Minimising power loss in a relay

[HackedFridgeMagnet]

Hi
Just wondering if anyone has tried reducing the current in a relay after turn on. Obviously this is an effort to save a bit of power.
So turn on at full voltage and then throttle down the current to 35-75% afterwards. The spec sheet says it will hold at 10% of voltage.
I could do this by using two gpio pins both driving the same bjt through different resistors.


The relays in question are both DC 12v coil, switching 230v.
either Omron G8P or Panasonic ALZ.

http://www.components.omron.com/components/web/pdflib.nsf/0/1F359FF7E15DAA0E85257201007DD46E/$file/D20G8P0800.pdf
http://pewa.panasonic.com/components/relays/power-relays/non-polarized-power-relays/alz/

followup question, though I realise there is no truly right or wrong answer for the info I am giving you.

The Panasonic ALZ is specced to switch 16A AC, but I calculate that it is ok to switch 5.6amps inductive using the rule of thumb 0.35 x full current.

Would this relay be ok for 6 amps max on a small motor type load? I realise I am right near the edge.
Importantly the relay would switch only once per day and almost certainly wouldn't have 6 amps inductive every time it switched.
If I have to I could also calculate the currents zero crossing, but I would like not to be relying on this.


thoughts anyone.

[ve7xen]

If power consumption of the relay is important you may want to use latching relays, especially if you're switching infrequently as it sounds like you are.

[HackedFridgeMagnet]

Thanks that seems more what I want.

[DutchGert]

Why not go for a solid state solution?

[David_AVD]

What about a resistor in series with the relay coil for the reduced current.  A capacitor goes across the resistor to give it a full power kick at switch-on.

I have a feeling I've done this trick a long time ago and fairly sure it worked ok.

[DutchGert]

What about a resistor in series with the relay coil for the reduced current.  A capacitor goes across the resistor to give it a full power kick at switch-on.

I have a feeling I've done this trick a long time ago and fairly sure it worked ok.

But it is not very enery efficient. If u want it to be energy efficient and reliable go for a solid state solution or u could use a latching relay if u switch very few times a day

[David_AVD]

That's correct, but it did answer the OP's question on how to do the reduced current draw after switch-on.   

[HackedFridgeMagnet]

I have thought of solid state but can't seem to find the appropriate one. I want isolation too and inductive loads, under $10.

What about a resistor in series with the relay coil for the reduced current.  A capacitor goes across the resistor to give it a full power kick at switch-on.

Seems at least as good as my original solution.

Definitely switching very few times per day so latching relay seems good.

Not really sure about the single or double coil though. I will have to do more reading.
And with latching relays, if they use no power once latched can they stay switched indefinitely?

[G7PSK]

Used to be able to get relays that had 2 coils one drew a large current to pull on the other a much smaller current to hold the large coil was switched off automatically by the relay on closing. Not seen one around recently though.

[David_AVD]

Not really sure about the single or double coil though. I will have to do more reading.
And with latching relays, if they use no power once latched can they stay switched indefinitely?

Single coil require a polarity reversal, so more fiddly to drive.  Double coil is just like driving two separate relays.

Latching relays are bi-stable, so there's no time limit.

The only bad thing with latching relays is if you lose power with the circuit "on", it will stay on when you power up again (until you turn it off).  That could be an issue depending on your application.

[HackedFridgeMagnet]

The only bad thing with latching relays is if you lose power with the circuit "on", it will stay on when you power up again (until you turn it off).  That could be an issue depending on your application.

Yes I think that is the deal breaker.
Oh well either the cap or the twin gpios.

[David_AVD]

Using 2 IO pins means you have total control over the timing instead of fiddling with capacitor values, but ...

One thing that I'd be concerned about when using different base current drives, is that the gain of the transistor will play a more critical role in setting the lower voltage.

Using 2 transistors (each with a base resistor) with a dropper resistor (or zener so the coil R value has little effect) in the collector of one would be be more consistent.

The cap + resistor method should work well, especially if the coil is a relatively high resistance so the cap doesn't need to be too big.

[dannyf]

A typical approach here is to use a r/c network in serial with the relay. The resistor helps lower the current through the relay and the capacitor provides the initial kick to latch the relay but gets out of the way after that.

I do think, however, a latching relay is a far better solution in your case.

[HackedFridgeMagnet]

I think I will go with the series r/c network.
You are  right, I don't want to muck around relying on current gain. Basically with series the cap just needs to be big enough, so it is easier to size.
I will double check the pulse is long and large enough and add a good margin.
the latching relay would be better if i am sure it would always turn off, but it would mean I would have to put too much effort into working out, then explaining the consequences.

thanks david and dannyf, and the others.

[firehopper]

I think I will go with the series r/c network.
You are  right, I don't want to muck around relying on current gain. Basically with series the cap just needs to be big enough, so it is easier to size.
I will double check the pulse is long and large enough and add a good margin.
the latching relay would be better if i am sure it would always turn off, but it would mean I would have to put too much effort into working out, then explaining the consequences.

thanks david and dannyf, and the others.

if you sense the power failure. you can try and turn the relay off before power goes away, have a capacitor to store enough energy to turn the relay off, and have that as the last thing that happens.

[HackedFridgeMagnet]

Sorry I am not going to do the latching relay, not really for engineering reasons but just to make it easier to explain to the client. Probably the extra cost too.

I tried the cap idea in simulation only and it seems the cap needs to be pretty hefty and difficult to size.

So I am back to the twin gpio idea.

I did a sim with the relay at 155 ohms and 0.5 henrys. I don't really know the inductance of the relay yet and it will change when it is engaged.
So the following seems to work and if it doesn't I can just use the regular setup of one gpio anyway.

[dannyf]

The rc network concept also works in a driver setup: with one gpio pin, you can use it to produce stronger drive in transients and weaker in steady state.

[David_AVD]

I don't recall the relay coil resistance and capacitor values now as it was used some years ago.

Maybe I had a much higher coil resistance.  Have you measured (or checked the D/S for) the resistance of the intended relay coil?  That 155 Ohms sound low for a 12V medium size relay.

Keep in mind that many common relay types are available in a "high sensitivity" version.  This may help in with the capacitor size or just the overall power usage.

[NiHaoMike]

If the microcontroller has PWM, you can use that for driving the relay.

[HackedFridgeMagnet]

If the microcontroller has PWM, you can use that for driving the relay.

That's another way to do it. Obviously nice and efficient too. I will give it a try maybe at about 100khz.
I will make sure I am on a pwm timer pin. Have 2 standard duty cycles.

I don't think the extra noise will be an issue in this case.

[dannyf]

That's another way to do it. Obviously nice and efficient too.

Nope, nope and nope.

[HackedFridgeMagnet]

Lol, why not?

[vk6zgo]

Back when relays commonly had a lot of contacts,a normally closed contact would be used to short out a series resistor so the coil would get an initial higher current to operate it,then as the relay operated,the short was removed,so a lower "hold-in" current was provided.
 

[David_AVD]

That's a good idea to keep in mind.  Very elegant if you have that spare contact.   

[ve7xen]

Sorry I am not going to do the latching relay, not really for engineering reasons but just to make it easier to explain to the client. Probably the extra cost too.

O.o what kind of client are you working for that a simple mechanical device is harder to explain than an RC time constant reducing drive current by some mechanism?