Relay specifications

[crutschow]

The electrical life of a relay is usually given as the number of contact operations, not hours. The life of a contact is virtually infinite if the contact isn't switched.

The switching power is normally given as the apparent value of the load (VA). It has nothing to do with the contact resistance. Inductive or capacitive loads can, of course, increase contact erosion and thus reduce it's life below the rated value.

 

[Sceadwian]

crutschow, the datasheet clearly shows an electrical lifetime and a mechanical one, which one is which? Because they both can't be number of switching operations the mechanical lifetime is an order of magnitude lower.

 

[alphacat]

The switching power is normally given as the apparent value of the load (VA). It has nothing to do with the contact resistance. Inductive or capacitive loads can, of course, increase contact erosion and thus reduce it's life below the rated value.

The apprent power rating of the load is a constant number written on its enclosure, but the power that it actually consumes every moment changes.

So isnt the switching power depended on when you switch the load in relation to the mains voltage sinewave?

 

[Sceadwian]

Yes it is Alphacat, but it's a mechanical system and reacts very slowly compared to an electrical system, which makes it very very difficult to properly time contact make/break. If your timing is off by even a little bit you'll be at the wrong end of an AC swing and the contacts will weld or vaporize if you try switching under too much load.

 

[crutschow]

crutschow, the datasheet clearly shows an electrical lifetime and a mechanical one, which one is which? Because they both can't be number of switching operations the mechanical lifetime is an order of magnitude lower.

You have it backwards. The mechanical life is given as 1 x 10^6 (1,000,000) and the electrical life is 1 x 10^5 (100,000).

 

[crutschow]

The apprent power rating of the load is a constant number written on its enclosure, but the power that it actually consumes every moment changes.

So isnt the switching power depended on when you switch the load in relation to the mains voltage sinewave?

Yes. But the rating is based upon the apparent average (VA) power, not the instantaneous power. You are trying to complicate a simple rating.

 

[Sceadwian]

Sorry crut got it backwards. That means a million switching operations and 100 thousand hours. Or 10 years of continous operation, that sounds about right to me. Keep in mind the number of switchings it can survive are also related to how frequently it's switched.

 

[crutschow]

Sorry crut got it backwards. That means a million switching operations and 100 thousand hours. Or 10 years of continous operation, that sounds about right to me. Keep in mind the number of switchings it can survive are also related to how frequently it's switched.

No, it's 1 million mechanical switching operations with a small load, and 100,000 electrical operations at full load, which is the contact rating. It's not hours. The number of hours a relay operates has little effect on its lifetime if it's not switching.

I doubt that the frequency of switching much affects the number of switchings until failure.

 

[mneary]

So isnt the switching power depended on when you switch the load in relation to the mains voltage sinewave?

In theory, the exact time of switching could yield different switching power. With a real relay you cannot predict with certainty when the contacts will meet and/or separate.

A theoretical relay might close or open instantaneously. But all real relays take a finite time. Upon closing, the contacts usually bounce, sometimes up to 10 milliseconds of rapidly closing and opening. Upon opening, the air gap is increasing at a finite rate, and at the beginning of that time the air gap may be small enough to conduct (i.e. spark).

 

[Sceadwian]

Ahh, now I get it crut.

 

[alphacat]

In theory, the exact time of switching could yield different switching power. With a real relay you cannot predict with certainty when the contacts will meet and/or separate.

A theoretical relay might close or open instantaneously. But all real relays take a finite time. Upon closing, the contacts usually bounce, sometimes up to 10 milliseconds of rapidly closing and opening. Upon opening, the air gap is increasing at a finite rate, and at the beginning of that time the air gap may be small enough to conduct (i.e. spark).

Yeah i got you.
By software, I switch the relay 10ms before Zero-crossing.
Since the bouncing time of the relay is 10ms, the load starts operating at zero-cross of the voltage sine-wave (I checked it with the oscilloscope).

But i understand according to what you and Crutshow say that during that 10ms, the contacts also conduct and therefore also consume power, and for that reason i cant say that the switching power is when the bouncing has stopped?

 

[Sceadwian]

Alphacat you have no way of predicting that that contact bounce won't change as the relay ages or that it gets hung up one time and you end up switching during the peak of an AC cycle. While attempting to switch at zero crossing is a good idea to prolong the life of the relay you shouldn't expect that it will reliably allow you to overrate the switching power of the relay safely.

If the contacts have stopped bouncing the relay will pass it's fully rated power.