question on electromechanical relay specifications

[Gus]

I'm looking at some miniature electromechanical relays that use a DC voltage
to energize the coil. The DC ratings for the contacts are given for
resistive, inductive, lamp, and low level loads and are wildly different.
For example with a resistive load its 1A, but only 100mA with a lamp load.
Why the difference ?

I've tried to ask a vendor's application engineer but they haven't answered
yet, and I'd appreciate any insight from this group.

 

[Tim Wescott]

I'm looking at some miniature electromechanical relays that use a DC voltage
to energize the coil. The DC ratings for the contacts are given for
resistive, inductive, lamp, and low level loads and are wildly different.
For example with a resistive load its 1A, but only 100mA with a lamp load.
Why the difference ?

I've tried to ask a vendor's application engineer but they haven't answered
yet, and I'd appreciate any insight from this group.

* A resistive load will start flowing current according to I = E/R, and
will do so indefinitely.

* An inductive load will want to keep flowing current when you break
it's circuit because V = L dI/dT. When the relay opens and I wants go
to zero instantly (dI/dT = -infinity) the inductor will generate enough
voltage to keep the current going -- even if this means sparking across
the relay contacts (snubbers reduce this, sometimes dramatically).

* A lamp is rated for its current once it is hot. When it is cold its
resistance is significantly less than when it is hot; the relay needs to
supply that inrush current without welding the contacts together.

* I dunno about the low-level stuff.

 

[Tim Shoppa]

For example with a resistive load its 1A, but only

100mA with a lamp load. Why the difference ?

Most lamps are rated for their steady-state current. When the filament
is "cold" they can draw many more times more current than their
steady-state current. For the relay contacts it's this switched current
that matters more than steady-state.

A factor of 10 is about right for many small vacuum-filled (how do you
fill a container with nothing?) lamps.

There are tricks to reduce the switch-on surge including "prewarming"
the filament through a resistor. Prewarming also greatly increases
total number of lamp on/off cycles before the filament burns out.

Tim.

 

[John Larkin]

* A resistive load will start flowing current according to I = E/R, and
will do so indefinitely.

* An inductive load will want to keep flowing current when you break
it's circuit because V = L dI/dT. When the relay opens and I wants go
to zero instantly (dI/dT = -infinity) the inductor will generate enough
voltage to keep the current going -- even if this means sparking across
the relay contacts (snubbers reduce this, sometimes dramatically).

* A lamp is rated for its current once it is hot. When it is cold its
resistance is significantly less than when it is hot; the relay needs to
supply that inrush current without welding the contacts together.

* I dunno about the low-level stuff.

In theory, the nice gold or silver plating on the contacts is good for
low-level switching, but once you arc the contacts with bigger loads,
it's trashed, so may not be suitable for very low-level use after
that.

That's what I've heard, anyhow.

We're using tiny sealed relays for microvolt-level switching, latching
relays to keep the thermal EMFs down.

John

 

[Adrian Tuddenham]

Most lamps are rated for their steady-state current. When the filament
is "cold" they can draw many more times more current than their
steady-state current. For the relay contacts it's this switched current
that matters more than steady-state.

It is flowing during the 'contact-bounce' time, before the mechanical
bits stop moving. This can cause welding of the contacts or a
micro-welding and tearing apart sequence, which rapidly erodes the
contact surfaces.

 

[John Larkin]

Most lamps are rated for their steady-state current. When the filament
is "cold" they can draw many more times more current than their
steady-state current. For the relay contacts it's this switched current
that matters more than steady-state.

A factor of 10 is about right for many small vacuum-filled (how do you
fill a container with nothing?) lamps.

But most of them are filled with nitrogen.

There are tricks to reduce the switch-on surge including "prewarming"
the filament through a resistor. Prewarming also greatly increases
total number of lamp on/off cycles before the filament burns out.

I've read that cycling doesn't actually reduce filament life; life is
dominated by evaporation.

John

 

[w_tom]

"Prewarming" a filament has no effective on power cycling. But
again, even the industry's own specifications do not define power
cycling as destructive. Light bulb life expectancy is a function of
voltage (filament temperature) and hours of operation. This even
provided by industry standard formulas.

When a light bulb is cold, it will demand a very large current. If
the bulb is kept slightly warm, then the intial power on current is
lower. This means smaller value transistors and less transients on the
ground plane that would otherwise caught state sensitive devices to
change logic states.

It is mythical that power cycling causes incandescant bulb failure.
But that initial power on current when the bulb is cold can be
problematic to the control electronics.