Sorry for the delay. I'm glad there is interest and talent out there. Another component of the effort is to produce fuels locally. Please check out the following link; http://www.epa.gov/otaq/07-meoh.htm Minnesota has a huge methanol fuel production capability (another process control ap.), the President visited our state promoting fuels development, the US EPA promotes it, and we are entering a war against Saudi expatriots on a battlefield of their choosing. But the Minnesota politicos fight it. The way it works is; the politicos clobber me, claim their new great idea, muster a ton of money, link with big corporations, and hire people like you. After over twenty years in this game, I accept that a scientist advances science: business(wo)men, lawyers, and politicos grab the money and glory. My goal is to prepare the infrastructure and escape with my hide. >>>>>>>>>>>>>>>>>> Original Message <<<<<<<<<<<<<<<<<< On 9/21/01, 9:38:26 AM, Daniel Taylor <dante at plethora.net> wrote regarding Re: [TCLUG] Linux controller for fuel cells; progress!.sdm: > On Fri, 21 Sep 2001, Rick Engebretson wrote: > > Some top politicos are now urging the Gov to reach out to fuel cell mfrs. > > Ford was mentioned, with a potential billion dollar investment. I > > mentioned this group's talent pool for developing a controller (forgive > > me). > > > > Some serious work is being done on embedded Linux systems. Real mode > > kernels are being used. Most development is focused on small appliances. > > A large process controller would not be so constrained. > > > > My feeling is; > > > > 1) Use protected mode, multitasking Linux PC to "manage" > > > > 2) multiple real mode Linux x86 add-on cards. > > > > Quite a variety of x86 embedded cards exist. Some are complete 486 > > computers on a chip with full I/O, memory, etc., and plug into ISA slots. > > Perhaps DOS could be used for the embedded card? Lot's of questions! > As one of the available embedded Linux programmers on the list, > I really appreciate this. > OpenDOS is a good choice for some embedded systems. But for systems > where a PC104 is acceptable cost wise, Linux provides a better > infrastructure layer. Good proprietary solutions include QNX and > VxWorks. In fact, QNX is pretty much best-of-breed IMO, and worth > the money if you don't mind a proprietary solution. > Specificly WRT your comments above: a single processor of 486/100 > power and efficient software is capable of controlling an amazing > quantity of hardware. For large configurations (say coordinating > the activities of several fuel cells and microturbines) a slightly > more powerful system (say P5/200) capable of running a sophisticated > user interface and dynamic control software may be desirable. > Daniel Taylor