L matching network for ceramic antenna

I am designing a board with the IN100 NanoBeacon. My knowledge of RF is limited to making use of reference designs and I am trying to figure out a design recommendation.

According to the datasheet,

The device does not require external RF matching components when
interfacing to a 50Ω impedance

I am trying to use the YC0010AA antenna from Quectel. The antenna’s datasheet recommends an L-matching network.

While the chipset’s datasheet claims that no matching network is needed, the antenna recommends one.

Since I want to use the antenna in question, should I follow their recommendation? Is there a resource that would help me understand how to arrive at such decisions.

The name of the game is impedance matching. When the output impedance of your SoC is perfectly matched to the input impedance of your antenna, there are no reflections (read: no losses), and you get the designed performance out of your antenna.

This means that the SoC has a 50Ω output impedance, so when used with a 50Ω antenna, you can direct-connect them together (using a 50Ω transmission line). All the impedances match, so there’s no reason to use a matching network to bridge them together.

The datasheet for that Quectel antenna is a bit difficult to follow. My interpretation is that the antenna itself is not 50Ω (which is not uncommon), but when you use the matching network that they provide, it is 50Ω. I think that the return loss and impedance measurements that they show in section 4.2 rely on the matching network in section 4.9.

My recommendation would be to add the footprints for the “L-network” and populate it with their recommended matching components, using RF grade components from a vendor like Johansson. To get the best performance, you’ll want to repeat the VNA measurements shown in the datasheet on your PCB with your selected components. You may need to adjust them a bit.

If this is for a commercial product, you may need to add extra filtering to suppress harmonics, if the SoC is too “hot” out-of-band. In that case, you’ll want to also think about using a 3 or 5 component “π-network” to suppress unwanted harmonics that will fail EMC, or use a 50Ω band-pass filter to remove the harmonics. Tuning antennas for optimal radiation performance and maximum harmonics suppression using a discrete 5 component network is an elegant solution (that I used to do at my last job), but requires a lot of work. The easiest way to approach this is with a “π-network” default-populated with the vendor-recommended parts, and then do a pre-screen at a EMC lab to see how bad your harmonics emissions are. You may be fine without any added filtering.

1 Like

Thank you. It makes sense to me now. This is not for a commercial product. I am doing it for a personal/hobby project to learn RF design.

This happens to be my second RF design. My previous one was an ESP32-C3 based design. I blindly followed their “π-network” recommendation and my board worked. I didn’t attempt to understand the signal characteristics.

I will let you know how it goes.