Olimex AVR-CAN development boards

84 views
Skip to first unread message

Luke Weston

unread,
Apr 16, 2012, 11:40:09 AM4/16/12
to lunar-...@googlegroups.com
Hi everyone,

I'm sorry that I was unable to make it to the meeting that John August and Andy and others (?) had a couple of weeks ago.
Was there anything important of note that I missed that I might be able to catch up on via email?

I see that we have that pair of Olimex AVR-CAN boards that John provided. I will bring those over when I catch up with Andy on Tuesday and give those back to him.

These boards basically consist of an Atmel AT90CAN128 microcontroller on a small board with an RS-232 serial interface on one end of the board and a CAN bus interface on the other end of the board. (The female DE-9 is the RS-232 and the male DE-9 is the CAN.)

(I know RS-232 is a bit anachronistic and hard to work with on most modern PCs... I personally wouldn't choose RS-232 for any modern hardware design unless it specifically had to be connected to some old legacy RS-232 hardware.)

Lots of documentation on these boards is available here. Schematics, manuals, pinouts, software examples, you name it, they're pretty good with their documentation:

http://www.olimex.com/dev/avr-can.html

The good news is that the pinout of the CAN pair and power rails on the DE-9 CAN connector are compatible with my CAN boards, so you can plug them together and they should work. :)

What I don't really understand, and I'm hoping I can have it explained to me, is what exactly is the intended use-case for these boards?
How do they fit into the design and development of the overall system?

Is the flight computer in AUSROC2.5 or a similar vehicle going to talk to RS-232 ports, which have a couple of these boards attached to them to convert them to CAN interfaces?

Or are these intended more for just testing and development of CAN devices, replicating the flight computer's CAN interface on the ground and sending data over the CAN to the throttle controllers, as a testing/development platform?

I assume it's mainly the latter.

However, there is one little problem here.

My boards were designed to have 24V supplied over the power rail (pin 9) on the DE-9 connector. This was intended to be 24V because the specified supply voltage required by the Rutex R2020 boards is 24V. My boards regulate the 24V (or less) down to 5V on board using a switching regulator, which has negligible power dissipation.

The Olimex boards also use pin 9 on the DE-9 CAN connector for their unregulated power supply rail. However, they regulate it down to 5 volts on board using a LD1117 linear regulator IC, and their official documentation specifies 12 V as the maximum supply voltage. (24V might work, but it is outside the official specs, and the little regulators are likely to get very warm.)

We can overcome this by having the power pins (pin 9 on the DE-9 connectors) connected to the cable that is connected to my CAN boards NOT connected to the DE-9 connector on the Olimex board, and having this wire connected to a 24V power supply, and having a separate power supply (9V-12V or so) connected to the power pin on the DE-9 connector (or one of the other power supply headers) on the Olimex board.

There is a little cable with a 2-pin female 0.1" header connected to a 9V battery snap connector for this purpose which was included with this kit of boards and cables from John.

Connecting all the devices on the CAN bus to a single power rail on the CAN cable and connecting it to 24V runs a good risk of cooking the regulators on the Olimex board(s), but if you reduce that voltage below 24V then you're no longer supplying the specified voltage supply to the Rutex R2020. So you've got to split it into two different power supplies. (Plus the 50V supply for the motor drive H-bridge supply on the Rutex R2020s as well, which is another separate thing.)

Keep in mind that if you're working in a particular development/testing context where you only want to talk over the CAN interface from the Olimex CAN board to my CAN boards and you're not running the Rutex board and the motor at that particular time, then the 24V supply doesn't matter at all, my board in and of itself is perfectly happy with 7V-12V or so (regulated down to 5V), so you could use the same power rail supplied via the CAN cable to supply the Olimex board(s) and my board(s) with a single 7-12V rail in one hit and everything would be happy.

Keep in mind that if there are 3 or more devices on the CAN then the terminators should be disconnected for the devices "in the middle" and only connected for the two devices at the end of the bus. Both my boards and the Olimex boards have semi-permanent solder-blob jumpers on the board for this purpose of including or disconnecting the CAN terminator resistor.

I hope this is valuable.

Cheers,
  Luke




Andy Gelme

unread,
Apr 16, 2012, 9:01:53 PM4/16/12
to lunar-...@googlegroups.com, john....@mail.optusnet.com.au
hi Luke,

On 2012-04-17 01:40 , Luke Weston wrote:
> I see that we have that pair of Olimex AVR-CAN boards that John provided.

I hope so ... given that *you* have them :)

> What I don't really understand, and I'm hoping I can have it explained
> to me, is what exactly is the intended use-case for these boards ?

Initially, prototyping of the communications protocol and ensuring that
we can establish communications between the Rutex motor controller
boards designed by yourself ... and anything that John August / ASRI
might do for AusRic 2.5.

> How do they fit into the design and development of the overall system ?

Initially, prototyping. Beyond that ... best for John August (CCed
above) to answer.

> Or are these intended more for just testing and development of CAN
> devices, replicating the flight computer's CAN interface on the ground
> and sending data over the CAN to the throttle controllers, as a
> testing/development platform ?

Yes.

When we get to this stage (controlling a running motor via the remote
CAN interface boards), will probably make a specific CANBus cable that
doesn't interconnect the two separate power supplies.

> Keep in mind that if you're working in a particular
> development/testing context where you only want to talk over the CAN
> interface from the Olimex CAN board to my CAN boards and you're not
> running the Rutex board and the motor at that particular time, then
> the 24V supply doesn't matter at all, my board in and of itself is
> perfectly happy with 7V-12V or so (regulated down to 5V), so you could
> use the same power rail supplied via the CAN cable to supply the
> Olimex board(s) and my board(s) with a single 7-12V rail in one hit
> and everything would be happy.

This is a good way to begin protocol development / testing ... prior to
using a running motor.

> Keep in mind that if there are 3 or more devices on the CAN then the
> terminators should be disconnected for the devices "in the middle" and
> only connected for the two devices at the end of the bus. Both my
> boards and the Olimex boards have semi-permanent solder-blob jumpers
> on the board for this purpose of including or disconnecting the CAN
> terminator resistor.

This won't be a problem for the current goals, e.g having one of John's
boards and protocol controlling one of your motor control boards.

--
-O- cheers = /\ /\/ /) `/ =
--O -- http://www.geekscape.org --
OOO -- an...@geekscape.org -- http://twitter.com/geekscape --

Reply all
Reply to author
Forward
0 new messages