Okay, I gave it a shot. You didn't mention the wattage, efficiency, or light output of the HPS fixture.

If I knew that, I could try to figure out the lumens of X output on a 81Sq.Ft. target 12 ft away.

But since I don't, I'll just guess ! It's much easier than math.

Have you seen the lights used on Movie sets? They generally have a rectangular reflector with adjustable flaps.

Make one of those. Go to Harbor Freight and get a hand-cranked winch, the hardware store and get some rope, a pulley, some chain, and some screw hooks.

Put the screw hook or whatever up in the peak of the GH, put the pulley on the hook, run the rope through the pulley.

Mount the winch at a comfortable height, put the rope on the winch and the other end of the rope, attach to the chain on your HPS light.

Now, crank the light up and down until you find the right height.

As for the second question, you can get a photocell plugin for a few dollars. They're designed to turn on lights after dark.

If you could find a really good one it might even have an adjustable setting.

I put in a floodlight at our business last year. I climbed down from the ladder, threw the switch and waited for it to get dark.

Finally, the light came on ! I was so happy.

And then the light went out.

I wasn't happy anymore.

And then the light came back on !

But this time I was suspicious.

I watched it go on and off, on and off, and finally realized:

The light went off every time a car drove past the front of our business !

I stood there, thinking furiously. The sensor must be too sensitive, I decided. The headlights of cars were tripping the light sensor to the OFF position.

So, I could go buy another sensor.

Nope. Didn't want to do that.

Can't take it back either, the floodlight had sat in it's box for a year before I got around to mounting it.

So instead I climbed up the ladder, stuck a plastic flower pot over the sensor, and lashed it in place with black electrician's tape.

Now, the headlights can't trip the sensor. Only true daylight is bright enough to shut the light off.

So if you can't get an adjustable sensor, try the flower pot trick.

Jkirk described the reflector setups you'll need. You have absolutely no chance of designing or building the ideal curved reflector for your job without some sophisticated machinery. Using a masked daytime/nighttime photocell is a good idea, try taping over the sensor to make it a sunlight/shade sensor. An actual calibrated foot-candle sensor will be quite expensive. I'd be tempted not to even bother with the sensor, just run the lights continuously during the day and maybe use a little less intensity.

Here's how to calculate foot-candles. Assume 20% loss, this is absolutely the best you can hope for, between absorption at the reflector and misdirected light going out the windows. Divide total lumens by total square feet, equals foot-candles. Example, 100,000 lumens over 100sf equals 1,000 foot-candles. This assumes an even light distribution, which you won't get. It will be brighter right under the bulb (for HID or standard fluoro tube) and fade out towards the edges and corners.

"It will be brighter right under the bulb (for HID or standard fluoro tube) and fade out towards the edges and corners."It will be brighter right under the bulb (for HID or standard fluoro tube) and fade out towards the edges and corners."

Hey, Shrubs... I'm thinking I once saw a light-diffuser... a grill of 1/2 inch slats like Venetian blinds...

Does this ring a bell? A way to avoid the hotspot right under the lens and bounce light to the corners?

Yes, there are light diffusers of various kinds. The problem is that they aren't very precise, so the chances are you'll bounce more light out the sides of the greenhouse than into the corners.

I figure I really need to add about 1000 or more foot-candles to be effective. My greenhouse will be 10 X 18 as it now looks. Here's what I came up with:

4 800W HPS lights, 2 in each reflector (several companies sell these double reflectors), hung 9 feet apart in the center to overlap a bit in coverage. Assuming I get a 10 X 10 distribution for each one:

Each "double light" in it's reflector will put out about 220,000 lumens or so (1600 watts a piece). Assuming I get 65% efficiency (estimating low), we have (220,000 X .65)/100 square feet (10 X 10 coverage area):

1430 foot-candles. I could probably get away with a bit less, but I am willing to spend a few extra bucks to actually harvest citrus in winter. Otherwise the greenhouse is no brighter than my windowsill-maybe 400 foot-candles inside the GH on a cloudy December day. Granted, at that rate my lighting in winter will cost more than my heating does!!!

I was going to try and get a sensor to kick it on when light is dim, but still have it off at night and it appears this is too cumbersome. I'll probably just run them 10-12 hrs a day from Nov. 1 thru April 1 or something like that-spring thru early fall we should have enough natural sunlight.

Whoops-I screwed that up. They don't make 800W HPS, do they? I meant 600.

Which is about 168000 lumens, at 65%-

(168000 X .65)/100 = 1092 foot-candles. I thought 1400 seemed high! But 1092 f-c is more like what I came up with last night.

Is that going to be enough to make the citrus fruit? Its only about as bright as a cloudy day in summer.

BTW, I don't think heating is going to be a problem once you install all those lights ;)

It'll have to be-I can't afford to run 6000W worth of lights!!

If you add that to the average of 400 natural foot-candles, that's almost 1500. That's getting there, and the occasional sunny day will boost it over 5000. If it doesn't work I'll add another light in the middle. I'd be interested in seeing what Florida gets in January.

Solar intensity at noon in southern Florida in January is almost 90% of what it is in June, but the days are a little shorter and they are more cloudy. The sun is roughly as strong in Key West in January as it is in Cleveland in June :)

Of course it is! But-citrus can't utilize more than about 3000 foot-candles-anything over this is irrelevant. I was talking from terms of total daily "units" of solar energy. Basically a function of day length, intensity, and cloudiness. There was a map online that showed this and in July Cleveland and Miami have similar numbers (because that's Florida's rainy season) but Miami's January numbers are identical to Cleveland's April-May numbers. It's that November - March period I have to worry about.