A distributed and scalable autopilot architecture is attractive to me due to space and sensor positioning demands on my airframes. The existing range of autopilots simply doesn't fit. The uNode (microNode) is the result of my desperate attempts to fix this. uNode is a very cutdown version of Pixhawk with a little bit of AUAV-X2 added.
Board contents are:
- STM32F427 Processor (as per pixhawk)
- 64Mbit onboard flash - no sdcard
- 4.5V to 10V input power supply or power from USB.
- No USB connector, only 0.1" holes
- MAX21100 6dof sensor - an experiment and about the only sensor I have space for.
- No I/O buffers. ESD and EMI protection on everything.
- Power for GPS module
- SRAM Battery backup
- Pinouts for
- 2x I2C
- SPI
- 3x UART
- 2x CAN with transceiver onboard
- 8x servo outputs
- PPM input
- SBUS or Spektrum module input.
Total manufacturing cost in volume may be ~$25.
These boards are designed to fit into the servo hatch of a wing. They have about the same area as a slim wing servo. There is no USB connector because there is absolutely no space for one. There is no sdcard because it is not accessible. Some other potential uses:
- As a CAN connected failproof board
- Failproof remote servo expander
- Remote sensor interface - I2C sensors over long wires without interference.
- A backup, get you home autopilot. Just add GPS module.
The board uses the same processor as the pixhawk so that the same OS and software is used. You pick whatever you need to run in what node location. The dual CAN can then be used to setup a robust network to transport the data between nodes. The PX4 firmware is structured to be very suitable for this.
The prototype hardware functions well. The onboard flash works and I have had bi-direction file transfer over mavlink-ftp working. I had it working... The PX4 codebase has taken another leap forward, broken my branch and I somehow need to catch up again.
That's all for now until I manage to make it do something useful.
Comments
@Roberto, Not much difference apart from size. If it doesn't fit inside the aircraft it is as much use as all the other options.
AUAV-X2 comes in at 30*45mm which is smaller than the uNode. That will fit nicely in the autopilot space I have. Unfortunately it is too deep and too expensive to fit in the wing spaces.
Hi Matthew ,
why do you think that your design is so different of our original design VR Brain 4.0 available until June 2012 :)
http://www.virtualrobotix.com/profiles/blogs/vbrain-coming
Actually available in rev 5.2 in micro and standard size ? My opinion is that in your design you don't have important future like sd card that are usefull for advanced code .
In our old design we don't have sd card but add the support in rev 4.5 the main difference respect of pixhawk is that our design is on single processor like your . I tested maxim sensor but onestly is cheaper but i prefer a lot inversense acc/giro after test st, freescale and maxim ...
The VR Brain 4.0 the first 32 bit platform that supported APM copter firmware 1 year before Pixhawk or PX4 with two revision of code one without OS and last with nuttx os support.
Best
Roberto
Hi Matt,
Nice work. Congratulations.
I am not surprised that you decided to design the system that you wanted.
How large of a market do you think there is for your design? Are you going to produce and sell them? Perhaps approach someone else to manufacture them?
Best regards,
Bill
@Mark MAX21100 is one of the lower cost sensors. Around $3.50 @ 1k volume. It doesn't make any sense not to put it on. It will make a nice backup autopilot with an extra gps somewhere on the bus system.
The 64Mbit memory is about $1. I would put a larger memory on if I had physical space for it.
@James Pike - That is pretty much my thinking also. Thanks for the feedback.
@ben @thomas The FFC or sensor damping idea is nice but simply not for this project. Any additional connectors add too much size. I am already concerned that the power supply capacitors are making the board too thick. If you want high integration please look elsewhere.
Another aspect of the STM32F429 processor. It allows the use of a suitable RTOS which assists in flexible process management. That means each board can flexibly run different parts of a process with no problem.
The implementation of ORB (object request broker) helps to share data around the network. It is quite memory intensive because you need a data buffer for every process. The higher capacity processor enables this.
The atmega parts are completely out capacity so there are much more restricted on what you can do. Pay an extra $5 and get much more bang.
@thomas butler - Now that is a comment of worthiness! I have seen the Naza teardown showing their use of this idea - it does mean that external vibration mounts are not needed and you can just stick it down, calibrate, and fly.
I also agree that the lowly atmega chip still has life in it left - apm 3.2 is so rock solid with just a neo-6m gps, I think if there is an internally dampened micro APM board released it would be great news for F330 & F250 fliers. (but I think the 32-bit chips are useful for addon items like optical flow, sonar, lidar).
So maybe 3DR can have a budget atmega board, but also the advanced board (Pixhawk), instead of just dropping the former. I don't know, I also guess the pixhawk price is dropping to early apm price too so who knows. Internally dampened pixhawk ftw!
@Matt So the idea is to make an inexpensive CANbus node that can also double as a complete autopilot for minimal size/cost apps? I guess the onboard flash and MAX21100 don't add much to the BOM cost?
Pixhawk replacement. I would prefer an offboard barometer/mag- allows for better weather sealing and mag isolation. Redundancy on the pixhawk seems lame considering how reliable processors are compared with the rest of the vehicle. Best thing small size and only solder throughhole connections. Huge source of failure is mechanical connectors and that source gives no warning and is impossible to test for.
@James Pike. Out of curiosity what would you use it for?
@Kees. I have not decided on the configuration yet bu it is at least three. Two for the wings and one as a failproof. Then the autopilot of choice plugs into that system. If all the sensors (except the high data rate 6dof) are mounted in the network then changing the autopilot doesn't change my sensors. This is good because I need the baro and airspeed sensors out on the wings. The mag can go anywhere but preferably a long way from the 2.5kW motor.
The Swedish word "lagom" translates as "just enough" and is deeply ingrained in the culture. I didn't get off to a good start with lagom by choosing the XXXL. I am not sure this project is helping either.
Matt,
Nice one. Looks like a Swedish pocket-knife.
How many in a Thermik XXXL?
Looking forward to see more.
Kees
-
1
-
2
-
3
of 3 Next