Led Lamp and relay problem

[Delphes]

Hello,

I face a very strange problem since few days without any solutions, I used a relay to switch a led lamp (led lamp is 48 SMD led direct 240 Volts GU10), after few switching let says 10 the relay contact stayed ON for ever, the relay i Used first was Finder serie 32 which can handle easily 5 amp under 240 volts, I decided to test Tyco relay and the problem was the same may be later after 20 switch ON/OFF.
I tried to put a MOV to protect relay contact again overvoltage during switching, problem was the same.
Could somebody can explain me why a simple 3,5 watts led lamp can break a 5 amps relay, and How I can solve it ?

regards
David

 

[Diver300]

Welcome to the forum.

I suspect that the lamp may have a reasonably large capacitor in it to keep the LEDs alight for the whole of the mains cycle. What may be killing the relay contacts is that the inrush current is welding the contacts closed.

A simple resistor in series will probably help a lot but one of these:-
EPCOS|B57236S800M|THERMISTOR, NTC, 80R, 1.6A | Farnell United Kingdom
has a lower voltage drop as it will heat up and reduce its resistance.

 

[Delphes]

Thanks Diver300, I will try as soon as Farnell deliver me, in the mean time I tested a SSR from Crydom to see 'whats happening" with this Led Lamp, When my relay is off my led lamp is blinking (YES !!) and when my relay is on, of course normal light, I do not know if this information can help to understand what's happening ?

regards

 

[Diver300]

I think that there is a suppressor circuit in the SSR. That leaks some current into the LED lamp. The current slowly charges the capacitor in the LED lamp, and when that gets to certain voltage, the circuit in the LED lamp turns on, and lights the LEDs. However, that takes more current than the SSR leakage, so the capacitor discharges.

If there is a capacitor in the light, it could well be causing this and the contact welding.

 

[Delphes]

Hi Diver300,

I finally tested with 100 ohms resistance in series to see whats happening, with only one led lamp and about 1000 switch ON/OFF I was not able to reproduce the problem, so it seems that your solution solve it, I have to confirm with more than one lamp in parallel, myquestions are :

1) Why did you select a thermistor instead of a normal resisstance ?

2) If I put 10 led lamp in parallelle (which is worst since the total capcitance will increase), does the 100 ohms resistance value will stay the same ?

3) Could you explian clearly why this resistance in serie avoid relay contact to stay ON forever ?

Many thanks

 

[Diver300]

1) A thermistor reduces the voltage drop. You need an NTC thermistor, so that the resistance is largest at turn on, and then the thermistor heats up and its resistance decreases, so there is less voltage drop.

It depends on your load, and how sensitive it is to supply voltage, as to whether a thermistor would give you any improvement at all over a simple resistor.

2) The 100 Ω series resistance was a guess. To get the same voltage drop with 10 lamps you will need 10 Ω, and that could well be too small to protect the relay. You still haven't said what the power rating of the lamp is and without that I don't know what resistance it could stand. You would be able to use more lamps with an NTC thermistor than with a resistor.

The peak current with a 100 Ω resistor and a 240 V supply is about 2.4 A. However if the lamps are 15 W, and you have 10 of them, there will be a significant voltage drop with a 100 Ω resistor.

3) The relay is failing because you are massively overloading it. When you connect a capacitor to a supply, it charges extremely quickly. The current is only limited by the resistance of the wires. You could easily be getting 100 A peak current, which is far more than he relay can handle, and it occurs as the contacts close, so there isn't much contact pressure. You are simply welding the contacts together.

Relays are fairly robust, and you don't have to keep the current spikes to less than 5 A for a 5 A relay, but it is a good idea to keep the peak current to no more than a few times the continuous rating.

I have had the same problem with a PIR light switch turning on a fluorescent lamp that contained a PFC capacitor. A thermistor stopped it happening again. There is also an issue turning on low voltage halogen lamps as they can take massive currents when the filament is cold.

 

[Delphes]

Dear Diver300,

Thanks for your quick answer, my led lamp is 48 SMD led 3,5 watts so 10 lamp in // will power around 35 watts, so which thermistor value in that case will be suitable ?

regards

 

[Diver300]

OK, that will take about 35/240 = 0.145 A. If you accept a 10 V drop, then you can use a 10 / 0.145 = 68 Ω resistor. That will limit the turn-on current to about 4 A, which is fine as it is less than the relay rating.

The resistor will dissipate about 1.5 W so a 5 W resistor would be fine.

The thermistor that I quoted before (EPCOS|B57236S800M|THERMISTOR, NTC, 80R, 1.6A | Farnell United Kingdom) is about 80 Ω when cold, and will be a lot less when hot, but I can't work out how much less. Anyhow there would be less voltage drop.

The 50 Ω version would also protect the relay as the current would be less than 5 A and would have lower voltage drop.

Neither costs much, and a 5 W resistor will not be much cheaper.

 

[Delphes]

Dear Diver300,

I was able to test in real situation this afternoon, and the result was not black not white, let me explain, first I got this from Farnell :
MURATA|PTGL07AR700H8B52A0|POSISTOR, PTC, CIRCUIT PROTECTION | Farnell United Kingdom since I discovered I had few in stock here, i do not know if the caracterictic is great but let me explain the result, with one led lamp no problem I was able to switch ON/OFF 100 times with no problem. BUT with the 10 led lamp I made the following test switch ON WAIT 3 sec and switch OFF after let says, 1 minute, the led lamp was just blinking instead of staying ON for 3 secs, I saw that the Posistor was little hot and smelled like hot, I stopped my test. In every case relay was protected and never stayed ON, so it was good. Do you think this posistor was not ok due to his limit current to 0,4 A ? I will try to test your thermistor , any idea about whats happened ?

regards

 

[Diver300]

I think that the lamps are taking a lot more than 0.0145 amps each. They might be using the capacitors as current limit devices. (google "transformerless power supply") and so the power factor is very poor.

You need a lower value PTC, fewer lamps, or you could just use a 2 or 3 PTCs in parallel.

Edit:- This was wrong. I didn't notice that Delphes had used PTCs where he should have used NTCs

 

[cowana]

No! A PTC is a positive temperature coefficient resistor. That means that as it heats up, the resistance increases, so it dissipates more energy, heats up more etc.

You need a thermistor - a NTC negative temperarture coefficient resistor. That means it will limit the inrush current, but then heat up, the resiatance drops, it heats up more, the resistance drops more and it dissipates less power.

A PTC is ideal as a permanant current limit for low power applications (eg protecting a PIC circuit - as it heats up, it will drop more voltage).
A NTC is ideal to limit startup current for your sort of project.

Andrew

 

[Diver300]

Thank you for that. I had suggested NTC, but Delphes had used a PTC. I didn't notice that he had changed that and repeated his mistake.

NTCs are what is needed here.

I guess that even a small resistance will reduce the inrush current and save the relay.

 

[Delphes]

Dear All,

I just ordered 9754229 at Farnell, will test and keep you inform, these led lamp will give me crasy, you have to think that I used 5 amps relay wich can handle let say 1000 watts at least, and that my 10 led lamp are supposed to get 35 watts !!!

Thanks again for your help.

 

[cowana]

That looks like a good solution.

At room temperature, an initial current of 2.5A would make an initial energy dissipation of 82.5v, so 200W. That will heat it up quickly, the resistance will fall quickly, and the power dissipation will fall to something very low (~3W). Just be careful not to touch it - it could be pretty hot!

 

[Delphes]

There is'nt any way to measure the inruch current ?, so we can be sure where we are going. I am little afraid to have very hot NTC has you said it is located in a kind of critical place, any other value will do the same ?

rgds

 

[cowana]

The only way to measure inrush current is with a shunt resistor and storage scope. As your load is only 35W (150mA) - it probably won't get very hot. You'll have to try it and see, but I wouldn't worry about it until you've tried it.

Andrew

 

[Diver300]

NTC inrush limiters will get hot, and the lower the value, the cooler it will run. However, the temperature stabilises because if it gets hotter, there is less resistance so less heat. An NTC of half the resistance when cold will generate somewhat less power, but still more than half the power.

Any NTC will generate less power than a fixed resistor that has the same limiting effect.

You could use two relays. One to turn on the light, and one to short out the resistor or NTC. That method is usual for much larger loads and isn't generally needed for small loads.

 

[Delphes]

Dear,

So I tested with NTC as you receommended and stress test with 10 led lamp did not make any relay failure, so great news, the NTC was not hot as I expected. Now I just have to find a good NTC value for every case like one led lamp (from 1 to 4 watts) and 10 led lamp (from 10 to 40 watts total) , I am not afraid by voltage drop since I think led lamp works with 220 volts also (220 to 240) , I prefer find a good value thats protect the relay in any case, so I think 50 ohms is good compromise, any idea ?

regards and thanks again