pixhawk / hardware Goto Github PK
View Code? Open in Web Editor NEWPX4 Hardware designs
PX4 Hardware designs
This issue is to track the efforts on a WIP CAN angle of attack sensor.
This is part of the core developers roadmap for PX4:
http://pixhawk.org/roadmap
@nickarsov @pkocmoud The airspeed sensor browns out on boot. The most likely reason is that the power sequencing goes bad during boot.
It only happens when powered through battery, it doesn't happen when powered through USB.
How does the 5V rail plot look like on boot? Please do not offer any judgements, I just want to see the real data. 3 plots:
The documentation shows 4 = GND and 1 = VDD_5 but this is reversed assuming the numbering on the overview diagram are correct.
We can access the HW flow control pins of USART6 on larger packages and hence should see if we can free that port.
Hello,
I have two Pixhawks. One of them doesn't show any RC input. I'm using a Spektrum satellite and have made sure to update firmware on both Pixhawks. Just wondering if there's anything I can do.
Thanks for your time!
Working on this with Phillip K. and Nick. Purposefully reduced to the bare minimum to allow a clear product-market-fit for racer pilots. Has CAN in contrast to small other variants like Pixfalcon. Will fill this with feedback and issues as testing becomes available.
This issue is to track feature requests for a next generation hardware revision.
This is part of the core developers roadmap for PX4:
http://pixhawk.org/roadmap
@proficnc This is to track our discussion about sealed connectors for Dronecode. Feel free to throw in any intermediate results.
The core dev team is currently working on a STM32 based, UAVCAN enabled motor control solution for multicopters. This will be augmented with CAN based sensors, e.g. for angle of attack.
This is part of the core developers roadmap for PX4:
http://pixhawk.org/roadmap
@nickarsov What is the full-scale from 0-3.3V on the voltage and current sense lines? E.g. if the voltage is 3.3V on both lines, what voltage and what current is the sensor seeing?
Note that you need to have the same scaling for both power modules, as your users will get terribly confused else. You only should offer a different scaling range for specific sub groups, e.g. very large airframes.
The Pixhawk 2 bundled 8S power supply unit we got (prototype) has 1.2V on the output. I would like to understand if this a manufacturing or design flaw, and how we move on from there.
The wiring for the Serial/UART is shown on this page WITHOUT GND pins. https://pixhawk.ethz.ch/px4/users/apps/mavlink
On this page it's shown WITH GND pins. https://pixhawk.ethz.ch/px4/modules/wiring#other_setups
Unfortunately the pic with GNS pins is hard to see where the wires go, but on the other pic it's clear where the RX/TX pins go, at least.
I had to check the https://pixhawk.ethz.ch/px4/modules/px4fmu#Connectors page to be sure where the (black) GND pins went into the (black) connector.
Solution: need better close-ups of the wiring, especially the GND/s.
@kd0aij @nickarsov @pkocmoud The voltage and current sense pins on the power connector are inverted vs. the 3DR power module (and most likely vs. your own power module as well).
Correct pinout:
https://pixhawk.org/modules/pixhawk
Your pinout:
https://pixhawk.org/_media/modules/pixracer-rc12-12-06-2015-1330.pdf
Note how voltage and current are swapped.
@nickarsov @pkocmoud @kd0aij I'm tracking this here for now. We can't seem to be using 470R for the backwards direction, we need to use something with much higher resistance, e.g. 1.5K or 2.2K. We need to try it against real hardware.
Can you please attach (just drag images onto the text pane to attach) scope shots of the port?
@pkocmoud @nickarsov Is there something I can help with in terms of the HW test routines? I presume you use something like tests sensors
- what else do you need?
"According to ISO 11898, placing the termination resistor on a
node must be avoided because the bus lines lose termination if the node is disconnected from the bus.”
As far as I understand the s2740vc ESC:s are internally terminated. Doesn’t this mean that they “have” to be placed at the end of the bus?
This is not yet open sourced, but our unit fails with 1.2V
@kd0aij @nickarsov Do we need to connect the 5V signal to the receiver? That might be problematic and we might want to only connect S.PORT and GND.
Hi, I'm trying to make a px4 fmu board for my fixed wing. Now I met some problems with the .sch file. In the latest PX4 FMU 1.7.1, what's the value of RN1 (CTS 742c083)? I can't find this value in this file or BOM. It appears to be a current-limit resistor between SPI3 and MicroSD. I wonder If any one knows something about this ? Thanks a lot . :-)
@pkocmoud @nickarsov APM and PX4 both boot into "debug" mode without the microSD card inserted. So unexperienced users who do not insert a card will not be able to connect to the board.
Which essentially means you need to bundle one, because the product is non-functioning without. Hobby king and 3DR both include a card in their default kits.
We should only expose these once to save board space. They can't be used in parallel, and its better if users get two different cables, which will make it obvious they will have to decide.
We should also work out what the correct cabling to FrSky is.
I've stacked my PX4IO board unto a PX4FMU board. On startup, I get "PX4IO not detected", "PX4IOAR not detected" etc. Is there a step I'm missing? When I run the command "px4io start", I get the message "PX4IO ready" or "PX4IO: already loaded". However, when I choose to update the firmware using "px4io update", I get "error updating PX4IO - check that bootloader mode is enabled".
There have been some issues identified with certain boards (i.e. missing zero-ohm links, poor soldering). It would be helpful if boards were stamped with a batch number of some kind to help identify boards with known manufacturing issues.
E.g. batch number could be composed of a manufacturer (e.g. 3DR) and a production run date stamp (e.g. YYYYMMDD).
With the move to the MPU9250, and the potential for high rate reading. The need for a large isolation system is reduced.
Some form of mechanical and thermal isolation is still needed, but we should consider some options here.
PX4FLOW emmits very intensive electromagnetic noise, so u-blox GPS loses lock in 15cm from it, GPS signal SNR drop is noticable even on 1 meter.
There are two sources of noise - PX4FLOW board itself and UART. Uart is not buffered so all noise from CPU pins goes to UART wires. PCB should be redesigned completely to reduce emmitted noise, maybe to hide all lines between two screen layers on multilayer PCB...
The pinout document says that there is a second USART2 located next to the power module. However, it is actually connected to the UART5 pins. Also RX and TX are swapped relative to the UART5 pinout.
@kd0aij @pkocmoud @nickarsov The QAV250 doesn't include mounting material for the autopilot. So if you order it and a Pixracer you can't mount it. What were you planning in terms of mounting hardware?
All Pixhawk series boards should buffer the RTC RAM. We have now hard fault logging:
[boot] Fault Log info File No 3 Length 3257 flags:0x00 state:1
[boot] Fault Log is Armed
[hardfault_log] Fault Log is Armed
Which means even if the whole OS crashes it will power-cycle, offer the last system state for an in-flight restore and then once the system is back up online write the fault to the microSD for post-flight analysis.
Pixhawk 1 and Pixhawk 2 have the buffer power supply, and so should any other board in the series.
I have an issue with the PX4 not booting when applying power to the quadcopter in either the form of a 3S Lipo or a lab power supply @ 11.1V.
When applying power, the PX4's green LED is on immediately and all other LEDs are off (and never flash during power on). If I press the reset button it boots fine. Alternately, after connecting power, I can unplug the PX4 from the CC BEC (5.1V) during power up and connect it afterwards and everything works fine. Powering via USB works fine.
My wiring works as follows:
Lipo (3S) -> Powerboard -> DJI ESCs
Lipo (3S) -> Powerboard -> CC BEC (5.1V) -> GR-16 Receiver -> PX4 via 5VDD on multiport.
Mike's Response:
This is not expected behaviour; the board should reset cleanly regardless of the power source. It sounds like there might be some glitching or other noise on the rail that is causing the PX4FMU to not boot correctly. If you have a scope and you can watch the VDD3v3 rail you might be able to catch this, but it may be tricky.
If you repeat your unplug/replug test but disconnect the BEC from the powerboard (i.e. so you are re-powering the BEC) which result do you get? Which CC BEC are you using?
Can you repeat the test with the ESC's out of the circuit? I realise that if you're using the DJI setup chances are that they're all soldered in, but if you can just power the BEC/receiver/FMU setup from your bench supply that would be a good start.
One more test; while the board is in the green LED only state, if you connect USB, what happens? In particular does the board boot? If not, do you see the ST bootloader on the USB bus, or is it totally silent?
Apologies for the problem; any information you can provide here may help us narrow down the issue, and I will try to reproduce your setup here as well.
Spektrum connector VCC should be switchable under program control. With the current px4io version, the only way to implement bind support for DSM satellite receivers is by routing the VCC line though one of the available relays, thus requiring the modification of the receivers cable and tying up one of the relays. VCC could easily be controlled with a bitport and possibly a inverting buffer since the receiver's power needs are relatively low.
The black sonar variants seem to have supply challenges.
@pkocmoud @nickarsov Its not clear what your customers get with each Racer. What exactly is in the box? Which list of cables? Mounting material (the racer frames do not come with any)? Foam?
The hysteresis time might be too long, leading to the autopilot seeing excessive voltages. Since no test results are available, I will do this myself.
Thanks much for PX4FMUv2.3.sch and PX4FMUv2.3.brd, it would be very helpful and "Open Source" friendly if you guys shared the BOM (Digikey or Mouser part numbers for example).
Gerbers can be generated, however it would be nice to have known good Gerber files on hand as well.
Please see the example in the OpenPilot CC3D and the associated "Open Source" files that include BOM and Gerbers.
http://git.openpilot.org/browse/OpenPilot/hardware/Production/CopterControl%203D/Gerbers/
http://git.openpilot.org/browse/OpenPilot/hardware/Production/CopterControl/Gerbers/
@pkocmoud @nickarsov @kd0aij I'm measuring 1.8V output on the GPS power port with just the airspeed sensor attached. Can you confirm that? I'm measuring 0V with nothing attached.
Right now we only have a serial port dedicated to a OSD, but we haven't a default / recommended setup or instructions to help users getting started.
The housing for Pixhawk 2 is an integral component to to design. Where do we stand on its inclusion as a PX4 design? The foam as well...
Since manufacturers are also doing power modules and ESCs it would be good if we could identify a low-cost automotive type we can recommend. I think the drone industrial space will need the same kind of standardisation as automotive needed.
For now just looking for recommendations, its too early to try to standardise as we hardly understand the needs and market.
@proficnc Feedback welcome!
Some common types:
@nickarsov The connection to PB0 should be straight from the raw and non-inverted signal (RC_INPUT), not from behind the inverter.
Independent on how we do the new inverting circuitry that pattern will allow Spektrum binding.
This post is misguided in assuming this would be an ok approach to fix it, but the analysis is still correct:
http://diydrones.com/forum/topics/solution-proposal-for-pixhawk-imu2-related-bad-accell-health?xg_source=activity
The PX4 airspeed sensor breakout is using a 5V sensor for a 3.3V bus, which is not compliant with the I2C spec and leads to thousands of read errors in a short time frame. It can be fixed in two ways:
Scope shots and further discussion here:
PX4/PX4-Autopilot#3544 (comment)
We should identify the default part.
Active discussion on Dronecode mailing list.
@pkocmoud @nickarsov @kd0aij This is to discuss schematic revision 13.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.