FEDEVEL
Platform forum

tomas.kaplan , 06-25-2026, 08:41 AM

Hi there.

I am currently in the middle of fierce battle to implement 2 identical ceramic chip antennas for WiFi and Bluetooth to a very tight and small pcb.

I am aware of not-optimal sizes of ground planes around the antennas, but I am wondering, if I can mirror the footprint for 1 of the antennas like on the picture. I am using YC0009AA where 1 pad is for connecting GND & feedpoint, and 2nd pad is for tuning/GND. WiFi antenna is connected according to the datasheet. The Bluetooth antenna is the one which i rotated by 180° and swapped the GND/feedpoint within the pad 1. Is this something I can do?

Rotating the antenna helps with the layout massivelly. I need matching PI network placed near the feedpoint and there is also filter/matching IC in front of Bluetooth MCU (STM32WB55).

So far, layout is pretty roughly done since i am not sure about this.

Thanks a lot for any kind of help.

, 06-25-2026, 08:43 AM

🤖 In the meantime, here’s a response from the [FEDEVEL AI HW Assistant]().
It helps to find relevant resources for your question, but it may make mistakes as it’s still in early development.

This is the response from the AI Assistant:
While the videos don't specifically address footprint mirroring, we can derive some key insights from antenna placement principles:

1. **Antenna Symmetry Matters**: Chip antennas often rely on specific ground plane configurations and feed orientations. Mirroring the footprint changes the antenna's spatial relationship with the ground plane and nearby components, potentially detuning the antenna. [Relevant ground plane discussion](https://www.youtube.com/watch?v=vX3l0iRxi8c&t=0)

2. **Feed Point Criticality**: The YC0009AA's pad 1 appears to combine GND and feed - swapping these could create impedance mismatches. Even small changes in feed point location require re-tuning matching networks. [Matching network importance](https://www.youtube.com/watch?v=F6iBKXAfC5Q&t=861)

3. **Ground Current Paths**: Rotating the antenna 180° alters current distribution patterns in the ground plane. This can affect radiation efficiency and pattern, especially in space-constrained designs. [Ground plane current discussion](https://www.youtube.com/watch?v=F6iBKXAfC5Q&t=1689)

4. **Practical Workaround**: If mirroring is unavoidable:
- Maintain identical keep-out zones
- Re-tune both antennas' matching networks independently
- Verify isolation between antennas (minimum λ/4 separation)
- Consider asymmetric ground plane treatment if board edges differ

5.

, 06-25-2026, 08:43 AM

**Verification Required**: Even with careful layout, you must:
- Measure S11/S21 parameters for both antennas
- Check pattern correlation using a reference design
- Test real-world throughput in final enclosure

For tight layouts, consider alternative strategies from the videos:
- Shared ground plane optimization [Ground plane sizing](https://www.youtube.com/watch?v=vX3l0iRxi8c&t=183)
- Strategic component placement under antenna (if datasheet allows)
- Using different antenna types for different radios [Antenna comparison](https://www.youtube.com/watch?v=cOtv0ddR3aI&t=2679)

Without seeing your specific layout, proceed cautiously - mirrored antennas often require extensive re-tuning and may still underperform compared to standard orientation.

You can continue chatting with the AI Assistant here:
https://fedevel.com/hw-assistant

Robert Feranec , 06-30-2026, 05:33 AM

i am not an expert on antennas, but what datasheet says? mostly I have seen PCB and ceramic antennas in parallel with GND .... when i searched around it looks like rotation will influence radiation pattern and rotating it by 180 may not be a good idea. It's an interesting topic

Robert Feranec , 06-30-2026, 06:58 AM

So I have got answer from Katerina: No, this flip is not good and will result in antenna either not working at all or very detuned. She is the one from my videos that AI suggested, you can watch them for better antenna understanding.

tomas.kaplan , 06-30-2026, 12:11 PM

Thank you Robert for your answer (and also thanks for forwarding it to Katerina, i have watched some of the videos). Datasheet says to put plenty of ground around the chip antennas (of course), but from my experience, it still can work in very tight designs. This antenna flipping thing is not documented anywhere though, so I was wondering if there is any kind of symmetry within the chip or if it would be completely killed this way. Interestingly, I spoke to Quectel people and they were not conserned by this.. the PCB is already in manufacturing process, so I will get back to you after bring-up procedure (i will be either crying, or laughing).

tomas.kaplan , 06-30-2026, 01:58 PM

To add one thing: I have an emergency plan. I am quite sure, that WiFi antenna will somehow work, because its pretty much according to the datasheet. In a case, where BT antenna performance is poor, i am planning to test Pulse (Yageo) W3008 on the exact same spot. It is actually symmetrical antenna with its specified layout being quite close to the one I have on the PCB. The orientaton would be also correct.

tomas.kaplan , 06-30-2026, 01:59 PM

This is exactly what I did with the Quectel antenna, This picture comes from the W3008 datasheet.

tomas.kaplan , 06-30-2026, 02:01 PM

If only RF would be more seethrough, rather than voodoo magic..

tomas.kaplan , 07-14-2026, 05:41 AM

Hi @Robert Feranec , i am back with some results.

Mirorring the footprint for Quectel chip antenna actually works and can be matched pretty good.

WiFi antenna, as expected, works pretty well too. I am not sure, where the second resonance comes from, but I can live with it.

See the pictures of actual layout and S11 measurements. (Note the series unpopulated positions from antenna pad to GND - its intended for inductors for tuning of resonance).

tomas.kaplan , 07-14-2026, 05:41 AM

Bluetooth S11:

tomas.kaplan , 07-14-2026, 05:44 AM

WiFi S11

tomas.kaplan , 07-14-2026, 05:45 AM

Measurements were done inside the plastic enclosure as it will be in the production. The question ofcourse is, how the radiation patterns look and how it will behave in the real use case, thats the thing I need to test as a next step, but this matching alone is a good start.

Robert Feranec , 07-14-2026, 11:29 AM

nice!