Discussion about RF reference designs

Go To Last Post
5 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi all,

I have in the past successfully used the AT86RF230 radio chip. In that case I copied the layout quite accurately from the Atmel application note AVR2005 (http://www.atmel.com/Images/doc8092.pdf). This resulted in good enough performance of about 50+ meters line of sight range with a small chip antenna.

Now I am prototyping with Atmega128RFA1 and closest to a reference design I have found is the application note AVR2044 (http://www.atmel.com/Images/doc8339.pdf). I would like to design a system with similar level of performance (about 50 meters range is enough).

However I am a bit confused about which way to go. In the AVR2005 a star-like ground separation/splitting scheme is used and it's importance is clearly explained. Also, I have seen it work in my own designs. But in the AVR2044 the ground plane is not split at all, it's completely solid over the whole PCB. This application note also talks about separating the digital and analog ground currents, but I see no evidence of it happening in the layout. Digital and analog grounds get happily mixed everywhere. Even one digital ground pin (55) is directly connected to the paddle under the AVR which is internally connected to analog ground. Also, the crystal layout looks quite vulnerable in contrast to the AVR2005.

Do anyone of you have any comments or experience? What kind of range can one expect between two RCB128RFA1 boards? Do you know of any other reference designs for Atmega128RFA1 that we could discuss here?

Last Edited: Fri. Oct 16, 2015 - 02:24 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm not an expert on ground separation, but from the experience, RCBs have very good RF performance under normal conditions.

NOTE: I no longer actively read this forum. Please ask your question on www.eevblog.com/forum if you want my answer.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Do you have any experience about the line-of-sight range outdoors with the RCBs?

I also see that they have included a shielding cabinet over the AVR for protecting it from external radiating disturbances. I wonder if this is a because the AVR itself is vulnerable, or is there some other weakness in the system... I have always been able to use the AT86RF230 chips without extra shielding cabinets and they have successfully passed even the hardest industrial EMC tests. The problem is that these cabinets are very expensive in low-medium volumes and they force some mechanical and layout restrictions to the design.

I would really be interesting to see other ETSI proved designs out there with Atmega128RFA1.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In my experiments I've got about 30 meters, but with a wooden door and a few corners in between the devices (I don't have 50 m LOS in the office).

I think shielding is required for FCC certification, I run all my devices without a shield cover and see no problems with them.

NOTE: I no longer actively read this forum. Please ask your question on www.eevblog.com/forum if you want my answer.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Basically, you can apply the approximately same layout to the ATmega128RFA1
as you already did for the AT86RF230. The RF frontend in both chips is
fairly similar.

There are vastly differing opinions about whether split GND planes are better
than a single, low-inductance plane. Both can be made work, but if not done
properly, both could be seen fail.

As usual, the most important factor regarding the achievable range is typically
the quality of the antenna. I recently compared a number of antennas in my
own lab, and I'm still surprised with the results from the PCB antenna described
in appnote AVR2006. If you've got the space for that kind of antenna, it
outperforms almost everything else.

Jörg Wunsch

Please don't send me PMs, use email if you want to approach me personally.