Why Relays waste power and how to overcome that?

Question

Let's say that I am going to:

1. Use a relay to switch between 2 DC +12V power sources (one may interrupt so the other source is a backup).

2. I will use a capacitor on the load with minimum leak possible I can find to supply power during the relay switch operation.

3. The lines thikness is around 8 mm.

4. The current may go as high as 23A and as low as some mA.

I have been told that relays waste power, but it is not clear to me by how much as I don't know the best kind of relays to use for my application. I also read somewhere that an IC can be used to control relays so they waste less power, but that IC may also use power too which may result in more power loss probably.

Anyway, I want the most efficient solution for my specific applications.

Answer 1

Coil-held relays do indeed "waste" power as you say. However, you're paying for some value. The value is:

• It can make a connection even when fully de-energized. In fact, at total loss of power, it makes a connection. These are the "NC" contacts. This makes it a good choice for backup power; you connect normal power to both the coil and the NO contacts. You connect emergency power to the NC contacts.

• Positive interlocking. There is no case where NC and NO contacts are ever connected to each other. That is absolutely mandatory in certain power transfer applications due to the need to prohibit backfeeding.

• Positive multi-pole throw. There is no case where pole 1's NC contacts can be connected whilst pole 2's NO contacts are connected. That serves a whole bunch of applications where this must be positively so: to name a few,

  • Power transfer as discussed
  • reversing
  • Series-parallel switching

Remember, the relay does all that, and, when coil power is entirely lost, the relay actively throws with no power or assistance whatsoever.

Now, if you are willing to sacrifice the auto-throw-on-power-loss functionality, you can get all of the above with a latching relay where you throw with a momentary impulse, and then the relay throws. However, the relay will remain in this position indefinitely; it takes active action to throw it back.

Further, the 'active action' is a bit hard to manage. You need an impulse of energy for some interval of time. So you need to have electronics that sends the pulse for the required time interval to throw the relay over fully. Worse, you can't use extra contacts for that, because most such relays do not have the special contact needed: which would need to be closed until the relay is almost at end of throw.

Answer 2

Look for a relay (Farnell, Digikey, Mouser, RS etc.) that can handle a DC current of circa 25 to 30 amps through the contacts. It's important you choose the "DC rating". Then reduce your candidate list to those requiring a coil voltage of 12 volts.

Then reduce your list further to those that require the least amount of coil current to meet your efficiency needs - the coils waste power when activated.

Then decide what your switchover circuit will consist of i.e. op-amps, comparators etc. and decide how much power this circuit might waste (it should be a fraction of what the relay coil wastes by the way).

If your efficiency still looks poor then maybe you go for MOSFET switching instead of relay switching but you might only get something like up to a 10 times improvement in power losses on light to moderate loads and maybe only a 2:1 improvement on your heaviest load for this extra complication.

Answer 3

1. Relays need lower current to hold the state once turned on.
2. There are latching relays which only need a pulse input to switch the state

Answer 4

As mentioned already, there is a big danger of sticking/welded contacts if an AC rated relay is used for DC. Above 15A, a contactor is normally better suited. So a solution would be to search for a latching DC contactor rated for 25A, but those might be hard to find. Another solution would be a circuit with 2 or more Power MOSFETs, which are controlled by voltage and not current - the tiny control current is neglegible (nano or micro Ampere). But there is some more lost since the on-resistance is higher compared to a contactor. And special emphasis should be made to the order of switching, i.e. can both sources be parallel during switch over or is a "break before make" needed in the MOSFET control circuit.