watching. Please contact us at https://ozeninc.com/contact for more information.