nRF52840 Meandering Antenna Placement & Layout Questions

[mpkasp]

Hi,

See attached photo for layout placement of the trace antenna.

Altium PCB File: 108-0013-01 PCBA MAIN.PcbDoc

Background:

This is a design for a wearable watch style product. Tight design mechanically, and hard to find an ideal placement for the antenna. There are 3 switches on the left side of the PCB, 2 switches and a USB connector on the right, two large FPCs for the display at the top, and a vibration motor in the cutout on the bottom.

A first pass at component placement has been completed, traces have not been completed. The layout design is based off a hybrid of the 'nRF52840 Preview Development Kit - Hardware files 0_9_0' and 'nRF52840-QIAA Reference Layout 0_5' (the 'var5' project), as well as the old nRF51 dongle for the meandering antenna.

The SMA connector recommended in the Nordic designs is placed (and shorted) for testing/tuning only (trace will be cut when using it). It will be DNP in production.

Questions:

1. I would expect the added trace by placing a series SMA connector would affect RF performance. How do you recommend to account for this? Is the effect as simple as adding the length from the added trace that shorts the SMA conector to the length of the antenna when optimizing it for quarter wavelength? Is the shunt capacitor enough?

2. There is a concern for placement of the antenna right below the switch (MPN: TL1014). Sounds like there shouldn't be an issue placing the antenna on the other side of the PCB through a via. For the wearable device we don't need the full range of BTLE. Most use cases will be a user syncing to their phone while wearing the watch. Should we expect reasonable performance given our use case?

[AmbystomaLabs]

For starters just lose the provision for the sma connector. Any data derived from it will be useless since it will be so far out of phase of the real circuit and will have it's own set of parasitics to deal with. The correct way to tune something like this is with an .047 semi rigid pigtail. You cut the microstip in half then use the pigtail to look at the antenna with a vna and then switch it around to look at the output of the nRF and finally come up with a match for the two.

The switch over the meander will significantly detune it and will ruin the radiation pattern. Also, you could get false switch triggering due to all the RF that will be injected into the gpio. If you truly don't care about range, this can work. But, place a shunt 10pF or so on the gpio that will get coupled to the antenna.

As far as the via feed for the antenna, you will already have made a mess out of the antenna tuning with the switch so the via will just add to the mess you have to tune.

You might want to look into other antennae topologies some are better suited to this style of device. There are smd patch antennae that will have a ground plane underneath, monopoles always sit very close to a ground plane. The meander needs a pretty large area away from a ground plane to work correctly. Another thought could be just a short terminated leaky microstrip.

I'm sure if you google a bit you can probably find half a dozen different antenna designs that would help you out.