On Sun, Dec 23, 2007 at 07:52:11AM -0800, Andrew Greenberg wrote:
> Hi Mark,
> 
> > The PARTS (http://www.portlandrobotics.org/) folks are kibitzing around
> > doing a robo-Magellan http://www.robothon.org/robothon/robo-magellan.php
> > kit.  The design that folks are gravitating toward is to put together a
> > platform that has a 6DOF IMU + 3 Axis Magnitometer + the typical robot
> > related GPIO's and peripherals interfaced to a micro controller type of
> > processor which is then interfaced via USB (full speed bulk) to a user
> > provided USB host (laptop, Eee, OLPC, PC104, NSLU2...)
> > 
> > I'm thinking we could join forces.
> 
> This is, as always, a great idea and we'd love to do it. But usually,
> our avionics system ends up begin a little to do weird for use with a
> on-land robotics system. Flying robots with rocket engines and wheeled
> robots just end up having very different electrical systems :)

Yeah, I know.
 
> Our "generic" LPC2148 node was designed a year ago or so by a capstone
> team. They did a great job, and the complete EAGLE schematics along with
> documentation:
> 
> http://psas.pdx.edu/CapstoneLV2bProjectReport/
> 
> Since then we've written some peripheral libraries for the LPC2148, but
> unfortunately we haven't laid out the board yet. That's also on the
> books for the next few months.
> 
> I think there's a lot to be shared with sensor design:
> 
> - Whatever GPS we end up using can certainly be shared, although it
> might end up being too expensive for most robot use.
> 
> - The 3D magnetometer sensor can certainly be re-used.

ack.

> 
> - The IMU sensor design might be good thing to share as well. But our
> dynamic range and bandwidth is WAY bigger than what most robots will need.

ack

> 
> But, there are differences:
> 
> - We have a very specially power distribution system. Instead of using
> the standard USB 5V supply, we directly distribute a 14.4V Li Ion
> (probably polymer) battery pack, and have switching supplies on each
> board. For justification for that, see the Capstone report.

The for the robot we would try to have the sensors / IMU be USB powered.
 
> - We're very concerned about volume and weight, which means small,
> packed, four layer boards. Since we have some free use of Screaming
> Circuits' facilities, we'll probably end up using > 2 layer boards with
> enthusiast-inappropriate chip packages like uBGAs and such.
> 
> > What is the status of the LPC based IMU design and firmware?  Where is
> > it documented on the PSAS pages?
> 
> The IMU in the planning stages right now. We'll probably pick the
> sensors in the next two weeks (see http://psas.pdx.edu/SensorComparison/
> ), and begin the design shortly thereafter. All of the LPC nodes are
> essentially vaporware right now, and we plan to fix that this winter :)

Oh.  That should be fun.

> 
> If PARTS picks the LPC2148, then I think there's *lots* to be shared
> back and forth. It probably won't be board design, but I bet we could
> share sensor design and a lot of low-level firmware.
> 

I was thinking this too, it would be nice if both used common micro
processors.

> I'm not sure what the PARTS timeline is like, but we hope to be done
> with the generic node this spring and be working on sensor design
> throughout the summer and have a complete package sometime later in
> 2008. If PARTS is in more of a rush than that, then we may be borrowing
> designs from you, rather than vice versa :) But hopefully this is
> something we can collaborate on!
>

Well timelines are a function of who you ask, but some folks are
thinking of building something by the end of this summer.  I'm not sure
I like talking time lines.

--mgross 

Attachment: signature.asc
Description: Digital signature

_______________________________________________
psas-avionics mailing list
psas-avionics@lists.psas.pdx.edu
http://lists.psas.pdx.edu/mailman/listinfo/psas-avionics

Reply via email to