Parallel triac (and loads) from single optotriac driver?

Gents,

I am building a simple circuit to turn my main garage lights on with my garage door opener. The model I have does NOT have a normal light and instead uses RF to turn on a small (almost useless) remote light and that only works half the time. Anyway, that part of the circuit is being handled by an ATtiny25 and is pretty trivial.

The lights I want to fire are three, 8 foot dual tube 110w/tube florescent light fixtures. The triacs I have on hand are Q2015L5. From looking at the temperature derating with the current I would need to pull, along with the heat sinks I have available, using a single triac to drive all three lights would be pushing the case temperature to the limit on a hot day when the lights had been left on for a while.

My first thought was simply to use three triacs and three heatsinks, one to power each fixture at 2 amps. I would like to use a single MOC3063 ZC optotriac to switch all three triacs on. If I have read the datasheet on the MOC3063 correctly, it should be able to easily provide enough current to properly switch all three triacs at the same time (or close enough for me!).

The question I have is this: Are there other "gotchas" that I should be aware of? I know I could simply use three optotriacs and trigger them from the same I/O pin, but I would like to keep it down to a single one if possible (aka: I have a lot more triacs readily at hand than I do MOC3063!). So, will this work, so should I go to Plan B?

Also, I am planning on connecting the heatsink thermally to the light fixture to allow it to dissipate heat that way and keep the temps down. Do you think that will buy enough margin to give it a try with a single Q2015L5 and a single heatsink? I guess worse case it the lights quit working automatically after something dies on a hot day and I get to redesign the circuit.

Thanks for the help!

Florescent lighting often have problem with triac switching because of poor power factor of florescent lights, at least on startup even if they are run state power factor corrected.

Most electric eye photoswitches use a bi-metalic contact heated by a half watt resistor in series with photo-resistor.

Surest solution would be to use a relay.

Even with electronic ballasts? What should I test to make sure these lights are happy with the triac?

I find that a small resistive load (such as a 20-40W incandescent bulb) in parallel helps to make most triac based controllers happy with fluorescent loads

I worked up a test circuit with the MOC3063 and three of the Q2015L5s triggered in parallel from the MOC. Everything triggered fine. I attached a "load" to one output (two 27w CFL lights) and they worked fine. I left them on for several hours with no issues and then turned them off for several hours, all with no funny behavior and the triacs never even got warm (with only 1/2 amp, I wouldn't expect them to get TOO hot!).

When I get my monitor for my scope back (my TDS 554A has a bad display, so I just use an external monitor on it), I'll take a look at the voltage and current to the gates and loads and see how things look. The next step will be to trigger one of those fixtures. Thankfully, my workbench has one of those exact fixtures sitting right above it...