From:	Wayne Radochonski [rawa@mbari.org]
Sent:	Wednesday, February 20, 2013 10:02 AM
To:	Andy Hamilton; Gene Massion
Cc:	Jim Birch; Ken Johnson
Subject:	Wayne's Status Report for February

Andy, et. al 
 
History 
 
PC-104 Carrier Board for ESP 
 
A design review was held for the PC-104 carrier board.  The reviewers decided that the design 
was ready for  
fabrication.  As a group, we also decided to add Brent's much requested audio interface. 
 
Due to my vacation and the request for audio, fabrication of the PC-104 will start on April 1st.  
This will  
give the PCB layout contractor enough time to finish the layout.  First article is expected mid 
April.  I won't  
start bring up until after the last Power Buoy deployment. 
 
Coastal Profiling Float Embedded Controller 
 
UART interrupts now appear to be working correctly.  The last two weeks have been extremely 
frustrating  
dealing with C/C++ library size and stack overflow issues. 
 
Although the GNU tools are very capable (and free!) for software development, their C/C++ 
libraries assume  
that you are running on Linux (and have plenty of storage).  NewLib and ucLibC are GNU 
libraries intended for use  
on Linux and embedded systems, respectively.  ucLibC has not been ported to the STM32F407 
and Code Red. 
 
Nearly all of the devices used in the buoyancy engine prototype are driven by character strings 
through a serial  
port.  The WiFi driver, in particular, makes heavy use of the C library functions "sscanf" and 
"sprintf".   
 
Using any "printf" related library functions pulls up almost 500K of code (95% unused) into the 
image.  Given  
that the processor has1MB of flash, the code bloat isn't of much short term concern.   
 
However, under certain circumstances use of these library functions will cause a stack overflow 
on the STM32F407  
processor.  Tracking down the stack overflow problem took several days to diagnose as I was 
left with a CPU with  
its paws to the sky and not many breadcrumbs left to determine why the CPU was executing 
some strange code in the  
middle of an unrelated and not even called member function. 
 
Given that NewLib is 200K lines of code, fixing it seemed futile.  Porting ucLibC (60K lines of 
code) to our tool chain  
and testing it in a reasonable amount of time also seemed like a daunting task. 
 
I punted on fixing NewLib or porting ucLibC (for now, at least) and wrote my own routines for 
"sscanf" and "sprintf" as  
well as a simple C++ string class. 
 
I am just about to start testing the WiFi driver with my own "string library".  Total code size is 
about 20K for all of the  
Coastal Profiler code. 
 
Power Buoy 
 
Team Power Buoy decided that I should re-enable the software that drives the make up pump in 
the "downstairs" PTO  
software.  We also discussed how to check and re-work some soldering issues on the PTO 
controllers for chips with  
exposed ground pads. 
 
Mystery 
 
I'm still investigating tools and processes for code testing and inspection.  Hopefully, we can 
find a middle path between 
increasing the development effort and quality.  The LRAUV project licensed Coverity's static 
code inspection tools.  I 
am investigating, if those tools could be used on the Coastl Profiler project without violating 
the terms of our license. 
 
Ballast 
 
I attended a "free" (as in vacation rental pitch "free") ORCAD training in Milpitas.  I learned 
about all of Cadence schematic  
capture and PCB layout tools (not needed).  I even managed to hook up a few library 
components with ORCAD schematic  
capture.  I have much more to learn before I am even 10% as productive with ORCAD as I am 
with DesignWorks.   
 
I will be on vacation from 2/25 through 3/8.   
  
*** Start of the statement for the group *** 
 
Wayne held a design review for the PC1-104 carrier board to be used by ESP.  The reviewers 
approved fabricating the  
board.  Wayne continues to develop "bare metal" software to support a WiFi spigot or the 
Coastal Profiler. 
 
*** End of the statement for the group *** 
 
Wayne
"The whole problem with the world is that fools and fanatics are always so certain of 
themselves and wiser people so full of doubts." -Bertrand Russell