Concurrent RFSoC Demo PG.mp4: this mp4 audio file was automatically transcribed by Sonix with the best speech-to-text algorithms. This transcript may contain errors.
Ray Hoare:
Ray Hoare here with Concurrent EDA. A quick take on the RFSoC system on a module. This module contains a Zynq RFSoC from AMD on a small module that you can embed into your device. These are in full production, and a couple of key features here is the RF front end. So on the RF side, these analog-to-digital converters, there are eight inputs that can directly sample RF spectrum at over 4 gigasamples per second. And also to transmit, there are eight RF DAC channels — digital-to-analog converters. And those can run at 6.5 gigasamples per second. So this is quite the beast. In addition to that, it's a full Zynq. So you have four cores on there running Linux and a full system with 4 gigabytes of memory as well as a bunch of supporting I/O as well as down here on the I/O, if you have some digital I/O or LVDS, lots of things on there. This example I want to show you is the SoM — the system on a module. That's underneath the black heat sink — and it does require a heat sink because it's doing that many gigasamples per second — mounted to a simple PCIe carrier board. And so that PCIe carrier board is not doing much except for transitioning the inputs to the board.
Ray Hoare:
In this demo, what we have is a loopback. So we're transmitting and receiving data. This is a nice little format. This is a mini form meant for external GPUs, but we used it for this demo. What's behind it — that's a power supply, a regular power supply, and it's going and talking to a PC. So if you want to put this inside of a PC, this works great. Or if you want to put this in your own enclosure, you don't need the PCIe bus. If you want to use the network coming out the back, that's also viable. So this example I'm going to show you here is the input signal. You can't really tell what it's doing. But what this is is a carrier frequency and a signal frequency combined. And specifically, if we look at: This is what we're sending out and what's coming in, we did an FFT, and you can see we have spikes here. So we have two different frequencies that the FFT is showing us. And then at the bottom, we did some filtering, some bandpass filtering, and pulled out those frequencies. So what we show here is the signal. I apologize. This is a recording of a recording, but it's much faster. So you can see the signal wiggling at the top and you really can't tell what it is. We do the FFT to figure out where that signal is. We then apply the bandpass filters, you know, 1 megahertz, 3 megahertz. So this is a simple example. And then we pull the signals back out. And so that way when you have an RF signal, you can sense what's out there with the FFT, and then we can perform the filtering. We can do this in the megahertz, gigahertz. We really can push it. And this is a simple example, but it proves the point. Simple block diagram. Here's the traditional Zynq. So it's very familiar. And then what you have is this special front end where the data is going in and out, and that's the RF front end. A side view of the same thing. Big power supply behind it. Not necessary. It's a 5 volt supply. You don't need all of that power, but that came with the external carrier. Applications of this system on a module might be when you're looking for drones and you need to be looking at the RF frequency that they might be using to communicate, or if you're in a certain environment where you're trying to say, "Well, there shouldn't be any cellphone communication in this environment," you can use the RF sampling to look at those frequency domains and to figure out what's going on. Or if you have a noisy system or you're trying to do some debugging. Maybe you need to be looking at the RF frequencies that your system is emitting that didn't pass your environmental qualifications. But this module is designed to be small, designed to be embedded. If you need a variation of this or you need a custom carrier or you want some FPGA design on "How do I get that data and send it out?" — we've done work there too. So that's a quick take on the RFSoC module from Trenz Electronics. If you're interested, send me a note. We are a distributor for Trenz. And if you need some design work, let us know too. Thank you very much. Bye.
Automatically convert your mp4 files to text (txt file), Microsoft Word (docx file), and SubRip Subtitle (srt file) in minutes.
Looking for the AI transcription tools? Read our in-depth best meeting transcription software and compare top picks in our guide to Happy Scribe transcription review. See how Otter vs Rev stacks up, plus our full breakdown of Dragon speech alternative.
Our customers range from interview-to-text to enterprise transcription — and everything in between. Explore our capabilities for Teams meeting transcription, conversation intelligence, and how to make English subtitles.
Try Sonix free today. No credit card required.
