RCWL-0516 Reverse engineering Motion detection sensor

[fedimakni]

Hello,

I wanted to reverse engineer the RCWL-0516 by designing my own PCB. I followed the link below and i understand how the pcb work
https://github.com/jdesbonnet/RCWL-0516
Attached a pic of the top side of the pcb Capture5
Attached a pic of the bottom side of the pcb Capture6
I also desoldered the same component from one working pcb and solder them to my pcb but without success.
I couldn't understand what could cause the problem? Do my antenna design could contain some mistake?
Do anyone have an idea how to fix it or what i should do?
Do anyone have the gerber file for the real working pcb so one could intergrate them easily in his design?

Looking forward to your help. Thanks

[JBeale]

I gather that board operates around 3 GHz.  How much experience do you have with microwave PCB circuit design? I don't have much, but I've worked with others who do. Every detail of that board, including the components, the exact shape and width and thickness of every copper trace and via, soldermask and silkscreen, the resin and fiberglass and processing of the PCB itself can have some effect on circuit operation at those frequencies. Developing such boards requires a lot of cut-and-try patience even with adequate test equipment like network analyzers covering the frequency of interest to measure the actual signals, behaviors and interactions.  Even working at much lower frequencies (433 MHz) I know there are significant batch-to-batch variations of boards from the exact same design files, presumably due to dielectric constants of the particular FR4 material each time, and maybe process tolerances and things like humidity.  Doing this kind of development without such test equipment would be a real challenge.
I don't know what the current state-of-the-art is, but at least a few decades ago I know that the standard approach was to use expensive teflon-based boards around these GHz frequencies (eg Rogers Duroid), because typical cheap FR4 material is so variable and poorly characterized it's hard to know how a circuit will behave when femto-farads of change in capacitance start to become significant.

[Clarisah]

Hello, I’m working on the same project as I would like to embed it in my PCB design for another application. You design looks like a perfect match, did you get it to work in the end?