RFD900, RFD900+ - New long range radio modem

Hi All,

I would like to introduce you to a new radio modem that we developed for very long range datalinks!


Some of the key features of the RFD900 are as follows:

  • Multi point and point to point link capability.
  • Long range >40km depending on antennas and GCS setup.
  • 2 x RP-SMA RF connectors, diversity switched.
  • 1 Watt (+30dBm) transmit power.
  • Transmit low pass filter.
  • > 20dB Low noise amplifier.
  • RX SAW filter.
  • Passive front end band pass filter.
  • Open source firmware / tools, field upgradeable, easy to configure.
  • Small (30 x 57 x 13 mm), light weight (14.5g).
  • Compatible with 3DR / Hope-RF radio modules.
  • License free use in Australia, Canada, USA, NZ.


These modems are designed to support long range applications, while being easy to use and affordable.  

These modems have been flying in various platforms and have demonstrated excellent performance in real applications. 

RFD900 modems are now available at: http://store.rfdesign.com.au

Support within APM planner and the radio configurator from Michael Oborne is already available.

It works seamlessly with APM planner, all radio Mavlink parameters are available.

Update, December 2014:  The RFD900+ with improved specifications is available now at:


Seppo Saario



Views: 111728

Reply to This

Replies to This Discussion

Thanks for clearing that up.  I just saw the 30mhz crystal and assumed they were overclocking it by 5 mhz.

That MAX2235 looks like a pretty decent PA.  47% efficiency isn't the best, but it seems like you just can't find everything you want in one PA chip anywhere.

Are there any heat issues with it?  I'm guessing it just has thermal vias under it?

Yes, it was a bit of a mission finding a good set of parts and Yes there is a good array of thermal vias under the PA :)   At full power, high duty cycle, the board does get warm without any heatsinking.  So I mounted a small heatsink under the board and implemented duty cycle throttling code with Tridge to prevent overheating.  Only in very hot applications would I expect throttling to kick in.   I was a little worried that not many other modules were using heatsinking, but then digging through some info on Digi X-tend (1W) radios, they start limiting at high ambient temperatures. 

Testing on the board is going well, production test system boards should ship tomorrow.  Lots happening :)

Seppo good to hear the testing is going well. I am looking forward to getting hold of the production model once your ready to start taking orders. I am not that far away in Canberra.

At 1W XTends run very hot.  You can't touch them.


I thought I would check in and see how your testing is going?  Are you getting close to production units yet :-)

Hi Greg,

Testing on the new boards is going well.  The power amplifier PWM code is working well, enabling fine power control steps.  I have set 15 power control levels by default, from 0 to +30dBm for the time being.

The production test fixture (bed of nails) works well too, making bootloading/programming the boards quick.

Still waiting on processors to arrive, scheduled at the end of the week, all going well, first full batch on the SMT line should go through late next week !, then testing and mounting through hole parts.

Still planning on the end of the month for availability :)



Seppo all sounds good and I look forward to being one of your first local customers.  For your 5 power control levels by default, from 0 to +30dBm will you have some operating current draw figures at some point?

There are 16 power control levels, ranging from 0dBm up to 30dBm,

Some draft data for power use is below, the TX measurements are while the transmitter is enabled.  The average current draw depends on the duty cycle while data is being sent.  I have seen Mavlink sessions range from around 20% to 50% duty, depending on what is being sent.

From the below, at +30dBm, average current at a relatively high TX duty cycle would be around, 0.8 * 950mA = ~760mA

0dBm, TX ~165mA, RX ~57mA

22dBm, TX ~390mA, RX ~57mA

28dBm, TX ~640mA, RX ~57mA

30dBm, TX ~950mA, RX ~57mA

RFD900 Modems are now available for purchase from the RFDesign online store : http://store.rfdesign.com.au

End of last week we ran through SMT production in Melbourne, Australia and this week we have run through testing and finishing of the radios.

So far they have flown with a few open source autopilots, with nice range :) 

What kind of range can be achieved?

We did some preliminary tests with the radios at full power, 1 Watt to see what is possible.

This was of course done in the middle of the night in the freezing cold (Winter in Brisbane, Australia).

One radio was on Mt Cootha, just to the northwest of the city, and the other radio was out of a car window on a wooden stick, connected to a laptop. (at ground level).  This is quite nasty for multipath and noise at ground level.

We achieved a bit over 25km with this configuration, at standard air data rate settings (64k).

The picture below shows the two locations where we had a stable link.

The range can be significantly improved with getting the antennas raised above ground level and reducing air data rate which will reduce noise in the receive channel due to narrower filters.

The antennas used were the standard 1/2 wave dipoles at each end of the link.

Next tests will be to an airborne platform, with the GCS antenna elevated a couple of meters.

We decided to be cruel and do some extreme testing to see how the RFD900 modems behave under extreme cold conditions. 

The modem was shut down and allowed to cold soak at -60 degrees C. Then powered up to maximum power and had a look at data transfer and connectivity tests. All works just fine.

The modems are available from Australia from the RFDesign web-store, but we are also making arrangements for a wider distribution network. -- Announcements coming soon!

Reply to Discussion


© 2020   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service