General theory on EFI

[Dennis Grisez]

I have seen PC104 based equipment live for quite a while in 
service as in-vehicle test instrumentation.  A certain system I've
worked with runs two PC104's, one for video capture, one for
data, both write to removable hard drives on the fly. 
That having been said, I still avoid using any PC based equipment 
in durability tests where bump courses are involved.  There are better
solutions for that type of thing and PC's are far too fragile.  Not to say
they couldn't be made more rugged, of course.

Dennis Grisez
Project Engineer
Transportation Research Center
Ph (937)666-2011 Ext 367
Fax (937)666-5066
grisezd at trcpg.com


>>> Eric Bryant <BRYANTE at ghsp.com> 11/06/00 03:33PM >>>
> From: Jeff Webb [mailto:mustang at ufl.edu] 
> Subject: Re: General theory on EFI
> 
> Two things to consider:
> 
> 1) Most PC's are not suitable for automotive use, due to temperature,
> vibration, and reliability concerns.  I would imagine that laptops
> should be better than desktops, and PC/104 boards may be the 
> best bet. 
> If you're just playing around, then you can probably make 
> things work. 
> If you want a dependable driver, then you should research this
> carefully.  Just something to think about.
> 

This is the understatement of the year.  Most PC's aren't validated at all.
OK, I'm exaggerating a bit - they do fire up the prototypes and ensure that
they can run Windows and the game of the moment.  That's it.  No
environmental, EMC, or mechanical durability testing to speak of.  Laptops
aren't better than desktops unless they're one of the "hardened" industrial
ones, and most of us can't afford those.  

I'm not familar enough with most vendors of PC-104 hardware, but there's a
chance they'd be a bit better.  Since that market isn't moving along at
quite the rocket pace of the PC industry, there's a possibility that some
validation is being done.  Anything shy of aerospace-quality components are
going to fall short of automotive standards, in my opinion.  

The likelihood of running into environmental problems may not be all that
bad, but I'd be concerned about EMC problems.  It's pretty easy to figure
out if you're going to run into temperature problems, but EMC problems can
be significantly more difficult to figure out.  Maybe I'm being a bit too
cautious.

> 2) Linux is not a real-time operating system.  Look at www.rtlinux.org 
> if you want to use linux for controlling an EFI system.

Seems less costly than investing in a bunch of OSEK tools:)

Eric Bryant
mailto:bryante at ghsp.com 
http://www.novagate.com/~bryante 
----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org 


----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org