Homebrew 6GHz Pulse Compression Radar

Home projects are sometimes the best way to learn about certain technologies. For example, many have gained knowledge in RF and wireless technologies by experimenting with various software-defined radio (SDR) platforms. However, home projects can also be a place to hone your craft and design circuits the way you want without any restrictions and requirements other than the ones you see fit. For example, cost is generally something anyone making circuits at home will want to keep to a minimum. Finnish electrical engineer Henrik Forsten recently wrote a very detailed blog on his journey designing a DIY 6GHz pulse compression radar, something that was done out of pure interest in the technology.

The idea of radar may seem simple. Transmit a signal, receive the reflected signal, and measure the time delay to determine the distance from the radar device to an object. But as Forsten shows there are many small details to consider. From a high level the first thing to consider is the type of radar to use and the architecture needed to support it. There are largely two different forms of radar. One is continuous wave radar and the other is pulse radar. Typically speaking, continuous wave radar is a simpler solution that uses two antennas and work well for shorter range applications.

On the other hand, pulse radar can use a single antenna and can achieve higher precision and operate more reliably over larger distances. Forsten mentions that he has designed and tested multiple frequency modulated continuous wave (FMCW) radars that have worked well in the past. As a result, for this endeavor he chose to build a pulse radar, or more specifically a pulse compression radar.

The architecture chosen for his project is a direct conversion transceiver architecture using a transmit/receive switch for one antenna while using an additional switch to receive on a separate antenna. This was chosen to allow the system to operate as a continuous wave radar if desirable. In addition, the direct conversion transceiver reduces cost while providing adequate performance. Other radio architectures that exist such as a superheterodyne radio can provide better spurious and image rejection performance but come at the expense of higher cost. Anyone familiar with SDR devices may also notice this architecture is also commonly used among many popular SDR devices.

Forsten's post also goes into lots of details on the choices and trade-offs made while choosing components for the project. For instance, while choosing data-converters he explains the importance of sample rate and resolution on ADC SNR performance. Also the correlation between sample rate and price one can expect to run into when choosing a device. Additional insights include the anti-aliasing filtering design for the ADC and DAC, the FPGA and digital system design, the RF design, and the PCB design. While all useful, the PCB design tips are very insightful. At higher frequencies a PCB can make or break a design, and to avoid costly mistakes simulations are reviewed along with proper termination techniques and data line routing.

The PCB was ordered through a Chinese manufacturer, which also offers assembly. They can fabricate the PCB and place and solder the components available at their factory. It was noted the quality of the PCBs was good considering the price point paid for everything. Although the cheaper FPGA device did have some suspicious markings covered on the cover. Overall, with the PCBs in hand, the programming and testing was able to be completed.

The PCB bring up consisted of creating the digital interfaces between the data converters as well control code for the RF analog circuitry. While evaluating the circuitry the output power was able to be calibrated, the receiver noise was measured, and non-idealities such as DC offset and LO leakage were compensated for in the digital domain. Software was then developed for target detection and tracking and the system was put to the test on the side of a road using homemade horn antennas that were already on hand. Everything worked well and the design was successful.

Forsten notes that, “the designed radar is fundamentally similar to modern large radars. It utilizes digital signal processing, supports arbitrary waveforms and has a very large maximum unambiguous target Doppler velocity due to high pulse repetition frequency. Only the maximum range is shorter than large radars due to low output power and small antenna.” The entire write-up and project is very impressive. To gain some knowledge and insight into radar design, be sure to check out the blog for the full details.