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:
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.
The radio is a clean-sheet design. It is based on a Si1002 transceiver from SiLabs which allows a software solution called "SiK" to be used on it.
SiK is also used on the 3DR radios, so protocol, framing, hopping etc. is identical which allows them to talk to each other (900MHz versions only).
One thing that I needed was long range and diversity. Diversity being especially useful as it automatically chooses the best antenna for each packet it sends/receives. For UAV or mobile applications, having two antennas, one in each polarisation, reduces the chance of link loss significantly. I have not seen any ISM band radios to date that support diversity by default.
The RFD900's will be available locally and through distributors. Support is also available locally to make they operate as intended. An Australian based web-store is in development where modems and cable/antenna kits can be purchased.
Will keep you updated on when everything goes live !
Nice work! The 3DR radios are decent, but they didn't really incorporate that many of the features I had initially planned when I proposed the idea of creating a Si1000 based telemetry radio. They chose to go with the HopeRF module instead of creating an original design or using the reference design. It's good see there's still some good engineers out there willing to do more than slap a module on a carrier board.
A few questions come to mind...
Have you done any tests to check your actual receive sensitivity? 3DR claims to have a sensitivity even higher than the HopeRF module they use, despite the fact that I pointed out how foolish and obviously false that claim is.
Which amplifier chip did you use (RMFD, Skyworks, Maxim, etc.)? Or did you use a discrete design?
Is there a passthrough? That is can you turn the amp on and off?
Can you vary the RF output by changing the Si1002 output? Or is the amp gain controllable in some way?
How much current does the unit draw at various output levels?
I see there seem to be at least 6 pins exposed. Have you thought about using your unit as a RX replacement? That is do you think you could output PWM or PPM or even some simple relay/GPIO control?
Sorry to hammer you with so many questions, but it looks like you've actually produced a unit very similar to the one I envisioned and suggested on this forum last September and I'm excited!
Thanks for the feedback ! -
I have done receive sensitivity tests on the prototype hardware, but I will repeat them on production boards with improved enclosures to get definitive figures. (With datarates, bandwidths and BER specified).
The PA used is a MAX2235 from Maxim, it has some nice features such as auto-ramping to reduce transient spurious, reasonable gain and cost. Some parts from RFMD were tempting though :)
There is no passthrough? (PA bypass), but the amplifier is turned off when it is not transmitting.
The gain of the PA can be controlled by PWM output. Currently it has 8 bit resolution, but can be 16 bit with software update. The output power from the Si1002 into the PA is still controllable, but by default it is set to a fixed level, and 16 power control levels are available.
Yes, there are some extra IO exposed :)
- These are usable for almost anything which is nice (crossbar selection of pin functionality).
RE: RX replacement..... something interesting is already in the works. :)
In terms of general IO control, there are already plans to have AT control of these extra pins.
Production PCBs are due to arrive tomorrow (30th May), a verification build is scheduled towards the end of the week to make sure everything is still good! On the production boards I will measure data on sensitivity, current consumption, etc. This will be incorporated into the datasheet.
Will post updates on how things are going and updates on availability shortly - in the meantime, anything else I can help with, let me know :)
Amazing work Seppo! You're really pimping the potential of the Si1002!
I had been looking at using Maxim PAs, but then SiLabs released some reference designs and eval/dev cards using a RMFD chip and one with a Skyworks chip. And I like to stick close to reference designs since I'm no RF engineer. RMFD actually claims their chip was specifically designed to work with the Si100X chips, so I figured it would be worth a look if I could actually find a distributor for their chips.
With my current aerial test platform I can't really spend too much current on the radio and it doesn't have the range to matter much anyways. But on my gas platform I will definitely need the range. I was hoping that the transmit power (and current) could be controlled to only use what it needs for the range in any case.
Hopefully Tridge's code is modular enough to trim it down for some control features to be added. I mentioned awhile back that it's a shame that the most powerful processor on the plane is relegated to just running the radio. I don't know enough about using external memory, but it would be pretty cool to run the whole autopilot and radio from the Si100X. A single chip solution for all the processing on the plane would be pretty slick (and cheap and would use a lot less power).
I look forward to learning more about your project. It seems you've now produced THE top end telemetry radio on the market.
One last question off the top of my head... I notice the HopeRF module in the 3DR radios have the Si1000 overclocked to 30mhz. What speed are you planning to run at? And do you think overclocking the Si100X chips is a good idea or even needed?
There are a couple of oscillators on the Si100x, 24.5MHz is the default internal oscillator that the processor uses. There is a 30MHz external crystal for the transceiver which is as per the datasheet. The RFD900 uses the 24.5MHz internal oscillator for the processor and 30MHz for the radio, same as the HopeRF / Si100x reference designs.
Thanks for clearing that up. I just saw the 30mhz crystal and assumed they were overclocking it by 5 mhz.
Do you intend to manufacture RFDxxx for Europe....868MHz...433MHz..etc.?
There is definite interest in looking into European radios, but will focus on the 900MHz initially. The next month will be dedicated to a smooth launch of the RFD900. After that, what would you forsee as the best band to develop?, 868/433? - bearing in mind duty cycle limitations at 868 and lower power at 433MHz
I think 433 would be the better choise.