Controlling CFL with SSR?

[g2c]

Hello,

I intend to automate some function and implement an alarm in my house using an Arduino.

First i'd appreciate an advice on how should i control the lights, mainly CFL rated at 5W to 40W, 220VAC. I was thinking of optoisolated SSRs as in https://www.futurlec.com/SSRAC.shtml but from the web i could see people comlaining that SSR's destroy CFL quite quikly. If mechanical relays are preferred, any special precaution?


Same for electric roller shutters: are SSRs recommended? and if mechanical relays are recommended, would i need a protection network?

Many thanks in advance

 

[cowboybob]

Hello, g2c.

I'd use low voltage mech relays for both. Socket them (makes for easy replacement). Very much simplifies the whole process.

And I wouldn't think a protection network would be necessary, although I am unsure what it is you're trying to protect. If it is for the Arduino, then simple monostable 555 timer interfaces and diode "kick-back" protection on the relays should be sufficient.

 

[g2c]

Thanks,

I don't know a big deal about the CFL. I do know that if the relay sees a capacitive load then it should be protected against hi di/dt at 'make', using, say a series inductance and if it sees an inductive one it should be protected against overvoltage at 'break' using, say, a MOV.

I did not understand why won't the arduino (or i/o expender / buffer / level shifter) energize steadily the relay -as long as it want a 'make'. why use a 555?

 

[cowboybob]

Thanks,

I don't know a big deal about the CFL. I do know that if the relay sees a capacitive load then it should be protected against hi di/dt at 'make', using, say a series inductance and if it sees an inductive one it should be protected against overvoltage at 'break' using, say, a MOV.

Since, I assume, you'll be using DC relays, a diode across the relay coil will suffice for inductive kickback. Any inductive problems on the motor side would probably be handled by an MOV.

I did not understand why won't the arduino (or i/o expender / buffer / level shifter) energize steadily the relay -as long as it want a 'make'. why use a 555?

Primarily relay isolation and relay current draw (from the arduino) considerations. A transistor switch would also do the job.

I might add that the shutter motors might also require a "heftier" secondary contactor, fired by the smaller, arduino driven relay/transistor option.

 

[Dean Huster]

CFLs are troublesome devices when you link them to anything with an SCR or TRIAC output. You'd think it would only be a problem with a partial-on TRIAC condition (dimming) where the chopped-up waveform messes with the switching circuitry in the CFL. But I've found that the typical CFL absolutely will not work in my outside motion-sensor light without going totally nuts. And then there's some will will work with a TRIAC dimmer. Maybe shopping for a compatible CFL is the route to go vs. finding an expensive SSR.

 

[alec_t]

I've had no success when trying to use an off-the-shelf solid state switch (PIR sensor or security timer) with a CFL. Either the CFL refuses to light or it kills the switch.

 

[g2c]

Test results

Condition: The whole installation is 220V, 50Hz

CFL test

First, to determine the aging cycle i measured the temperature. Test shows with the 20W CFL goes to 75 deg C in 3 minutes and reaches 88 deg C in 10 minutes. Afterwards, no more temp increase. With 10 minutes off, temp goes down to 32 deg C.

The chosen edging cycle was 10' on / 10' off with two CFLs: one is 20W, the other is 32W, and they are powered in opposed phase (when one is on the other is off). No significant aging done with the 9W lamp

Results

Zero failures in 1184 cycles for each of the two CFLs

9W CFL
Test purpose was to check that when turned off it indeed goes off

Result: in the dark, seems completely off

Fans test

Fan consumption: 180mA

Aging cycle: 1' on / 1' off

Results: zero failures after 3166 cycles

A second 245mA fan tested with ~ 100, as above cycles with zero failures

Rolling Shutters test

Tested shutter consumes 1.2A when going high (as well as when going low)

With ~ 1W/A, triac dissipation is ~ 1.2W, Tj =30+50*1.2=90°. As max rating is Tj mx=125° => no need for heat sink. Furthermore the measured triac’s case temperature, as measured on the tab of the TO220 connected to MT2, did not increase over 40deg C during the shutter travel from lowest to highest position.

Results: 10 cycles with zero failures

Conclusion

Driving fans, CFLs and rolling shutters with circuit of attached schematics seem to be remarkably reliable.