Re: [Paparazzi-devel] Introduction, Q's about STM32 development

[Jake Stewart]

Thanks to everyone that responded!  I didn't really know if the list would be 
active and friendly or not... It sure is!

I'll try to respond to everything in this post.  If I forget something feel 
free to hit me with an email.

To Christoph N.... I'd love to get all sorts of boards and play with them.  
I've kind of started with my ultra-cheap dream components and figure I'll play 
with them first.  I had figured on scratch building most of the system until I 
really looked into how much had already been done.  I saw a lot of autopilots 
out there and more cropping up almost daily.  I really didn't like any of them, 
but was drawn to the PPZ platform since it seems closest to my instincts on how 
to do things.  I was almost ready to order a board when I found the RoboVero 
board.  It uses a similar processor (LPC) and would probably be a minimal 
effort port.  For $99 with the RoboVero you get a nice dev board with a built 
in IMU!  It seemed really similar to the PPZ boards with most of the same 
functions available, and cost a lot less than PPZ board + IMU.  So I thought 
I'd try and get PPZ running on that.  I still think it's a pretty good idea, 
especially with the ability to connect easily to a Overo Gumstix computer.

Somewhere in that timeframe I ordered a fair bit of gear and don't actually 
have any funding available yet, just personal. I also started liking the ST 
accel/gyro chips back when I was just starting.  The ST chips have a prebuilt 
library running on STM32s for sensor fusion.  Maybe it's a lot of hype, but my 
math skills are not up to the programming of that sort of thing (didn't do well 
years ago in calc).  So I really have been jumping around with ideas.  In the 
end I figure I better start playing with the stuff I already have before I get 
married to a platform.

Anyways, I've been playing with the STM32 VL Discovery board and the more I 
study it the more it looks like a reasonable canidate.  It's a little over 
twice the size of the LISA/M and has less memory and a slower processor speed.  
That's why I'm hoping someone can definitively say weather compiling the 
firmware for this processor/board is feasable.  

I'm just not sure what these boards actually do.  They essentially seem to be 
small form factor processor boards with the proper pins tied to convienient 
connectors on the board.  I haven't noticed that there's really much in the way 
of specialized hardware that I couldn't do without.  I'm using all 3.3v, so no 
need for level shifters and so on.  I'm happy making my own connectors from the 
header pins for whatever I need.

For all I know, all that mgiht need to be done is reassign some pins to make 
the connectors easier and a few minor STM32 F105 -> F100 conversions.  Then 
again the different clock speeds might screw the timing all up and so on.  I'm 
also not against using two Discovery units in the plane if need be.  Might try 
to use one for IMU sensor fusion and one for the rest.


Chris G. asked about my sample collection idea...  Hopefully there's no 
competing microbiologists out there, I think I've come up with a fairly simple 
collection system.  The system involves one or more sample pods.  Each pod has 
a small electric motor with a pully/spool attached. From the spool a line is 
attached and loose coiled below the motor.  The bottom of the pod has a servo 
actuated door or a simple push solenoid to kick off the bottom pod cover, I'm 
still working on that out since I'd like to use autoclavable (polyproplyene) 
micropipette tip containers for the pods.  The rope has a weight (fishing 
sinkers) at or near the bottom and a sample collection apparatus which is 
essentially a piece of tape or agar coated fabric or any other manner of sticky 
material which will pick up a sample.  Water samples would be done with a 
sponge.

The idea is that the door kicks open and the rope falls to the ground where it 
is drug a short distance where particles stick to the collector, then the motor 
kicks in and winds the sample back up into the pod.  Then the plane returns to 
base.  Using variations of that system should make it easy to snag lots of 
samples from remote locations for cheap.  Obviously the danger is that the 
sample rope gets tangled and the plane gets dashed into the ground at 30+ mph.

So I do need cheap units.  Not only will some portion of flights crash in 
remote locations, but the crashes would probably also be at fairly high speed 
at a right angle into the ground.  Additionally I'll need some fire support 
aircraft around to put out any fires that crashed piles of electronics might 
create.  Has anyone done any work with fire support aircraft?  I see that as a 
problem unless I can make the aircraft in a fire proof manner.  Lipos can 
really burn hard under the wrong conditions.  There are probably some solutions 
like letting the string go manually by having a weak link or electrically, but 
I've got to be prepared regardless.

Can anyone tell me what the memory requirements are for the STM32 firmware?  In 
a few days I'll get my linux setup working (looks like BT3 is Debian based) and 
figure out myself, but I'm not sure how to guage the working memory 
requirements without running it.

Thanks to everyone for the advice!  I'll probably play with my hardware until I 
get frustrated enough to buy a pre-fab board, it's just my nature.  I'll be 
sure to let the list know how that goes.


-Jake 
 



> ----- Original Message -----
> From: Christoph Niemann
> Sent: 01/20/12 03:35 AM
> To: address@hidden
> Subject: Re: [Paparazzi-devel] Introduction, Q's about STM32 development
> 
> Dear Jake,
> 
> since you didn't do anything with paparazzi at all, I would recommend you
> to use something, that has been used with paparazzi before, because that is
> challenging enough. What about using an established or at least tested
> Autopilot first like TWOG or Umarin? I calculated the costs for an Umarin
> with 120-130€.
> Please don't get me wrong, I don't want to discourage you, but once you are
> used to the System and you have an overview about how to build a new AP,
> you can Still invent new, fancy APs as well.
> 
> Cheers
> 
> Christoph
> 
> 
> 2012/1/19 Jake Stewart <address@hidden>
> 
> > Hello Paparazzi list members! I'm a microbiologist from Idaho (USA). I
> > have a basic electronics background with some experience with RC flying and
> > a little bit with Atmega8 programming. I'm very interested in using UAVs
> > to collect unique biological samples from inaccessable areas like remote
> > areas, high altitude areas, open seas, arial samples, etc..
> >
> > Your project is amazingly impressive! So much so that it is pretty
> > intimidating to a non-expert such as myself. I have a few questions about
> > where to begin. I've done a fair bit of reading on the project, but
> > perhaps need some advice on getting started. Feel free to let me know if
> > I'm out of line, or reply off-list, or ignore any or all of my questions if
> > it's something I should be able to figure out myself.
> >
> > I am currently playing with STM32 VL Discovery boards from ST. I got a
> > handfull of them from Arrow for the insane price of $2.20 each! Link and
> > datasheet...
> > http://components.arrow.com/part/detail/49099771S8561546N7728
> >
> > http://download.siliconexpert.com/pdfs/2010/11/13/7/31/8/620/st_/manual/cd00277245.pdf
> >
> > It's essentially a STM32F100B chip with all the pins exposed on headers,
> > on a board with a STM32103 for USB programming support, a couple lights and
> > buttons, and with all the voltage regulators and whatnot that you'd expect
> > for a simple dev board. I'm curious what I could do with such a cheap
> > board. I figure that they're essentially disposable for all intents and
> > purposes, so no worries about going soldering iron crazy and trying to
> > prototype some interesting circuits.
> >
> > I'm curious about the extent that the code is tied to the hardware...
> > What hardware features are there beyond a proto/dev board that are
> > necessary for the project?
> > Would it be possible to implement something on a cheap board like the one
> > I mentioned?
> > Is the code dependent on any of the hardware on the LISA boards?
> > What are the memory requirements for the firmware?
> >
> > I ask because I'm currently using Atollic in Windows with my board and
> > haven't yet set up a linux environment/toolchain to work with. Any tips or
> > pointers to info on that would also be much appreciated, although I think I
> > saw some info on that somewhere. My only linux install at the moment is
> > BT3. Should I try with that or install debian/ubuntu?
> >
> > I reallize these are all pretty noobish questions, but I just wanted to
> > introduce myself and see if I couldn't make a contact or two who have done
> > similar work or could easily give me some answers.
> >
> > I plan to put some serious work into this project and think I have most of
> > the tools needed. Besides my $2.20 dev boards I have soldering irons,
> > lot's of spare components, a few prototype boards, two o-scopes,
> > multimeters, etc.. For the project I've gotten a 9DOF IMU (ST chips), a
> > few pressure/barometer sensors, a differential pressure sensor for an
> > airspeed tube, a Fastrax GPS module, and a couple generic RF serial modems
> > (sub-G TI chip based). I also have an electric plane, laptop, 3 small
> > cameras, and a 900mhz video transmitter. As far as I can tell I have
> > everything to build a prototype UAV and just need to wire everything
> > together and program it! Easy right? (kidding)
> >
> > If running on different hardware platforms is not something easily done
> > with the project I'm hoping I can put together a more basic program running
> > a subset of the normal features. One reason I'm thinking of this is that I
> > need to do things on the extreme cheap side as I don't have a lot of
> > confidence that a very high percentage of my planes would actually be
> > returning due to the dangerous nature of what they would be doing. Maybe
> > I'm wrong there, but I don't envision a whole lot of my hardware returning
> > to me and I'd be happy with a 50/50 crapshoot if I can keep the hardware
> > within budget.
> >
> > At the moment I'm hoping to use the components I've gotten for testing.
> > Not sure I really even need a radio.
> > STM32VLDiscovery board - $2.20 (arrow)
> > 9DOF IMU - $50 (pololu.com)
> > Fastrax GPS - $28 (mouser)
> > Cessna 182 (electric, 1.4m wingspan) - $130 (hobbyking)
> > --------------------
> > Total $210.20
> >
> > So if I can get any sort of reasonable return rate (many missions would
> > probably be pretty safe) I should be able to get samples for fairly cheap
> > (<$500). That should be much cheaper than collecting them by hand and open
> > up a lot of areas that simply aren't accessable. Am I crazy to think this
> > might be possible?
> >
> >
> > -Jake
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > _______________________________________________
> > Paparazzi-devel mailing list
> > address@hidden
> > https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
> >