controlling CFLs with SSRs

[g2c]

Hello,

I'd like to automate some functions in my house, particularly, turn on and off CFL lamps with a triac coupled to the logic via a zero crossing optocoupler. I saw that many people have functional and reliability problems and wondered if one of you could point me to a robust design: i have no T&M equipment My CFLs are 220V 50Hz with power is in the 3-30W range.

Thanks in advance for your help

 

[tpetar]

Just switching is needed On/Off ? No dimming?

What is T&M equipment ?

Search for solid state relay on EDABoard.

 

[g2c]

Thank,

Yes, on/off only; no dimming.

T&M: Test and Measurement, in other words lab stuff

 

[betwixt]

Just use then as normal switches, CFLs only cause problems when people try to dim them. I switch them here with simple triac circuits without any problems.

Brian.

 

[g2c]

Just use then as normal switches, CFLs only cause problems when people try to dim them. I switch them here with simple triac circuits without any problems.

Brian.

Thanks, I am not trying to dim them. Can you describe your diagram? do you use a series thermistance? Do you control CFL of < 5W? Thanks again

 

[tpetar]

Try this circuit:

https://www.electronics-lab.com/projects/misc/005/

 

[g2c]

Try this circuit:

https://www.electronics-lab.com/projects/misc/005/

Off state courrent (R3,C1) is in the order of 7mA (@230V/50Hz). Good chances that the CFL won't be really off :???:

 

[betwixt]

I spotted that too. Normal current for a 30W CFL at 220V would be around 140mA so restricting it to 7mA might be enough to stop it working but it might flicker instead. I would be inclined to drop C1 to 0.01uF to increase it's reactance to ~320K and see if it worked. The other alternative is to put a resistive load across the CFL but that wastes energy.

Brian.

 

[g2c]

I spotted that too. Normal current for a 30W CFL at 220V would be around 140mA so restricting it to 7mA might be enough to stop it working but it might flicker instead. I would be inclined to drop C1 to 0.01uF to increase it's reactance to ~320K and see if it worked. The other alternative is to put a resistive load across the CFL but that wastes energy.

Brian.

Thanks Brian, What is C3 for anyway?

 

[betwixt]

I think you mean C1.

It's primary purpose is to stop the triac firing until it's the right time for it to do so. Triacs and SCRs can be triggered either by applying the correct gate voltage or by a rise in voltage across them faster than allowed. In other words they are very prone to misfiring in the presence of interference. Of course most AC supplies are full of interference spikes because of loads being switched on and off and you wouldn't want them to turn the triac on at the wrong time. It can be wired as a 'snubber', just a resistor and capacitor in series across the triac A1 and A2 connections but the MOC3021 is also prone to the same misfiring so connecting the MOC trigger to the capacitor also protects it from the same problem with fewer components. The idea is that a spike has faster edges than normal AC cycles and the capacitor is more able to absorb it (lower reactance) so it helps by diverting the spike current and slowing it's rise and fall time.

Brian.

 

[chuckey]

I have come across this flicker problem using a CFL in place of a filament bulb in a PIR controlled light. Superficially its output stage should be a switch but. . .
Frank

- - - Updated - - -

I have come across this flicker problem using a CFL in place of a filament bulb in a PIR controlled light. Superficially its output stage should be a switch but. . .
Frank

 

[betwixt]

Older PIR suffered that problem but newer relay operated types are OK. Many of the older ones were powered by leakage through the lamp, they only used two wires to the sensor circuitry and the lamp filament was used at low current like a resistor. When CFLs are used, the leakage is through the rectifier and reservoir capacitor inside the lamp and as it charges the PIR starts to become starved of current and at the same time the CFL gets close to ignition point. When it eventually strikes, the lamp flashes, discharging the capacitor again and the cycle repeats.

Brian.

 

[g2c]

I think you mean C1.

It's primary purpose is to stop the triac firing until it's the right time for it to do so. Triacs and SCRs can be triggered either by applying the correct gate voltage or by a rise in voltage across them faster than allowed. In other words they are very prone to misfiring in the presence of interference. Of course most AC supplies are full of interference spikes because of loads being switched on and off and you wouldn't want them to turn the triac on at the wrong time. It can be wired as a 'snubber', just a resistor and capacitor in series across the triac A1 and A2 connections but the MOC3021 is also prone to the same misfiring so connecting the MOC trigger to the capacitor also protects it from the same problem with fewer components. The idea is that a spike has faster edges than normal AC cycles and the capacitor is more able to absorb it (lower reactance) so it helps by diverting the spike current and slowing it's rise and fall time.

Brian.

Sorry, i ment C1. I understand that a parasitic capacitance between A2 and gate can, for a high enough dv/dt, cause a triggering gate current turning the triac on for half a cycle (or forever if the voltage transient is due to an inductance). Some manufacturer propose "snubberless" and "logic level" triacs (and triac drivers). I'll have a closer look at these

 

[g2c]

lab setup and tests results in https://www.electro-tech-online.com/threads/controlling-cfl-with-ssr.134251/#post-1154459

 

[betwixt]

Interesting but predictable results, thank you for doing the tests.
One thing you didn't mention was the residual leakage current. Although low it is still dangerous so you need to take extra safety steps when replacing a lamp or working on the other appliances. Switching off at the wall doesn't completely isolate the power so the wiring can still be a hazard.

Brian.