Hi,

Thank you for the initiative here: in my university we did not hear a single word how to properly document during the development process, but I think this is very important for all engineers.

it really extremely depends the type of documentation. I am working for an international company with big team size and working across timezones.

To document development process, we heavily depend on note-taking applications where you can enter formatted text, tables, figures, equations. input normally comes from many different sources, like simulation results, measurement results. You can easily document meeting minutes and timelines a hypothesis for something, give, Action Items, etc etc.
The notes are stored on the cloud and accessed anywhere, from laptop, tablet, smartphone.
Examples are Quip, Evernote, OneNote, and there are many other tools out there.

If the document is more result-oriented than process oriented, we heavily powerpoint or keynote. These tools are also great to document figures, easily draw and add text and equations. As the results in many cases are anyway discussed with other colleagues in a web-conference, this is already a good preparation for the discussion and the upcoming meeting. The slides help the author focus on one key point per slide. And as we know, figures often are more descriptive than plain text, anyway.

I was holding a university course in the past, where the students had to design some RF circuits. They where asked to document their design from the beginning with OpenOffice Impress, as they had to present the final result and final circuit to the group.

Hi Phil,

I am not sure that your idea is really feasible, due to some practical reasons:
Where do you get the 3D with channel capacity? Making such a map is huge effort, involves test drives and so on.

Of practical interest for microwave engineers are two areas:
1. AI based models, which can be used for simulation. This training data should be obtained by measurement campaigns over many different conditions. Practically, some physical behaviors are not easy to describe in closed form models, therefore AI based models are used. One example is power amplifier modeling of nonlinearities. The main goal of the model here is to analyze the performance of the device under test with this model, and be able to simulate its performance.
2. Another topic is AI based Radar detection. In this area, targets are classified by machine learning signal processing. One example is actually the latest Google Pixel, which is using a radar sensor to detect gestures.

BR

The Frequencies linked in the presentation are representing only a very small percentage of 5G-NR bands defined in 3GPP. Basically, nearly all LTE bands can be used for 5G-NR, plus some extra Sub8 bands (below 8GHz) plus the millimeter wave bands (24-60GHz). A good source can be found here, https://www.sqimway.com/nr_band.php which covers 3GPP Release 16.
The website links to the 3GPP specification, which is not really for beginners...

Yes, for sure. In 77G Automotive Radar, saturated PAs are heavily used. The design is much simpler and offers higher efficiency. As no amplitude modulation is required, a nonlinear class of operation can be used.
The main disadvantage of saturated PAs is their reduced Power Supply Rejection Rate. Spurs and noise is not suppressed as much as with linear PAs. This could lead to higher far-off noise and undesired spectrum emission.
The other disadvantage is possibly a higher harmonic emission, due to the nonlinearity of the PAs. This should be in any case be suppressed by proper output matching networks.
The third disadvantage is that output power control and output power calibration is more complicated, as the output does not follow the input level in a linear way.