Protecting a power supply from backfeeding ?

Hi there!

I was asked to repair a camping cooler (a box where you put water bottles and food so they stay cool). It can be powered with the 12VDC from the car's cigarette lighter plug, or from external power supply 230VAC to 12VDC 5A.

It was an easy repair because there was a cracked solder joint in the external power supply, I just resoldered it and it works.

But I want to improve it! Currently, at the end of the external power brick there is a female cigarette ligther plug, where you connect the male plug from the cooler when you want to run it from 230VAC.

There is enough free space inside the cooler's cover to put the electronic board of this external power supply, so I would like to integrate it, remove the female cigarette lighter plug and connect the 12V output directly to the "main board" of the cooler.

Of course I still want to be able to plug it into the car's cigarette lighter plug, so what do I need to add in the circuit so it doesn't backfeed the (now internal) 230V to 12V power supply ?

Is it OK to use diodes on the + of both power sources, like this:

And what kind of diodes would I need in my case ?

Or is it safe to just backfeed this power supply?


Or is it safe to just backfeed this power supply?

That's one of those questions where the answer is 'it depends'. Many you will get away with, some you might not. Play safe and use a Schottky diode with an If greater than the expected current. Sorry, I am on holiday,if I were at home I'd look what I have and suggest a specific diode.

Could you not use a SPDT power source selection switch?

PerryBebbington:
That's one of those questions where the answer is 'it depends'. Many you will get away with, some you might not. Play safe and use a Schottky diode with an If greater than the expected current. Sorry, I am on holiday,if I were at home I'd look what I have and suggest a specific diode.

I have one FCH10A20 (10A schottky barrier), and a few SR560 and SB5100 (5A schottky diodes), would that work considering the sticker on the cooler shows 12V 48W max (so 4A) ?

aarg:
Could you not use a SPDT power source selection switch?

Yes that's an option I had considered but I wanted to avoid another switch :slight_smile:

guix:
I have one FCH10A20 (10A schottky barrier), and a few SR560 and SB5100 (5A schottky diodes), would that work considering the sticker on the cooler shows 12V 48W max (so 4A)

5A diodes should be ok but check the current with each power source independently to be sure. I don't entirely trust current and power ratings on labels.

Of course, the simple answer is to use a 12 V relay connected to the AC power supply, to switch over the connection. This avoids diode drops.

Since the AC supply contains rectifiers, it is most unlikely that it would be harmed in any way by “backfeeding” so a diode should be unnecessary. If you have a variable supply, you could connect an ammeter (multimeter set to 200 mA) in series to the un-powered AC supply and verify that any current drawn (should not exceed 20 mA) does not increase significantly when you ramp slowly up to 16 V or so.

Paul__B:
Of course, the simple answer is to use a 12 V relay connected to the AC power supply, to switch over the connection. This avoids diode drops.

Since the AC supply contains rectifiers, it is most unlikely that it would be harmed in any way by "backfeeding" so a diode should be unnecessary. If you have a variable supply, you could connect an ammeter (multimeter set to 200 mA) in series to the un-powered AC supply and verify that any current drawn (should not exceed 20 mA) does not increase significantly when you ramp slowly up to 16 V or so.

Thank you, I didn't think about using a SPDT relay, I like the idea :slight_smile:

Even if backfeeding the PSU was not an issue, I wanted to use diodes anyway to prevent the 12V to be on the contacts of the male cigarette lighter plug when the PSU is active (obviously dangerous). But using a relay instead will solve all problems, theoretically!

Thanks all :slight_smile:

I'm not sure putting the psu inside the case is a good idea as it generates heat. As you can see form the heat sinks!

johnerrington:
I'm not sure putting the psu inside the case is a good idea as it generates heat. As you can see form the heat sinks!

And there is a fan in there. May just need to modify the airflow a little. :grinning:

Yes, in fact I changed my mind and will put the PSU here:

The piece of cardboard is temporary, I will shape a piece of plexiglass to direct the airflow. Then there is enough room on the right so I can make a compartment to tidy the cords in :slight_smile:

I'm so bored...

I think I’ve found a suitable relay (datasheet here)

Can you tell me if my schematic is correct or is there any modification I must do ?

What are the resistor and diode for?

Isn't the diode supposed to protect against reverse voltage spike when the relay coil is de-energized ?

I'm not sure if I need that resistor. It was originally on the module board where I took this relay, but after looking this board again, I think it was part of a "RC snubber". It's a weird board where there is two connectors, one with 3 pins labelled VCC ECO GND and the other with 2 pins, labelled AC, with a bridge rectifier and few other components. I have no idea how this board was supposed to work, applying 12V to VCC and GND didn't work so I just took the relay, the diode and the resistor. It's a board that I found in a Philips MCD908 "DVD micro theatre". I think "VCC" was a supposed to be a negative voltage or something...

Clearly I know nothing about electronics, so please enlighten me :wink:

When you interrupt the current to an inductor suddenly (the relay coil is an inductor) the effect of the collapsing magnetic field is to try to maintain the original current as it was just before the current was interrupted. If the inductor is open circuit then in order to (try to) maintain the current a high voltage is produced because there is a high resistance between the coil ends, so the only hope of getting the original current is a high voltage according to Ohm's law. Adding a diode provides a low resistance path for the current, preventing the high voltage spike.

guix:
Isn't the diode supposed to protect against reverse voltage spike when the relay coil is de-energized ?

Indeed, it is on the PCB you show for that reason, but you are not going to disconnect or switch off the relay in your circuit. When you turn off the mains supply, the voltage will simply fade down as the capacitor discharges through the relay coil; there is no "spike".

guix:
I'm not sure if I need that resistor. It was originally on the module board where I took this relay, but after looking this board again, I think it was part of a "RC snubber". It's a weird board where there is two connectors,

No, the resistor was there because the supply generated more than enough voltage for the relay. There is nothing weird about the relay board, AC goes into the two pin connector and the three pin connector then feeds DC power to some other piece of circuit which provides a positive signal to the "ECO" terminal to switch on the relay.

Paul__B:
Indeed, it is on the PCB you show for that reason, but you are not going to disconnect or switch off the relay in your circuit. When you turn off the mains supply, the voltage will simply fade down as the capacitor discharges through the relay coil; there is no "spike".

That makes sense :slight_smile:

Paul__B:
No, the resistor was there because the supply generated more than enough voltage for the relay.

I'm still not sure if I need a resistor at all, or a different value ? The PSU, with no load, outputs around 14.5V, but the relay is supposed to work with up to 15.6V and relay's coil resistance is about 270 ohms, according to datasheet.

So there you are. :grinning:

Sorry for delay, I was waiting delivery of a cigarette lighter cable, from which I replaced the cable anyway since it was so thin inside!

Project is almost done, I just need to cut an opening in the cover for accessing the cables


Thanks all especially Paul B for the relay idea, works great :slight_smile: