[ARC5] Keying suggestions

[Mike Hanz]

That's an excellent point, Dennis - and as you say, it has been an 
industry standard for a loooong time.  I used to call it the 60% 
solution, because it gives you a "pretty darn good" [pdg] reduction in 
relay dropout time and provides a comfortable 'protection device' 
envelope cushion, albeit at a price of the zener.  Having employed both 
this and the series diode approach in military and commercial designs 
for a number of years when I first began in the engineering profession, 
it really boils down to how critical the timing is, because even the 
zener method demonstrably slows the dropout time...just not as much as a 
paralleled diode.

The amount of energy in the magnetic field of the relay at the moment of 
field loss, coupled with its inductive  (and related dynamic pole gap 
characteristics through the few tens of milliseconds of dropout), is 
what determines the peak of the inductive voltage spike, absent any 
other circuit effects.  It is interesting to see that the peak is 
remarkably stable for a single relay type, based on hundreds of 
thousands of tests we ran back in the 1960s on relays we were using in 
nuclear weapons and associated equipment.  As a result, it is relatively 
straightforward to select a diode whose PIV comfortably exceeds the 
potential spike, which is what Steve Williams did for his 274N set.  To 
bring it down to a more practical level, for these small relays, a 1kV 
PIV is usually more than adequate.

Since I have spent more than 40 years in systems engineering, I must 
also bring us back to earth with the classic question that should be the 
first to be asked of any proposed approach, i.e., "What happens if the 
series diode fails?"  Guess what...in this particular case, not much.   
:-)    If it fails open, then your key stops working.  If it fails 
shorted, then the maximum keying rate may be slowed down, depending on 
how the relay is fed from the power source.  You may not even know of a 
failure...  That's what most folks usually call a win-win situation.

73,
Mike

Dennis Monticelli wrote:

>There is an industry standard way of clamping that inductive spike and
>still allowing for rapid release of coil energy.  You simply put a
>zener in series with the clamp diode.  When the diode engages, the
>total clamp voltage is 0.7V plus Vzener.  The value of the zener is
>not critical.  I would suggest something in the 10 to 20V range with a
>current rating equal to or greater than the coil current.  A zener in
>that voltage range would dissipate the relay coil engery approx 15
>times faster than just using a diode while clamping the max voltage at
>a predefined safe level.  To get the same benefit from an RC snubber
>would require some experimentation of values.
>