Topband: Vacuum relay powering #2

[Tom Rauch]

> I work for Agilent (used to be HP) and they used a LOT of 
> reed
> relays.  They come in form A (SPST, NO), form B (SPST, NC) 
> and
> form C (SPDT).  The form C relays need a permanent magnet 
> to hold
> the NC contact when not energized.  IIRC, the energizing 
> current
> is the same between forms A, B, and C, (being limited by 
> heat dissipation)
> and the result is that the contact pressure is half as 
> much for form C.
> Back when we had component engineers and a reliability lab 
> (before
> they were all fired to save money) the expects told us 
> that form
> C was inherently less reliable than A and B, presumably 
> due to the
> reduced pressure.  We were instructed not to use form C 
> unless
> absolutely necessary.

After some thought and a look at this last night, I'm going 
to strongly disagree with the idea that reduced holding 
voltage reduces current rating in most, if not all, relays. 
My reason does not conflict with Rick's data at all.

While I haven't seen every relay in the world, I've looked 
at a lot of relays. I have open samples of dozens of 
different relays collected over the years because, when 
picking a relay, I always look at internal construction. I 
also have glass and ceramic vacuum relays that are open, as 
well as open frame relays.

In every single relay I've looked at there is some spring or 
spring action that controls energized relay contact 
pressure. I can't find a single relay in all my relays that 
has a floating armature connected to a rigid contact. Even 
in very high current form X relays or double make double 
break relays, the contacts are buffered by something "spring 
like" or an actual spring that flexes.

Now think about how the electromagnet works. When resting, 
the armature is either out of the coil or up away from the 
pole piece some distance. The flux path has an air gap. As 
coil current increases the armature moves towards the other 
pole or into the coil, and flux density through that 
magnetic circuit increases rapidly just from the distance 
reduction alone. If you have enough coil current to overcome 
the return spring that holds the relay in the normally open 
position and open the normally closed contacts, you can be 
assured the relay armature will fully move until the plunger 
or armature bottoms out. I wasn't able to find a single 
relay that didn't behave this way.

So now we have exactly the same mechanical travel regardless 
of applied holding current as long as the armature is 
seated. ALL of the contact pressure is determined only by 
the spring tension against the contacts when actuated, just 
as it is determined by return spring tension while resting.

In small open frame 10-20 amp contact relays like used in 
antenna switches, the spring is actually the relay contact 
bars. Those bars, beryllium copper in some open frame low 
and medium current relays, are the springs that limit 
contact pressure when the relay is energized. If you have a 
relay like that push the relay closed and watch how that 
spring bends just before the relay armature bottoms on the 
magnet pole. No matter how hard you push on the armature the 
armature won't travel further, so the contact pressure never 
changes once the relay armature has bottomed.

You'll see the same thing is true with very high current 
open frame relays, like power contactors. The contacts are 
on rigid carriers, but the rigid carrier is buffered by a 
spring on each contact bar. Again this lets the contact seat 
just before the pole piece and armature bottom out. Once 
again, no matter how much coil current, the contact pressure 
is the same.

All of the vacuum relays I have use a similar method to one 
of the two above. They either use a separate spring or a 
"spring material" that flexes  in the contact support bar.

There are several very good reasons relays do this. First, 
the contacts or armature might be physically damaged or 
deformed  if pressure was simply limited by coil current and 
not a controlled source like a spring. Second, the 
manufacturer would have an impossible job adjusting the 
relay for correct holding pressure at a given voltage if he 
relied on flux density to control contact pressure. He could 
never get it right! The reason is temperature would greatly 
affect holding pressure. Not only does the coil change 
resistance a great deal with temperature, the materials in 
the relay will change size. Even at a constant temperature 
the contact pressure would vary as the coil heated. I 
measured 30-50 degrees F temperature rise over ambient in a 
small 1.2 watt relay coil in 15-20 minutes of operating in a 
0 degree F test chamber. It is impossible to have anything 
close to steady pressure without a buffering spring.

This perfectly fits with my observations of how relays work. 
A few weeks ago I spent a week or two  trying to find a way 
to force a better connection in a transfer relay. I had a 12 
volt relay with higher than desirable contact resistance and 
I tried to activate and hold the relay with 35-30 volts to 
see if reliability on closure would improve. It did not. The 
only thing higher voltage did was reduce transfer time.

Amplifier manufacturers like Heath and the rest commonly 
used open frame relays for antenna switching. The big 
problem with those relays was receiving through the relay 
when they were out in service for a while. This was caused 
by a lack of current through the NC contacts allowing a film 
to build up, and the fact two contacts were effectively in 
series in the receive path. I initiated a change in relay 
design that removed the unused center NC contact from the 
relay. This allowed more bending moment on the contact 
carrier bars (which are springs)  from the return spring 
pressure, and this increased wiping action. If you look at 
some Heath, Ameritron, and other manufacturers you will see 
the unused NC contact bent up out of the way. This was a 
first step until custom relays could be produced.

This actually all fits with what Rick said. The form X, or 
double make double break relay is less reliable for contact 
resistance when compared to normal single contact point 
relays. That arrangement has twice the number of contacts in 
series with the signal path. This means it is much more 
likely to have resistance problems at low contact (NOT coil) 
currents. You have to be sure two contacts stay clean, not 
just one. The second issue is if you use a SPST relay 
instead of a form X, all of the spring pressure goes to ONE 
contact. This doubles contact return spring pressure 
available to bend the flexible contact carrier and increases 
wiping action. This is for a constant coil size.

If I would have read this thread four weeks ago... the two 
weeks I recently spent looking at relays would have been 
more productive. Now, with a few hours work last night, I 
answered my own question why increasing coil voltage did not 
increase pull in reliability the slightest amount!! I hope 
this saves someone else work and worry.

73 Tom