Path: utzoo!attcan!uunet!cs.utexas.edu!rutgers!sunybcs!kitty!larry
From: larry@kitty.UUCP (Larry Lippman)
Newsgroups: sci.electronics
Subject: Re: Latching Relays
Summary: The consideration of relay contact _timing_
Message-ID: <3406@kitty.UUCP>
Date: 24 Sep 89 04:54:15 GMT
References: <2030.AA2030@americ>
Distribution: na
Organization: Recognition Research Corp., Clarence, NY
Lines: 69

In article <2030.AA2030@americ>, erk@americ.UUCP (Erick Parsons) writes:
> A very common latching circuit (for motor control) is to use one
> relay with two sets of contacts. It is a DPDT relay.         ^^^
> It requires three wires in the control conduit, two switches: SPST
> momentary contact, one normally open and one normally closed.
>                          _________                               Hot
>                         |         |                        ___    |
>           ------o-----------|/|-----------o-------------- .   . --o
>           |             |         |                               |
>           o-----o-----------| |-----------o-------------- .---. --o
>           |             |_________|
>           |
>           |                  C
>           |-----o------------O------------o-----------------------o
>                              I                                    |
>                              L                                   Gnd
>                     2nd set of contacts
>                          for load

	The above scheme is inadvisable, and simply will not work with
many relays.  The reason is that most relays intended for power control
purposes have a SPDT contact configuration known as "break-make" (BM).
Operating the NO "on" switch will energize the coil, but as the armature
begins to pull in, the break contact will break, interrupting the current
flow to the coil BEFORE the make contact will make to latch the relay on.
The net result is that the relay will chatter, but never pull in to a
fully operated and latched state.

	While a contact configuration of "make-break" (MB) will permit the
above circuit to be functional, such a MB contact arrangement is relatively
uncommon in a power control relay.  In reality, running the "on" switch
with a break relay contact serves no useful purpose.  The above circuit may
be used if the "on" switch is connected directly to the relay coil, with
the break contact being unused.  A motor contactor, for example, uses a
single make contact for the latch circuit.

	While the heyday of relay logic control circuits has now passed
due to the advent of solid-state logic, microprocessors, and industrial
Programmable Logic Controllers (PLC's), relay contact timing can still be
a problem if not considered in the design of some control circuits.

	The concept of relay and relay contact timing was honed to a fine
art in the days when telephone switching circuits were solely designed
using electromechanical technology.  As an example, a common style of
of wire-spring telephone relay having TWELVE contact sets could be
configured with any combination of make (M), break (B), make-break (MB),
break-make (BM), early make-break (EMB), early break-make (EBM), early
make (EM) and early breek (EB) contacts!  In addition, not only could
such a relay be designed with up to three separate windings, but it
could be designed using copper slugs as an additional one-turn winding
to effect either a "slow-release" or "slow-operate" relay with effective
"slow" timing as great as 250 milliseconds.

	I'd like to make one other point as long as I am here on my
soapbox. :-)  I do not fault the above poster for his circuit suggestion,
even though it is generally unworkable.  At first glance, the circuit
seems reasonable and it seems that it should work.  Many people seeing
such a circuit would scoff at the suggestion of actually building and
testing it since it seems "so simple".  The point is, the REASON why
this circuit won't work is NOT something that is taught today in most
BSEE curricula.  It is something that an engineer learns only through
first-person hands-on experience.  And such experience is something
which takes time, involves making mistakes, and is not learned in any
classroom.

<> Larry Lippman @ Recognition Research Corp. - Uniquex Corp. - Viatran Corp.
<> UUCP  {allegra|boulder|decvax|rutgers|watmath}!sunybcs!kitty!larry
<> TEL 716/688-1231 | 716/773-1700  {hplabs|utzoo|uunet}!/      \uniquex!larry
<> FAX 716/741-9635 | 716/773-2488      "Have you hugged your cat today?"