What is the real max frequency of a relay?

[Chriss]

Hi!
I thinking about to use a relay in a simple device which should be controlled with a uC.
I'm a bit confused with datasheets of relays.

What is a typical 12v relay max on/off switching frequency in real world?
I can see in the datasheet of JW RELAYS from Panasonic does they put two characteristic in:

Expected life: - Mechanical (at 180 times/min.)
                      - Electrical (at 6 times/min.)

What these two infos are mean?

Thanks...

[MK14]

It depends on the relay. But they usually take around 10ms (milliseconds) to switch on, or off.
So about 100 flips per second (50 Hz).

BUT such "fast" switching rates (for relays), is NOT a good idea.

Practical contact life is usually something like 10,000 cycles (under significant load, see applicable datasheet)
Also mechanical switching life, is usually around (1 to 100) million switch cycles (NOT electrically loaded).
But there are many different relays, some with specifications, different to what I said above.

Because of the relatively short switching life cycle and slow switch speeds, other methods, such as transistors or Fets (mosfets), make more sense, but are harder to design with.

[Kilrah]

I'd say "electrical" takes into account arcing and heating of the contacts at max rated load, while "mechanical" is with no load.

[MK14]

Hi!
I thinking about to use a relay in a simple device which should be controlled with a uC.
I'm a bit confused with datasheets of relays.

What is a typical 12v relay max on/off switching frequency in real world?
I can see in the datasheet of JW RELAYS from Panasonic does they put two characteristic in:

Expected life: - Mechanical (at 180 times/min.)
                      - Electrical (at 6 times/min.)

What these two infos are mean?

Thanks...

The figures you want, are in the right hand column of the datasheet. The figures you supplied, are just the test conditions, NOT the lifetimes.

Expected life Mechanical (at 180 times/min.) ... Min. 5×10^6
Electrical (at 6 times/min.) ... Min. 10^5 (at resistive load)

Which means:
5,000,000 mechanical (No Load), switches.
100,000 RESISTIVE load switches.

Capacitive/inductive load specs, are usually even lower, as the contacts can arc/burn/wear out etc. (As mentioned above).

Source:
http://www3.panasonic.biz/ac/e_download/control/relay/power/catalog/mech_eng_jw.pdf

Switching speed/times:

Operate time (at nominal voltage) (at 20°C 68°F) Max. 15 ms (excluding contact bounce time.)
Release time (at nominal voltage) (at 20°C 68°F) Max. 5 ms (excluding contact bounce time) (Without diode)

From "your" datasheet.

[German_EE]

The best way I have found to measure the switching time of a relay is to build a buzzer using the coil and a set of normally closed contacts. Then you just measure the frequency, compute 1/f and add a bit for the fiddle factor. I used this method to find the quickest relay I could for use in a radio transmit/receive circuit, most relays oscillated at about 500-700 Hz but the winner came out at just below 900 Hz and it's still in use a year later.

[SL4P]

Also keep in mind that every type of relay is different... mass, armature length etc.  All of which will affect switching frequency.
The 'on' time may well be different to the 'switch off' time.... I.e asymmetric.

[SeanB]

Release time also varies depending if you use a catch diode, which slows down the release time, or are prepared to clamp the drive transistor collector at a much higher voltage using either a zener diode or transient suppressor diode across it, and using a high voltage transistor so the spike can be very high, reducing the release time. As well if you use a current limiting circuit to reduce the coil current after pull in to a much lower value ( hold current) you can significantly reduce release time as well as you do not have to dissipate the stored energy in the coil you otherwise would have.

[Chriss]

Thanks guy's you give me a very clear and useful infos about my questions.

I wish to use a relay as a repellent, I mean, the ticking sound should be the repellent sound.
So, the relay would not have a load on it's contact's.

But, I also thinking about to use a buzzer from a pc mobo to generate the sound like the relay make.

I mean, if I will use a clicking relay driven in random time on and off and that click would be as a repellent noise than
I avoid to make a "sound generator" and an amp for the pc mobo buzzer.

But, on the other hand, with that "sound generator"  I could modify the frequency and that would be better for my project.

I would use an Attiny13 in my repellent project....

However, thank's for the useful infos about the relays.

[Zero999]

What's wrong with a speaker? A component specifically designed to make a sound and will be much more reliable and possibly cheaper too.

[Seekonk]

If you monitor coil current with a scope, you can see the slope change as the armature closes.  That will give you the real mechanical speed of the relay.

[Chriss]

What's wrong with a speaker?

Nothing is wrong with the speaker, I just think for the speaker I have to build a amplifier and write a more "complex" code to generate a frequency in some range.

Compared to the relay clicking code and parts it looks for me easier to build up but of course the speaker project would more fit my project because I can fine tune the sound.
Hmmm...

Is a buzzer from a PC mobo a piezzo buzzer or just a simple speaker?

[SeanB]

Most PC motherboards use a simple 30R speaker, with a frequency response of pretty much zero out of the 2-3kHz band, mostly now used to convey a power on beep and other basic sound from the south bridge.

[Zero999]

What makes you believe driving a speaker would be more complex than a relay? Just for simple tones, a speaker switched by an NPN transistor would do and should be loud enough when driven from 3V.

A motherboard typically has a moving iron speaker. More often referred to as an electromagnetic transducer by component distributors because the sound quality isn't high enough for most speaker applications.
https://en.wikipedia.org/wiki/Moving_iron_speaker

[sarepairman2]

look into vibrators. not the kind you might like, but the kind used in old radios.

[Chriss]

Thank you guy's I got so lot of infos from you that is amusing.

Thanks.

[PA0PBZ]

The best way I have found to measure the switching time of a relay is to build a buzzer using the coil and a set of normally closed contacts.

I'm not sure that is the way to do it, because the relay only has to travel so far that it opens the NC contacts, while real switching would involve travelling till the NO contacts close. Assuming of course that it is a normal break-before-make relay.

[Zero999]

The best way I have found to measure the switching time of a relay is to build a buzzer using the coil and a set of normally closed contacts.

I'm not sure that is the way to do it, because the relay only has to travel so far that it opens the NC contacts, while real switching would involve travelling till the NO contacts close. Assuming of course that it is a normal break-before-make relay.

I suppose you could drive the relay using a transistor and use the NO contacts to short the base to emitter, to turn it off.

[PA0PBZ]

I suppose you could drive the relay using a transistor and use the NO contacts to short the base to emitter, to turn it off.

Wouldn't that create the exact same problem but the other way around? It only has to open the NO contacts before getting energized again.
I think you need some kind of flippy floppy thing to do this right.

[Gyro]

Aren't relays designed to be as quiet as possible? You certainly wouldn't want them to make unnecessary noise - that implies excessive bounce, mechanical wear etc. not to mention nuisance value.

[retrolefty]

In troublesooting a thermocouple multiplex unit many years ago I had to measure actual contact bounce, which is the ultimate measure of the maximum speed limit for switching relay, you can't go faster when the contacts are bouncing.

 These were quality reed relays and I scoped contact bounce to less then a few microseconds, with no buzzing sounds at any speed using a function generator to drive the coils. These were industrial rated reed relays designed for the low voltage/low current generated by thermcouples. So probably not as good when higher switching higher voltages and current need as with slower relays. I tested using 12 dc contact voltage. Also these were all SPST NO which I know is faster then double throw contacts of rather mass contact assemblies.