[Diy_efi] RE: Military/ Automotive design

James Holland j_holland
Mon Feb 26 19:16:12 UTC 2007


> James,
> Can you, if this is possible, give us a few tips re: the basics of
> designing
> circuit boards for the high-temp. environment such as military and/or
> automotive?
> Scotty

Its not really that much different to designing for any other environment.
You have to calculate your tolerances over the entire temperature range.
Calculate your junction temperatures or follow the manufacturers guidance on
derating for whatever your max. temperature is.
The requirements for military stuff vary widely. Naval stuff that is used
inside the ship is generally in an air-conditioned environment and the
temperature range is pretty limited. You also usually have a nice big AC/DC
converter between your equipment and the ships generators so the supply is
pretty stable. Stuff used outside has to meet the temperature requirements
for artic use and equatorial use (ambient + solar heating).
The main difference between military and commercial automotive is that a car
manufacturer gives the customer a handbook stating 'Do not jump Start' Do
not run with the battery disconnected' 'Do not connect the battery the wrong
way round'. The military give the designer a spec that says. 'It must
survive....'
The challenging aspects are the wide supply range and lack of regulation,
the temperature range and the horrendous amount of noise on both power and
signals. Of course your equipment musn't generate any significant conducted
or radiated emissions!
If you consider a cars supply. You have to design your equipment to operate
down to 7V and up to something like 16.5V. The power dissipation increases
5.5 times. You can reduce power loses by using switched mode regulators and
PWMing solenoids and motors but then you start struggling with noise.
Low temps can sometimes be more difficult than high temps. A colleague had
board with a microcontroller on it that was designed by a subcontractor. It
wouldn't start at -40C. The problem was eventually traced to the decoupling
caps which were Y5V or similar. The capacitance of a Y5V drops dramatically
with temperature, at -40 they weren't doing anything at all. The processor
was coming out of reset and switching on. The current spike was causing the
voltage to drop and the Brown Out Detect was kicking in. The processor went
into reset. When it came out of reset it all started again! A change to X7Rs
cured the problem.

Cheers
James





More information about the Diy_efi mailing list