[Date Prev][Date Next][Thread Prev][Thread Next] - [Date Index][Thread Index][Author Index]

Re: Flight Computers



Hi all,

On Mon, 29 Apr 2002, Phil Karn wrote

[...]
 
> Given an 1802 CPU, IPS is clearly the way to go. It has proven itself

[...]

> But the number of people on this planet who are *really* good at
> writing IPS code can still probably be counted on the fingers of one
> hand. Writing IPS can be an arduous process, and there's never any
> such thing as a "minor" change when one teensy mistake, such as not
> keeping your thread "stack neutral", can be fatal to the whole system.

[...]

I know Forth ( i belive that IPS is similar to Forth ) and it isn't so 
hard to write and debug software. You can do very atomic operations like 
redefinig words without need to touch or recompiling other portions off 
the system. Its still interpreter like basic. And more, one of my 
private / for fun project I build on clone off 8051 worked more then year 
on the top of mountain. I sit on the front of my computer and changed alost
entrie operation system while CPU is located about 1500 km from me. i8051 
has brain demaged memory map and Forth isn't perfect operation language for 
this type od CPU but it worked. System is only about 9k + 32k ram for 
user words and data.

> And times have changed. The 1802 is starting to get scarce, but much
> more advanced CPUs are now available to fly in space. There are now

[...]

Yes, I agree. 
In the same project I must change CPU module to Advantech 486 class CPU 
becouse I need encryption and more power. Finaly we instaled Linux in 
flash disk ( 64 M ), and now we have VPN tunnel, ssh, snmp etc. It's still 
is working a top of the same mountain. But we have more power consumption 
and more heat problems in the summer. 

Now I'm in procces of writing new, operation system for it. I know that is 
crazy, but I need something like IOS from cisco. One monolityc 
file to upload. No more mess with many small programs. I still need 
preemptive mutitasking, file system, task menagement. And I need real-time. 
And this Linux isn't provide very well. 

After about more the 4 months of desk research I found excelent kernel 
for my project. Its uC/OS-II Real Time kernel, project from Jean J. 
Labrosse http://www.micrium.com . This kernel is ROMable, and ported so 
far on about 85 or more processor. StronARM is too on the list :)
Sorry I don't have ARM CPU ( in poland is very hard to get some boards 
with this type CPU to play and learn ). I focused onto Intel CPU. Yes, 
thea are so bulky and power hogs, but I have many intel class CPU around me.

I think that this kernel is OK for use as IHU kernel. Small, fast, good 
design. Nice code and excelent documentation. I order book, and now only 
about 2 months from start I have good looking and promising kernel with 
about 8 running task. Now its meter of about 4 to 8 months to get entrie 
system working in beta stage. OBC ( like IHU ) need only very small list 
off tasks, it dosn't need to be very fast. Tasks are mostly linerar and 
atomic ( pin 1 up, then pin 3 down, read pin 2, read ADC, assembly frame, 
send to rs232, check time, pin 1 up, pin 3 down etc ). This is internal 
life off an IHU ( or any other On Board Controller ). But IHU must respond 
in finite time and predictable state to every event. No meter what is 
happening. Or it is lost. Art of writing software and commanding 
distant Meteo Station / Packer radio swith is easy game then commanding 
satelite.
But in both situations our babys is about 1500 km from us. More, I'm in 
better situation, I can buy at about 300 USD Air Ticket, and go with 
scredriver and hammer to fix it. But no one can fix satelite 
( at last no amateur sat ) while it is in the space.
BTW, uC/OS-II is free for education use. And not, I'm not employer of 
Micrium just in case if some one asked.

IMHO software for IHU must be small, tested, tested and more tested. 
And when it is ok, don't touch it, no meter what new feature You like to 
add.  Payload and other on board experiments need separate CPU for their 
own use. Main IHU software task list ( loop ) need to be max 1 sheet od 
paper long. No more, or there will be troubles. 

It's my 2 c to the discuscion.

I can help writing software and even creating some hardware for some 
new nice Satelite. It's very big fun when You play with pice of silicon 
and brass and when this little hardware is working ok. And doing useful 
things. And some times it very frustrating when it dosn't.

Sorry for my English, I still need more practise in writing.

Vy! 73 de Artur sp5qan

Admin of http://www.sp5zcc.waw.pl club SP5ZCC web server.

----------------------------------
Artur Dworak  AD1272-RIPE 
artur(at)supermedia.pl
Security Officer - Super Media ISP


----
Via the amsat-bb mailing list at AMSAT.ORG courtesy of AMSAT-NA.
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org



AMSAT Top AMSAT Home