What is ArduRF?
ArduRF is a wireless electronics development platform closely based on Arduino hardware. The objectives are simple:
- Take the Pain out of Wireless Links
- True Arduino Compatibility
- Long Range AES-128 Encrypted Radio Link
- Battery Powered
- Operate from a Solar panel
Why Another Wireless Arduino Project?
Because it hasn't been done right yet. I hope to change this for my backers and future customers. Previous projects
- Use a different CPU
- Don't work from batteries
- Can't run all existing Arduino sketches
- Are not compatible with all the major operating systems.
- Can not work in the 915MHz, 866MHz and 433MHz license free bands
- Were very far from production ready.
We fixed all that.
This project will ship the early bird reward within weeks of the funding of the project with an Atmega328p CPU, battery charger and radios on board.
An Independent evaluator that has been testing the prototypes has reported a battery only powered ArduRf1s running for 80 hours sending a 40 byte message once every 5 seconds. This was done WITHOUT putting the ArduRf1s and the radio in Low Power Sleep mode.
The ArduRF Family of Boards
The ArduRF-family consists of three designs, each optimized for a different application environment. The ArduRF1 is a completely Arduino-compatible board. It has the same shape, connectors and CPU as an Arduino Uno R3 and operates at 5V even when battery powered. It has a LiPo battery charger and battery connector on board. The integrated boost converter provides 5V when operating off the battery to maintain shield compatibility.
The ArduRF1 and its CPU operates at 5V and 16MHz - even when running from the battery. The ArduRF1 has the Arduino Uno form factor and uses the FTDI FT231 chip for the USB interface. The form factor, pinout, CPU, and 5V operation ensures that the vast majority of shields will just work. All the experience and working knowledge of Arduino's are directly applicable. The standard IDE and examples continue to work without modification.
The ArduRF1s is a smaller version of the design and maintains all of the features of the ArduRF1 except it operates at 3.3V. All the pins available on an Arduino Uno are brought to connectors on the board edge . The pin pitch is 0.1" and the width is 0.8" making it possible to mount the board on a solderless breadboard and still have two breadboard holes on each side available for connection to other devices.
The third member of the ArduRF-family of boards is the PC2ArduRF1. It is the smallest and lowest cost member of the family and is specifically designed to wirelessly connect a PC to one or more of the other ArduRF1 members.
An off-the-shelf transceiver is used to minimize risk to the project. The transceiver is available in three different frequency bands, 433MHz, 868MHz and 915MHz. It has been tested on the first and second prototype (915MHz version only to stay legal) and found to have exceptional range more than 500m in an open field. (That is meters, not miles). The data rate is adjustable up to 300kbps and the transceiver supports hardware AES encryption making the transmissions secure. Simple wire monopoles are used as the antennas since they work well and there is no material benefit to using a manufactured antenna. Wire antennas save the cost of the antenna and the RF-connector - which cost more than the antenna.
The ArduRF1 contains a lot of circuitry to sequence and select the power source for the CPU and transceiver. The transceiver is a 3.3V circuit and a regulator is provided to power the transceiver. Level shift IC's are used to convert the 3.3V signals to 5V.
The main power input is a DC power jack followed by a switch mode power supply regulating the voltage to 5V. A switch mode regulator was chosen to allow the battery to be charged from the DC input and to allow higher input voltages to be used without requiring a heatsink on the regulator. The input voltage limit of the buck regulator is 42V. The 42V input allows the ArduRF1 and ArduRF1s to be used in automobiles and on motorcycles without a pre-regulator. This allows a low cost telemetry system to be designed around the ArduRF1(s) for racing cars and motorbikes.
A charger for a Lithium Polymer rechargeable cell provides 400mA of charge current to the cell and will automatically terminate the charging when the cell is fully charged. The ArduRF1(s) is fully functional while the battery is being charged.
The LiPo cell has a varying terminal voltage depending on the remaining capacity. For most of the discharge time of the battery the voltage will be approximately 3.7V but will fall to 2.9V before the battery protection circuit shuts down the ArduRF. To power the ArduRF, the battery is connected to a boost converter to provide a steady 5V supply over the entire range of the battery voltage. In the first prototype the regulator for the transceiver was powered directly from the battery. In this version the transceiver regulator gets its input from the 5V rail. The result is a small decrease in efficiency but the transceiver works optimally until the battery protection circuit shuts the system down. In the first prototype the transceiver started to have a shorter range once the battery reached 3.4V. This may be the exact time where the best range is needed to be sure that messages reporting the low battery condition get through.
The ArduRF1(s) contains circuitry to select the correct power supply under all circumstances and this was one of the most challenging aspects of the design. The ArduRF1(s) uses power from the available sources in this order:
- DC power jack,
- USB input,
- The rechargeable battery.
The charger runs from the on board 5V and the 5V is provided by the buck regulator from the DC jack, the USB input, or the boost converter from the battery. The battery charger does not try to charge the battery when the 5V comes from the battery via the boost regulator.
Some of the rewards include one or two LiPo cells capable of powering the ArduRF1 for an extended period. Right now, I can't be more specific since the tests are ongoing. The manufacturers label claims 1500mAh as the capacity for the cell but testing has shown the capacity to be closer to 1400mAh than 1500mAh. None of the rewards contain more than two cells since that is the limit on the number of cells that can be mailed.
Solar Powered Wireless Arduino
The ArduRF1 and ArduRF1s can operate from a DC input as high as 42V continuously, charge the battery at 400mA , power the transceiver and the CPU. You can power the ArduRF1 and ArduRF1s from a 12V solar panel without problems where the voltage is all over the map from 0V at night to 18V or more during the day. When the solar power is there the battery will charge and the 5V or 3V will power the CPU, transceiver and auxiliary circuits. When the sun sets or is obscured by clouds the LiPo cell will take over seamlessly. The LiPo-cell offered with the rewards power the ArduRF1(s) continuously for 24 hours while the radio is transmitting. Recharge time varies a bit with cell age but is around 3 hours from a fully discharged state.
The first prototypes were designed, manufactured, tested and beaten up. It had some issues that needed fixing. Right now the (hopefully) final prototype set is undergoing evaluation by a select group of developers to verify that it works as intended, is truly compatible, and is mechanically sound.
The ArduRF Prototypes
Five prototypes have been manufactured to bring the design to its current state, three for the ArduRF1, and one each for the other two boards. The other two designs were prototyped after the development of the ArduRF1 was complete and it was being tested. Since they are subsets of the ArduRF1 - they worked straight out of the oven.
The third ArduRF1 prototype has been manufactured to fix some minor issues regarding the legibility of the text once the connectors are installed. The connectors come very close to the text making the pin labels hard to read. The SP header pins were also not labeled in the first two prototypes. The position of the radio module was moved. A small change was made to the power switching circuitry to prevent two 5V supplies from being on at the same time. Occasionally, this lead to the boost converter supplying the circuit while the DC input was available.
When fixing the silk screen lettering gets to the top of the list of things to fix on the PC board the the design is ready for production.
On the first prototype (at the top of the photo) the transceiver sits on stilts to verify that the ground plane on the PCB didn't interfere with the transceiver. The raised transceiver did indeed have better performance than the transceivers mounted directly on the PCB showing that the ground plane did interfere with the operation of the transceiver. The transceiver was moved and the ground plane was altered to ensure that the transceiver works correctly when soldered directly to the PCB. You can also see that the reset switch migrated from its original position to the top left corner of the PCB. In the original position it was too hard to reach when there was a full size shield stacked on top of the ArduRF1. The LEDs for the battery charger and Tx/Rx indication was on the edge of the board of the first prototype but the standard Arduino LED was against the opposite edge of the board and almost all shields covered it. The standard LED was also moved to the left edge of the board where it is more visible when the ArduRF1 is covered by a shield and positioned so that the other LEDs are visible.
The photograph above shows a PC2ArduRF next to an ArduRF1 showing the simplification of the design when the PC supplies regulated 5V. The transceiver is soldered to the bottom of the PCB.
The project started with a single ArduRF1 board. I was asked numerous times for different versions of the main design. Two options stood out: the first one is for a simplified version to allow a PC to communicate with the full blown ArduRF1. Since it is tied to a PC it has no need for a battery charger or other supply since it can get its power directly from a USB port. The request was honored in the form of the PC2ArduRF. To ensure it is compatible with all the major operating systems, the FTDI USB to serial interface is still on the board. The module has a USB A-type connector to directly plug into a USB without the need for adaptors or cables.
An optional plastic case is available to protect the PCB and the circuitry of the PC2ArduRF.
ArduRF1s breaks the shield compatibility and the 5V operation but offers a lower price and smaller form factor. The ArduRF1s is the ArduRF1 without the boost DC-DC regulator and operates at 3.3V instead of 5V. It is much smaller than an Arduino board and is available at a lower price than the ArduRF1. The CPU is still the Atmega328P and the FTDI USB interface is retained but the board is reduced down to 2.25 x 0.9 inch. To save space (and money) the DC jack and boost converter was removed from the design. The battery charger, USB interface, and the step down switcher remains on the design. That means that voltage input range is still to 6V to 42V for the DC input and the full battery charge current is available over the entire input range.
The battery will also charge when the ArduRF1s is connected via the USB plug. Since the CPU operates at 3.3V the power management is much less complex than on the ArduRF1 resulting in major space savings and reduced design headaches.
The ArduRF1s the preferred board for development where 5V operation is not required and space is at a premium. The pin pitch of the connectors is 0.1" for ease of use.
What remains to be done is to buy boards in large panels to simplify the production. The first and second ArduRF1 prototypes were made on panels to allow the production reflow solder profile to be developed and to verify that the boards build and solder correctly. I also experimented with soldering the transceiver at the same time as the rest of the components using lower melting point solder paste. The experiments were successful but it turns out that the transceiver interferes with the programming of the boot loader software so that option was abandoned.
Once the project is funded and the designs are finalized they will be published online for download. I will publish the complete schematics as PDF files. I will publish the PCB layout as Gerber files. Unfortunately I don't use any of the hardware tools commonly used by the Arduino community so my source files are in the proprietary formats used by Pads and OrCad. If someone volunteers to capture the design in an open tool such as KiCad I will be more than happy to verify the schematic files and netlist against the actual PCB and then we can verify the new PCB against the KiCad schematic. The designs are published under a Creative Commons License.
The rewards consist of various combinations of the ArduRF family of boards. To keep the list of rewards manageable the various radio frequency options are not listed separately and the user choice will be determined after the pledges are made. Please note that I have no plan on shipping 433MHz radios to US and Canadian addresses since that option is not legal. The default option is the 915MHz/868MHz radio. The transceiver works well at both frequencies as long as the antenna is trimmed to the correct length. European users should program the radio for 868MHz and North American users should use the transceiver at 915MHz. Please also note that the encryption may be disabled and it may be necessary to disable the encryption to remain legal. Model rocketeers please pay special attention to your hobby's restrictions on the use of radio links.
All rewards include a pair of transceivers since these boards are mostly useless without two boards and two transceivers. User who may want us the ArduRF boards without the receiver will have to buy from my web site or Tindie once the Kickstarter is over.
The short answer is that development costs money and this design has been unusually expensive to bring to this stage. What started out as a project to make an Arduino clone with a good, long range radio link morphed into a family of products. One design is now three and the costs have multiplied accordingly.
To offer the Lithium Polymer cell at a reasonable price I have to buy a very substantial number of cells. Most factories have a 3000 piece minimum order quantity and the large number of cells has to be shipped as 'hazardous goods'. This alone makes buying less than several hundred cells at a time prohibitively expensive.
I will build the design in the USA on my own equipment to ensure a high quality product. To keep the price reasonable the amount of labor going into each board has to be kept to a bare minimum. On the prototypes the DC power connector was placed by hand because I don't have the correct feeder for a part this large. Pick and Place machines use feeders to position the part to be picked from a precisely known position and orientation. I have feeders for many standard parts but not for the DC power jack. Feeders are precision engineered and cost a lot of money. It is not uncommon to invest more money in feeders than in the Pick and Place machine itself. Since the project grew in scale I am now short of essentially all feeder sizes. A few more feeders will allow me to build all three boards in quick succession without having to remove reels from feeders and load different reels in the same feeders.
The next financial outlay is to pay for reasonable quantities of the components. Although most components can be bought in quantities smaller than full reels the cost can be prohibitive. Someone must count the parts at the supplier and put them on reels. This all goes into higher prices for small orders. For small orders the shipping cost makes up a much larger percentage of the overall cost than for larger orders.
If this Kickstarter is wildly successful - I will have to add to the workforce to be able to ship on time. With your support I hope to bring a truly Arduino compatible product to market with a great radio link at an affordable price. Please help make this project wildly successful by telling 5000 of your closest friends about it and urge them to back it.
Risks and challenges
The biggest risk to a timely conclusion of the project and delivery of the rewards is unexpected component shortages. The Atmel CPU is notorious for becoming scarce and expensive but presently the distributors are well stocked. None of the single source parts are difficult to find from multiple distributors. Parts were chosen where numerous distributors carry stock over lower cost parts that have single suppliers. I have a reasonable stock of CPUs on hand and will add to my stock if I see supplier stock run low or shortages are reported.
Another risk to timely delivery is equipment failure since I have only one set of assembly equipment. My Pick and Place machine and Reflow Oven are both made in the USA and are about one year old. Most spare parts are available with only a short lead time but if a major sub-system were to fail it may take as long as six weeks to obtain a replacement. In the event of such a failure I will sub-contract the assembly of the PCBs to someone (already identified) that has the same equipment as I do.
Even in the event of a runaway success of the Kickstarter the production is unlikely to overwhelm our capability to assemble and test the boards due to the capacity limitation of the production equipment. It may however become necessary to work an extra shift to keep to the promised delivery dates and that will incur extra labor costs. It is likely that the extra costs will be offset by the reductions in the component prices and PCB costs and by operating a production line for testing instead of the batch testing currently envisaged.Learn about accountability on Kickstarter
- (31 days)