I’m developing a custom DW1000 board as part of a university research project. I have tested a prototype on FR4 which has been adequate for basic ranging tests, albeit around 20-50cm over estimates, resulting presumably from antenna delay in the design. This could least in part due to the antenna traces being calibrated for RO4350B with a dielectric constant around 3.5 (for FR4 I believe this is more like 4.2 -4.5).
My question is, which is more important; substrate material or antenna traces? The university milling machine couldn’t cope with RO4350B so I am considering staying with FR4 for the next design iteration - could optimising the antenna dimensions for the higher dielectric constant improve performance enough to avoid outsourcing the milling?
I have attached the top copper Gerber to give an idea of my design, all feedback is welcomed! (I’m aware the track transitions are awful in this software, please ignore those…)