2
\$\begingroup\$

I use an Arduino to control a relay. This relay is on an 8-relay board. The relays are rated at 120 VAC, 10 A.

This relay powers on and off an AC-DC power supply. It has 110 V AC input and is rated at 12 V DC, 20 A.

The load is an LED strip. I have already 7 or 8 burned relays from different LED strips. It doesn't matter if it is a 1 m long LED strip or a 16 m long led strip. The current, in both cases, is below the rated 20 A of the power supply.

The relays usually stop working when turning on or off the relay. Just one stops working. The other relays on the board are still operational.

I'm pointing at the power supply as the the guilty one here as I have hundred other loads that didn't burn. And some of the loads are higher than this power supply.

What could be causing it? How can I avoid more relays being burned? Is there a "diode" for AC that could stop it from burning? Do you see anything else besides the power supply that could be causing it?

Simple circuit drawing

PS: By burning I mean that it just stops working.

EDIT: The relay board is a cheap AliExpress 8-relay board. The relays are JQC-3FF-S-Z. The datasheet that I found is: https://www.digikey.com/htmldatasheets/production/2071105/0/0/1/jqc-3ff.html

EDIT 2: for future internet users, as described on the answers below, the issue was with the surge current. I bought new power supplies that had a NTC thermistor device in series with the AC line. It takes a few seconds to turn on the Led Strip, but I guess that's better than burning relays.

\$\endgroup\$
8
  • 3
    \$\begingroup\$ What part of the relay is "burning?" Is it the coil? or is it the contacts? \$\endgroup\$ Sep 5, 2022 at 0:26
  • 1
    \$\begingroup\$ What is the specification of the relays? Can you supply a link to a data sheet? How are you controlling the relays? Where does the "Signal" come from? From what you describe, all the other components, the power supply and the various LED strips still work once the relay is replaced - correct? Without any further information I would guess that it is the inrush current of the power supply that is burning out the relay contacts and hence you need better relays. \$\endgroup\$ Sep 5, 2022 at 0:29
  • \$\begingroup\$ @SolomonSlow, how can I test which one stopped working? \$\endgroup\$
    – anonimou
    Sep 5, 2022 at 0:31
  • \$\begingroup\$ @PeterJennings , the relay is TONGLING JQC-3FF-S-Z . I'm controlling the relays through Arduino. Yes, everything is still operational. Just the relay stops working and replacing it, everything works again. \$\endgroup\$
    – anonimou
    Sep 5, 2022 at 0:32
  • 3
    \$\begingroup\$ @anonimou Do you hear the relay click? If not it's probably the coil, if you do it's probably the contacts. It might not click if the contacts become welded together, but you would know if that happened because the power supply would stay on. \$\endgroup\$
    – GodJihyo
    Sep 5, 2022 at 0:34

2 Answers 2

1
\$\begingroup\$

You should never use dry relay contacts on a power supply input that are not rated for the surge current of charging the capacitor and the inductive stored energy on flyback when shutting it off.

These surges greatly exceed the rated current.

You can however use it to control a resistive load like LEDs as long as it is within the current rating.

Many DC supplies also have an enable control for a logic level or contact closure enable.

Cycling the power supply also adds stress to the supply. Consider if it has a soft start circuit but is cycled too fast and does not have protection.

Most power supplies have a varied lifespan of power cycles due to the added random phase current surge. Your mileage may vary.

PC's have been designed this way for some time to use only a soft logic power switch.

The datasheet of your relay link shows what happens with high surge currents on lifespan. No surprise here.

Although OMRON Relays were 10x better.

enter image description here

\$\endgroup\$
3
  • \$\begingroup\$ Thank you for the answer. Is there a way to limit the surge currents? Would a varistor and a RC circuit better than nothing? \$\endgroup\$
    – anonimou
    Sep 5, 2022 at 1:04
  • \$\begingroup\$ Would a Solid State Relay be better in this case or it won't matter as the rating would still be lower than the surge current. \$\endgroup\$
    – anonimou
    Sep 5, 2022 at 1:04
  • \$\begingroup\$ There is no problem with leaving the power on idle with no load. This is better with an switched output . Yes a ZCS SSR is better but expensive \$\endgroup\$ Sep 5, 2022 at 1:06
1
\$\begingroup\$

The root cause of your relay failure is the high input surge current of the DC power supply due to capacitor charging at cold-start.

You would need to select a relay, from 'high-inrush current' models, considering the TV rating mentioned in the data sheet.

The TV rating, as per UL and CSA standards, is an assessment of the inrush current withstanding capability of relay contacts as follows:

TV3 - 51A, TV4 - 65A, TV5 - 78A and TV8 - 117A, TV10 - 141A, TV15 - 191A.

The 120V, 240W (12 V * 20 A), power supply, that you have used, could have an inrush current of the order of 150 A.

Hence the relay would need to have a rating of TV10 or TV 15.

The relay you are using has a rating of TV5 only.

\$\endgroup\$
2
  • \$\begingroup\$ Thank you for the answer. Looking for solutions, I found about NTC thermistors. Would using something like a NTC thermistors to reduce the insurge current also possibly solve the issue? \$\endgroup\$
    – anonimou
    Sep 5, 2022 at 16:32
  • \$\begingroup\$ Anytime, anonimou! You may try using an NTC thermistor. However, its functioning would be dependent on ambient temperature. You could have start-up problems when it's too cold and improper functioning when it's too hot. It may also be ineffective during power interruptions. You may also use an SMPS-manufacturer-supplied inrush current limiter that utilizes a wire-wound resistor (instead of an NTC thermistor) that is then bypassed by a relay contact. \$\endgroup\$
    – vu2nan
    Sep 6, 2022 at 4:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.