My buddy Dave has been pestering me to SUBSCRIBE to this newsletter  for
over a year now.

My name is Steve Wilke, I've been  involved on and off in aviation since I
was a kid.

I  spent over 20 years in Aerospace & Defense, both active duty in  the
Marines and as a civilian working on the radar and missile  systems of such
aircraft as the F-4J/S Phantom II, the F14A/F14D  Tomcat, the F-15 Eagle,
F/A-18
Hornet, the YF-22  prototype, the AMRAAM Missile, and many other very fun
toys.

I decided to subscribe because I am concerned about people  using
commercial-grade parts in an aviation environment.  At  the very least, they
should
be using automotive-spec  parts, but most preferably mil-spec parts.

I have a feeling that  many of your newsletter readers/posters are not aware
of what  happens during "load dumps", such as turning off your landing
lights.  You can get pretty high voltage "spike" transients on  the
electrical
system (60v) when you suddenly turn  off a heavy load.   This will wipe out
circuits that don't  have protection against such transients,  particularly
those  like the 4000 series CMOS.

I really don't care so much for the  "newsletter" type of things; it would
be preferable to have  everything on a forum - it is so much easier to
maintain   continuity that way.

I have been a member of  "AllAboutCircuits.com" as "SgtWookie" for several
years - and I have  managed to accumulate more posts than any other member,
by a   long shot.

Anyway, rather long - I guess "SUBSCRIBE" would've  been adequate.

Regards,
Steve in  Orlando.



It sounds like Steve has some useful and practical  experience on real
hardware..

None the less, at the risk of sounding  argumentative (though I'm not
trying to be), I have a little different  perspective.  I work as an
engineer at an IC manufacturer.  From  my perspective, I know that
_many_ commercial, automotive, and military  grade IC's often don't
have any internal differences, and in fact are built  on the same
production line and taken from the same lots.

Sometimes  the different product markings imply a slightly different
test  environment/spec.  For instance, automotive grade IC's are
spec'ed to  operate over a wider temperature range than commercial
parts.  Early  in a product life product destined for automotive
marking will be tested  over the wider range to verify performance.
However, once defects are  analyzed, in order to lower manufacturing
costs, all of the devices are  tested in the same environmental
conditions (usually the more relaxed,  commercial conditions).  Often
mil-spec and automotive spec parts go  through the production line
along side the commercial parts.

I agree  with Steve's assessment that many people don't understand load
dumps.   I question his use of landing lights as an example..  Load
dump is  caused by the storage of electrical energy due the inductive
reactance of a  circuit, and how much current is going through the
circuit.  In the  case of an incandescent light bulb, the inductance is
very low, so the  stored energy is very low, even though the operating
current may be  somewhat high.

The only thing that's sort of strange about incandescent  lights is
that they have relatively high inrush current when starting from  a
cold filament.  The rated current through the filament is limited  by
its resistance while at it's normal operating temperature.  When  the
filament is cold, the resistance is _much_ lower, so when you  first
flip on the switch, the current will be higher for a short time  (in
the milliseconds if I recall).  Much before the you get  useful
brightness from the lamp, the inrush current has dropped  off.

This inrush current doesn't have any bearing on turn-off  transients.
The filament takes a long time to cool (many seconds) to a  temperature
where its resistance is again low.  The load dump from a  simple
landing light will be exceedingly low.

Again, load dump  happen with highly inductive sources and inductive  loads.


Alternators can sometimes put their regulators at risk if  they are
running at full output and they are then disconnected from the  system
load (disconnect relay).  I agree that 60V is quite  possible
(probable).  Normal (factory) regulators are usually  reasonably well
protected from this.  The el-cheapo internal regulator  from
Checker-Schucks-Kragen might not be (but probably is).

In  actuality, an alternator load dump isn't of the purely  inductive
kind.  Wound-field alternators (the kind in your car) are  regulated by
adjusting the current going through the field windings.   Regulation of
the field current has damping which delays response -  improving
stability.  The delay causes the output voltage to lag  changes in
load.  When the alternator is disconnected from the load,  not only do
you get an inductive kick, it last a bit longer because the  regulator
needs time to respond to the changed load and reduce the field  current
appropriately.



A landing gear motor may cause a  purely inductive load dump.
Switching off a large motor with a small switch  can damage the switch
contacts.  That's why such loads are often  controlled by relays.



Finally, a very small scale load dump  always occurs when opening the
switch which controls a relay's coil.   The current through the coil
includes some stored magnetic energy.   Opening the switch attempts to
halt the current through the coil, and the  stored energy causes the
voltage across the coil to rise for a short time -  until the energy is
depleted.

A diode can be connected across the  coil.  With the diode, the voltage
is limited to no higher than the  forward bias voltage above the bus
voltage - eliminating the  problem.

The scale of the load dump on the coil is always determined by  the
current through the coil, and is completely unrelated to how  much
current the relay is controlling (through the contacts).

The  voltage caused by the dump is caused by the very small stored
energy in the  coil.  It lasts a very short amount of time  (decaying
pulse).

Normal system wiring doesn't conduct spikes  (decaying pulses) very
efficiently.  The spike will be flattened by  bus capacitance.  This is
why usually only the controlling switch is  put at risk.


I hope some of this helps.  If someone disagrees,  I'd be happy to
discuss, and even be shown wrong..   ;)


Regards,

Matt-



Hello Matt,

If your parts that are rated as being mil-spec are not being tested in
compliance with MIL-STD-883, I would be VERY interested to hear about it.

It's fine to test everything to a commercial threshold, if you're marking
everything commercial.  However, if your company is attempting to
rubber-stamp everything while not conducting the tests required for the higher
grades (especially mil-spec), then I predict some very unpleasant events to
occur.

I worked on a military communication satellite project years back; my
direct involvement regarded PROMS.  We started off with mil-spec proms that  had
to be scrounged from everywhere (dealing with qualified parts lists can be
a  nightmare), then programmed the PROMS, and then they were subjected to 9
months  of thermal cycling to simulate conditions in space.  Only 15% of the
254  PROMS made it through the test, but the failures were planned for.  We
even  had a couple of spares.

As far as load dumps, the inductive "kick" from the alternator when a heavy
load turns off is what I was referring to.  Not all aircraft have
retractable gear, but they all have landing lights.  Flap motors or fuel  transfer
pumps would have been a better example to have used.

A note on relay coil reverse-EMF supression diodes aka "flywheel" diodes -
it's typical for people to use a single standard diode to prevent the
spike.  However,  you'll get faster suppression if you use either a  TVS diode
or a 12v Zener back-to-back with a standard diode.

System wiring also has inductance.  A piece of straight wire 10mm long
(0.394") has an inductance of about 15nH at 10MHz.  This adds up rather
quickly.  A wire that's 10 feet long will therefore have an inductance of  about
4.6uH.  That can cause problems if you're not aware of it.

Steve