Sender: <marv@lancaironline.net> (Marvin Kaye)
To: lml
Date: Sun, 12 May 2002 14:58:20 -0400
Message-ID: <redirect-1235652@logan.com>
From: "Gary Casey" <glcasey@adelphia.net>
Subject: diodes across relays
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
Importance: Normal

<<I also believe that a diode is needed on the key switch/starter/mags, but
again I do not know where or why.

I have three questions:

1.    What do the diodes do for me when installed?
2.    What will happen if I do not use diodes?
3.    How do I install them (which way do they face) and what size/spec
diode do I need?>>

1.  Jim Cameron gave a nice reply, but I'll give one in other words that
might help.  Just imagine that a coil of wire (inductor to be precise) has
inertia - it takes time for current to get flowing, but once the current has
built up it wants to keep going.  Suddenly opening the circuit with a switch
disrupts the current and something has to give.  What happens is that the
voltage rapidly builds up until the energy is absorbed someplace.  One place
is in a arc across the switch contacts and the other is by the internal
capacitance of the coil.  During this process the voltage could easily build
up to more than 100 volts.  Imagine that the switch is on the "high side",
or feeding the relay from the battery.  When the switch is opened the relay
coil tries to keep going, essentially "sucking" current from the downstream
side of the switch, pulling the voltage negative or below ground.  This
sounds like a terrible thing, but it happens every time an inductive load is
shut off, usually not bothering anything too badly.  In cars no diodes are
generally used and the system is just beefed up to tolerate the voltage
spikes.  In an aircraft, however, the switches are pretty delicate and there
are lots of radios that pick respond to transients, so diodes are often
placed across relays.  We used to call them "free-wheeling" diodes because
they allow current to flow in a circle from the negative (grounded) end of
the coil to the positive end.  When the switch is opened the current in the
coil continues to flow around this loop until the resistance of the coil
dissipates the energy. (a picture would be worth a thousand words)

2.  If you don't use diodes switch life might be reduced and radios might
pick up some noise - you could possibly hear a "pop" in the audio.  Other
than that nothing will happen.  The normal failure mode for a diode is to
short.  The failure is typically caused by over-voltage (called
back-biasing) or an over-current.  Over-voltage can happen in a "load dump"
condition when the voltage regulator is unable to shut down the alternator
fast enough when a load is disconnected.  These events do happen and that's
why it is important for the diode to be rated at 100 volts or more.  High
current conditions are not possible as the only current the diode conducts
is the relay coil current and that is less than an amp in most cases.
Personally, I would not use a diode in any relay that draws less than maybe
.2 amps as the voltage transient isn't enough to worry about.  I said the
switch life MIGHT be reduced - a small arc on opening tends to burn off dirt
and oxide that builds up on the contacts so putting a diode on the relay
could actually reduce switch life.  For typical switches not enough of a
consideration to worry about.

3.  The diode needs to be across the relay coil so that it doesn't conduct
with the normal battery polarity.  Imagine the painted bar on the diode to
be the "+" indication so that goes on the "high side" of the coil.  The
diode could be installed in any location - at the relay or at the switch,
but it is slightly more effective at the coil as the current loop is
shorter.  Note that these diodes aren't magic as there is a slight time
delay before they conduct and there will still be a very short voltage spike
every time you open the circuit.  Jim is right in that the most commonly
found diode - the 1N4002 is as good a choice as any as it is, IIRC, rated at
over 1 amp and 100 volts.  (will conduct 1 amp continuously without
overheating in the forward direction and not conduct with 100 volts applied
in the reverse direction)

Incidentally, diodes are NOT TO BE USED at the mag connections to the
ignition switch.  This very voltage transient that the diode eliminates is
the spark.  In other words a mag uses the voltage transient from opening the
circuit of an inductor (the coil) to create the spark.  Put a diode across
that and no spark.  Why doesn't all this create radio interference?  The
wires are all shielded.

That brings up a question, though - why use so many relays?  They add
weight, complexity, failure modes, cost and build time.  For resistive loads
like the pitot heater there is no voltage transient to damage the switch so
all you need is a switch rated at 15 amps or so.  Even for a flap servo
motor the current draw is not all that great.  The voltage transient from
opening the master relay doesn't effect much of anything since most
everything is shut off.  The highest current relay coil is probably the for
the starter contactor and you might want one there.  These devices don't
exist in cars because of reliability issues, cost and reverse polarity
protection.  Each of these diodes will turn into a short if you connect
power backwards with the switch turned on.  Not likely, but possible.

Sorry about the long post - Jim's was much shorter and to the point!

Gary Casey
ES project (with as few relays as I can get by with - maybe the only one in
the airplane will be the starter contactor)