NANDfarm => RFarmer

Small, easy to use, cheap ZigBee developement board for the internet of things.

Similar projects worth following
This project is now on Tindie:

Let's be honest here, everything you do/buy/hack is price dependent and when it comes to adding to each and every object you think of, a wireless enabled chip, things start to get expensive really fast. Now, there are a lot of development boards out in the wild, capable of wireless communication but not many people are going to buy a 60$ device, to stick it in a flower pot. What if you have many flower pots? Things will get ugly. So why not create a small, cheap, open source development board with wireless capabilities?

Based on the relatively new 8 bit ATmega256RFR2 chip, it features:

- 2.4 GHz ZigBee Transciever with up to 2 Mb/s data rate, Hardware MAC and AES security
- 256K Flash, 8 K EEPROM, 32K SRAM
- 6 timer/counters
- 2 U(S)ART ports
- JTAG and SP

System design:

Got so far (v0.2):

- calibrated the matching circuit on the antenna

- added RF antenna out for matching and external antenna connection.

- all pins are 2.54 mm

- JTAG and SPI separate interfaces

- power led

ToDo (v0.3):

- finish the project website (

- evaluate the new the Atmel Cortex M0+ ZigBee enabled chips

- make it compatible with Atmel Over The Air Upload (OTAU) bootloader

- change the FTDI with a Cypress CY7C65211

- Kickstarter?

  • Tindie time

    Radu Constantin02/27/2017 at 10:32 0 comments

    This project is now on Tindie:

    If anyone is interested, just make a request and I will produce the boards as soon as possible. (There is no need to make extra boards if there is no demand right?)

  • Hold up

    Radu Constantin02/21/2015 at 18:33 0 comments

    Due to some personal matters I put this project on hold. I will release all the necessary informations/files in the "wild". The files are released under the beerware license.

    If you have any questions regarding my work so far I would be happy to answer them.

    Take care and have fun.

  • RFarmer v0.2 testing and antenna matching

    Radu Constantin08/08/2014 at 15:03 1 comment

    Unofficial rules for PCB trace impedance design:

    1. You can never have too many vias.

    2. Add more vias

    3. Follow the datasheet instructions

    4. Every PCB has different impedance so you may have different matching values for different PCB from the same batch.

    5. You can never achieve more than 20-30% efficiency without a Network Analyzer that works at more than 2.5GHz.

    6. If you are new at this, do yourself a favor and start with a dipole antenna.

    7. Before you buy the matching components make sure their tolerances are in specs. A 0.2nH tolerance may not sound like much but it may be the difference between a working device and a pile of nerves.

    As you may have guessed by now, I started testing the second version of the RFarmer. Once all the parts arrived I assembled the boards and I was ready to start tackling the matching circuit for the antenna trace. Imagine my frustration when I found out that the university where I can work with the network analyzer was closed for the summer vacation. What to do, what to do…

    There must be some other way to match my circuit, right? Surely there is a hack-able way to do this, right? Right? ... No there isn’t! Not a full match at least.

    Let’s talk some basics theory that was once learned and then forgotten.

    In circuit theory we learned that maximum power is transmitted through a load when the resistance of the load is equal to the resistance of the source. This is DC. When we talk about AC, the same rule applies, but resistance is called impedance. Now, impedance varies with the frequency of the AC signal, and this is where it starts to be interesting. When you start working with high frequencies things start to get a bit voodoo. The main idea is this. The higher the frequency, the more interference you get from the things around the trace (other traces, vias, pcb dielectric, etc.) When you design a design a trace for an antenna, the main (only) impedance you wish to have is from the capacitor formed by the trace itself and the ground plane on the other side of the PCB. But things are not that simple. As you increase the frequency, stray capacitance (read inductance) will start to form between the trace and other traces near it. Even if you have a ground plane around the your trace, the lack of vias between the two ground planes will screw with the total trace impedance, leading to weird stuff like the device working when your hand gets close.

    ---To be continued---

  • BOM for RFarmer v0.2

    Radu Constantin08/06/2014 at 13:42 0 comments

    If anyone is intereseted, the BOM for the second version of the RFarmer is below in GDoc format. Sorry for the french. :)

  • BOM for RFarmer v0.1

    Radu Constantin07/24/2014 at 08:14 0 comments

    If anyone is intereseted, the BOM for the first version of the RFarmer is below in GDoc format. Sorry for the french. :)

  • Before the RFarmer

    Radu Constantin07/23/2014 at 16:38 0 comments

    Before my ideea for the NANDfarm and the RFarmer; before via stithcing and impedance matching, there was The Base.

    A few months ago (March?) a client asked me if I could build him a set of wireless interconected devices that can be controlled by a tablet. The ideea was nothing new. There are a lot of readily available boards/modules/chips avalilable on the market to do just this. The only problem was to find the right one. So I started looking into some readily available modules and settled on the ATMEL ATZB-24-A2 ZigBit module. It's small, it has an ATmega1281 inside and best of all it has an antenna.  

    After several hours in Eagle I fired up the CNC for drilling, heated the ferochloride shower and I was ready to start the production. I ended up making 3 main boards for the tests and a small USB-to-ZigBit dongle for the tablet. The main device looks something like this:

    Aside for the passive components and the ZigBit module, I also used a Microchip MCP73871 Single Cell Battery Management Circuit and a FT230 for the serial communication. Usually I make my own boards because it's much faster than sending them to the fab. For this I use the CNC and the ferichloride shower available at a nearby FabLab.

    All in all it turned out to be a very nice project so far. Now it's time to commence the programming.

    Unfortnately, this is a work in progress, and I can not yet give the schematics or the BOM but I can happily answer all your questions. When other info is available, I will post it here.

  • RFarmer v0.1

    Radu Constantin07/23/2014 at 15:58 0 comments

    As with every first version of a project, you are bound to find errors and problems. And I had a lot. Thinking back on it I can not imagine why I sent the board to the fab. This being said, I present to you my shameful first attempt at board manufacturing with the hope that maybe some people will learn from my mistakes.

    The ideea was to start the prototyping process as cheap as possible. So first of all, I started ordering samples and after a few mishaps with the postal service here, I finally had some chips to work with.  So I fired up Eagle and started to work on the schematic. After a few hours of work I was happy with the look of the board (at that time) and I started working out the prices and the aquirement of the component.

    As anyone can see, there are a lot of problems with my design. First of all I did not put separate pins for device programming. So I had to connect all the necessary lines direcly to the PORTB headers. This worked for all the connections axept for the RESET. Luckily out of habbit I made the via holes for home PCB production (0.6 mm) and I was able to fit a small pin in them. 

    In my hurry to finish the design I completly forgot the basic rules for differential traces and impedance matching. I am ashamed. However, with the help of a good friend, and a Network Annalyzer from the electronics laboratory I managed to barely match the circuit which resulted in a whopping 10 cm of wireless range. I am still surprised it worked at all. 

    Aside from all the problems I had with this board, I am quite happy with the robustness of the small microcontroller. The single-chip sollution performed excellent even with all the short circuits bad connections, QFN hand soldering and impedance missmatching.

View all 7 project logs

Enjoy this project?



Jannier wrote 10/16/2014 at 20:40 point
do you have de PCB???

  Are you sure? yes | no

Radu Constantin wrote 03/08/2015 at 11:47 point

the v0.2 circuit and layout can be found in the resources. Late reply, I know.

  Are you sure? yes | no

Jasurbek wrote 08/05/2014 at 02:31 point
Radu, Will it be possible to use NANDfarms with mini OSs like TinyOS, Contiki OS, 6LoWPAN etc. ?

  Are you sure? yes | no

Radu Constantin wrote 08/06/2014 at 13:34 point
Yes it is possible. I know the Zigduino (or something like that) uses the contiki OS and is based on the atmega128rfa1, wich is the older version of this chip.

  Are you sure? yes | no

Jasurbek wrote 08/08/2014 at 05:12 point
Thank you Radu, look forward for v3. and I will try to build my v2 when you finish. For v3, I think ATSAMR21Gxxx would be a nice choice as you have stated.
I guess it will be a good kickstarter project, but before you have to make some buzz about this platform, then maybe you can get a lot of supporters.

  Are you sure? yes | no

Jasurbek wrote 08/05/2014 at 02:00 point
Very nice project!!! Thank you for sharing

  Are you sure? yes | no

Łukasz Przeniosło wrote 07/24/2014 at 16:31 point
Radu, do you mind leaving some contact information for me?

  Are you sure? yes | no

Radu Constantin wrote 07/24/2014 at 16:56 point
I updated my profile with my linkedIn and google+.

  Are you sure? yes | no

Łukasz Przeniosło wrote 07/23/2014 at 19:22 point
I was looking forward to creating a device where I could implement tje zigbee stack. I had no idea of this atmel chip existense, i thought i will go for a stm32 arm chip. If you gonna swap for an cortex m0 device, will it also be specialised chip? Or will you build the hardware outside?

  Are you sure? yes | no

Radu Constantin wrote 07/23/2014 at 21:18 point
I am hoping that I will be able to provide both of them; the RFarmer8 and the RFarmer32. They are both specialized chips, but last time I verified (June?) on the Atmel website, only the 8 bit had a preliminary datasheet, so I started with it.

  Are you sure? yes | no

Łukasz Przeniosło wrote 07/23/2014 at 21:22 point
Yes i havent seen the cortex m0 avaible for buying anywhere.
I would love to contribute, maybe order some samples tommorow and try some. If you need some help and would like company on this project im totaly in.

  Are you sure? yes | no

Radu Constantin wrote 07/23/2014 at 21:30 point
Just checked, and it looks like the SAMR21 M0+ preliminary datasheet is up. Nice! Time to sample... :)

I would very much appreciate any help I can get. Thank you.

  Are you sure? yes | no

KingOfKYA(Travis K. ) wrote 07/23/2014 at 18:19 point
Not in local but you might try:

You can really gets shipping cost down with thiner pcbs too:) And with DHL shipping you can get a new board spin delivered within a week.

  Are you sure? yes | no

Radu Constantin wrote 07/23/2014 at 18:25 point
Had some problems with elecrow for a different project with an official price quote. For this project I used OSHPark and aside from the long wait, I am fairly pleased. I will try 3pcb; looks nice. 10x

  Are you sure? yes | no

Elecrow wrote 07/28/2017 at 05:39 point

Dear Radu, This is David from Elecrow. Could you please tell me what the problems you met? I will feedback to the PCB engineer to solve these problem. Thanks in advance. 

Post here that I mayn't reply timely. my Email is

  Are you sure? yes | no

Radu Constantin wrote 07/28/2017 at 06:26 point

The problem was more of a financial one. In the firm I was working at that time we had no way of paying for the boards and then justifying the cost to our internal accountant. Thank you for the offer to help, but his was several years ago.

  Are you sure? yes | no

KingOfKYA(Travis K. ) wrote 07/23/2014 at 20:36 point
Ah too bad, i have yet to really have an issue with them but with all theses board houses in china its best to speak really clearly to make sure there is not a mistranslated. This is a good refrance on how to write for them.

Also i have bough a few thousand, worth of stuff from elecrow so i may be getting special treatment, no idea.

  Are you sure? yes | no

Radu Constantin wrote 07/23/2014 at 21:24 point
I tend to stay away from China manufacturers, but I can not deny the price they offer. My problem was of a more administrative-financial matter, so I don't think it applies here.

Are you sure the link is ok? It just gives me the Wikipedia front page.

  Are you sure? yes | no

KingOfKYA(Travis K. ) wrote 07/23/2014 at 22:27 point
Its the wiki simple page if you look closer just makes English easier for them to read/run though translate.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates