Hi MIKE,

On 19-Mar-04, MIKE ROSS wrote to George White:

 JB>>> relays have inherent in themselves a measure of mechanical
 JB>>> positive feedback.

 MR>> Yes a "holding current" rating.

 GW>> No, it's actually a genuine, mechanical, posetive feedback. (Yes,
 GW>> I've spent time designing relays in the past, when there was much
 GW>> more electro-mechanical stuff about). When a relay is open there
 GW>> is a significant gap in the magnetic circuit. When it is
 GW>> energised the movement of the armature towards the pole reduces
 GW>> the air gap, and provides the positive feedback. In fact all
 GW>> relays have some form of air gap built in to stop them holding
 GW>> closed on the residual magnetism of the pole, a peice of non
 GW>> macnetic material of some kind somewhere in the magnetic circuit.
 GW>> On old relays it was often a copper rivet, on modern ones usually
 GW>> part of the plastic mouldings used in the manufacture.

 MR> This magnetic path break is sometimes hard to find but the
 MR> armature is often mounted using a brass spring hinge. One thing
 MR> for sure is it takes a whole lot less current to hold a relay so
 MR> that indicates some kind of reinforcement mechanism.

Not brass, it is too brittle, but phosphor bronze. That is a better
way of supplying the "air gap".

 MR> However, a positive feedback entails returning energy to a system
 MR> during part of an oscillation cycle. In the case of a relay, it
 MR> really exhibits an electromechanical hysterisis.  Many things
 MR> exhibit hysterisis without ever requiring a positive feedback. If
 MR> it was really positive feedback then wouldn't a relay oscillate?
 MR> (bounce or chatter is something else)

Hmm. OK, I was too simplistic in my description. What you say is much
more technically accurate. I could claim that time (and a minor
stroke) have clouded my memory, but it is more correct to say it was a
hurried and not properly thought through response.
Relay contact bounce is a very significant factor, and is why, where
possible, relay contacts are designed to slide accross each other
slightly. It reduces the problems of material transfer when they are
bouncing (and hence arcing).

 MR> In my own understanding one description is that it exhibits a
 MR> positive reluctance (i.e. increase in magnetic conductance) but
 MR> not a feedback. Say, if one were to add capacitance to this system
 MR> could it oscillate? I would appreciate if you would explain this
 MR> better to me?

That would be better as a technical explanation. In fact older vehicle
flasher units were capacitor/relay oscilators. Time has dimmed the
memory of how it was done but it used a electrolytic capacitor and
relied on a controlled difference between the pull in and drop out
voltages of the relay.

George

--- Terminate 5.00/Pro