Modular AT90CAN128 avionics for AUSROC2.5

[Luke Weston]

Hi everybody,

Thanks to John and Andy for the discussion on Saturday, I thought it
was valuable and productive :)

Here are some ideas of mine, to think about and share and discuss.
Maybe the ASRI guys don't agree, but you know, we'll talk about it :)

- Most of the peripheral hardware devices on AUSROC2.5 such as the
chamber pressure transducer, temperature sensor on the LOX valve
maybe, pressure transducer on the He/Tridyne pressurant system, engine
valve controller, etc, will be modules that talk to the CAN bus, and
they will be all based on an identical microcontroller, probably the
Atmel AT90CAN128 AVR microcontroller with integral CAN interface. This
is the same microcontroller used on those Olimex development boards.

- The AT90CAN128-16AU costs $11.07 USD in single quantities from
Digi-Key, which is about what you'd expect. It's not abnormally
expensive.

- All of these modules will share a common connector and pinout for
that connector and a standard specification in terms of what voltage
rail or rails they need to be supplied with and on what pin on the
interface connector that voltage rail(s) is supplied, and what pins on
that interface connector are used for the CAN-H/CAN-L differential
pair.

- The different modules will share the same AT90CAN128
microcontroller, and the same CAN interface connector, and the same
power supply pinout, and the same ISP and JTAG headers (note that this
chip has a JTAG interface) but each board will include the specific
hardware attached to the microcontroller to do the specific job that
it has to do - for example an instrumentation amplifier for a strain
gauge, or a fast external ADC connected to a pressure transducer plus
appropriate power supply wiring or an appropriate amplifier for that
specific pressure transducer, or power FETs to fire pyros, or a GPS
receiver module or whatever it is for that particular module, designed
and engineered to do the job that it's specifically intended to do, in
the smallest, cheapest, most compact system, without sacrificing
reliable performance.

- Each module will include a Microchip MCP2551 CAN transceiver between
the CAN interface on the microcontroller and the CAN bus differential
pair. This is the same chip as used on the Olimex AT90CAN128
development boards between the microcontroller and the CAN bus, and
it's also the same chip used on the boards I designed to control the
Rutex boards, which on those boards sits between the MCP2515 CAN
controller (which is not integrated inside the main ATmega328
microcontroller, unlike the AT90CAN128 boards) and the CAN bus.

- Standard ISP header and JTAG header and TTL UART header (to suit
FTDI cable) with consistent standard pinouts on every hardware unit.

- Maybe if the specific hardware for a particular module doesn't use
up all the microcontroller pins then maybe every microcontroller
module might have a micro-SD card on board with the microcontroller,
since they're small and lightweight and inexpensive? That way there is
a lot of distributed, redundant "black box" memory for every little
instrument throughout the vehicle which could be useful in case of
catastrophic failure of the vehicle.

- This standard specification for the interface mean that different
people and teams and universities across the country might design and
build different avionics modules and they can be shared and swapped
between different groups and they just plug together using the same
cables and connectors and test interfaces and programming connectors
and all that sort of thing, making collaboration much easier.

(I'm talking exclusively in terms of hardware here although of course
the same sort of philosophy applies to software as well).

- One of the modules we might want to design and build is a
computer-to-CAN interface, consisting of an AT90CAN128 microcontroller
plus a USB socket and USB-to-TTL-UART chipset, so you can just easily
plug it into USB and have an interface to send and receive traffic
from the other CAN modules for testing and development and
experimentation.

Also, you might find this to be interesting and well worth reading.
Bdale Garbee was involved significantly with the team of AMSAT people
who developed/develop this:
http://can-do.moraco.info/Default.htm

:)

Cheers,
Luke

[philip thomas]

 A well thought out plan. The SD card idea, will they be hard wired into the boards or by socket?

The only concern is that they will dislodge due to vibration during launch.

Phil Thomas (newbie)

> Date: Sun, 17 Jun 2012 17:25:34 +1000
> Subject: Modular AT90CAN128 avionics for AUSROC2.5
> From: reindeer...@gmail.com
> To: lunar-...@googlegroups.com; john....@mail.optusnet.com.au; an...@geekscape.org

> --
> You received this message because you are subscribed to the Google Groups "Lunar Numbat" group.
> To post to this group, send email to lunar-...@googlegroups.com.
> To unsubscribe from this group, send email to lunar-numbat...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/lunar-numbat?hl=en.
>

[Luke Weston]

Personally, with respect to micro-SD cards, what I would do is design the hardware to use standard surface-mount micro-SD card sockets on the PCBs, and just plug in cards during development and testing. Then when it comes to an actual flight, plug in an SD card into the socket and cover the whole card, in its socket, with epoxy or something like that so it's not going anywhere.

This then requires a hardware/software mechanism to download the stored data, either through wireless telemetry or plugging in a cable, without removing the card.

Alternatively you might have some mechanical mechanism screwed or bolted to the PCB over the SD card socket which mechanically holds the card in place tightly so it cannot move, but it can be unscrewed and taken apart. This might be as simple as a small screw through a hole which is right up against the back of the SD card (the end that pokes out of the socket) once it is inserted in the socket, holding it in against the socket.

Cheers,

  Luke

[Luke Weston]

Another factor I would consider, and likely has already be considered by people associated with ASRI, with regards to the AUSROC2.5 avionics is that the GPS receiver should not be like the common cheap hobbyist module type with an integrated patch antenna.

The GPS receiver chip should have its antenna bought off the board onto an SMA connector or similar which connects to an external antenna on the outside of the vehicle, something like a quadrifiliar helix probably, so that it's fairly isotropic and omnidirectional.

Cheers,

  Luke

--

You received this message because you are subscribed to the Google Groups "Lunar Numbat" group.

To view this discussion on the web visit https://groups.google.com/d/msg/lunar-numbat/-/XR3MBAtpe6sJ.

[Luke Weston]

Hi everyone,

Maybe have a look at this and tell me what you think. It's just a rough starting point to get the conversation going and get ideas flowing and shared, basically - I hope it's valuable or helpful.

https://github.com/lukeweston/CANAvionics

Schematic is there, finished PCB layout is not done yet.

There are a few different peripheral devices incorporated on the board, in addition to the AT90CAN128 and the CAN transceiver IC.

- MPXV7002 differential two-port pressure sensor for connecting to a pitot tube etc. (Relatively low differential pressure, not for use at very high pressures eg. measuring He pressure.)

- Two MOSFETs to provide the capability to fire pyrotechnic bridgewire type devices. Also, open-circuit detection capability is provided, so the system can tell that there is an electric match etc. connected (assuming that it has a very low resistance before ignition) and the safe/arm circuit is closed - so if the wire is broken or something like that you're aware of it.

- Header to connect a GPS receiver.

- MicroSD card for redundant local data storage at each distributed node. Could be soldered, glued, potted etc. for mechanical security if needed.

- Interfaces for two K-type thermocouples for temperature sensing.

- Sensor on-board the board for local temperature measurement and atmospheric pressure measurement.

- Voltage regulator which can accept a flexible 7-24V input, whatever is supplied over the CAN wiring loom and regulate it down to 5V locally.

Not all these devices need to be assembled on any one particular board. For example, if you only want the differential pressure sensor and one thermocouple, for example, in any one particular node, then you only need to assemble those components. But the same PCB design can be re-used for multiple different nodes around the vehicle which perform different functions, with different components and sensors mounted to meet the different requirements at different parts of the vehicle.

There are a few sensor capabilities which it might be desirable to add which I have not yet added - for example support for reading the output from strain gauges (appropriate instrumentation amplifier) and the high-pressure pressure transducers for tank pressure / He / Tridyne.

Cheers,
  Luke