On Wed, 16 Jun 2004 06:47:31 +0000, The Outcaste wrote:

snip

Thanks for the links. I've been looking at the wrong manufacturers. Now
it's just a matter of deciding on a model and finding somebody who might
sell me one with minimal fuss. That last part is always trouble when it
comes to uncommon lighting hardware. It's like buying a used car...

I'd thought about a micro controller, but haven't looked into the cost
of one, especially the cost of the hardware to program one.

The tiny PIC microcontrollers start at less than $2 a pop. More
substantial models can run more like $8, but those would be overkill. My
guess is that something in the 16xxxxx series would have enough pins to
drive a DAC, take one or more inputs (maybe a potentiometer for
controlling the speed of sunrise/sunset, for example), and have plenty of
spare cycle time to do whatever else you might like for about $4. You can
find information on the whole line at http://www.microchip.com

As for programming hardware, I use a Warp-13 from Newfound Electronics as
it plugs nicely into the MPLab IDE provided by MicroChip. You can also
use it to program a number of different ROM chips. About $100 (you can
get cheaper programmers, but this is an excellent all-around product).
Once you get used to the microcontroller thing, you spend a lot less time
thinking about digital logic. The first thought that pops into your head
is, "I can do that with a $3 microcontroller and a few lines of C code."

I'm also considering using one of my old 486 PC's to do the job. As
they can be picked up for $5-$10, this might be a low cost option.
It's also work as the power supply for the DAC circuit.

I'm still trying to figure out the power supply part. A lot of
microcontrollers can deal with a wide range of supply voltages these days,
but I don't want to have to run more wires than I have to. I seem to
remember some ballasts coming with low voltage DC supply lines for
sensors. It's one possibility, anyway.

I have one that has 3 printer ports, so I could set it up as a print
server as well as to control two sets of aquarium lights, each with
their own program; wouldn't even need a keyboard, mouse, or monitor,
just a network card and then access it via VNC or PC anywhere.

Load it with DOS and write a QBasic program, or load Win95/98 and use
Visual Basic. Just have to spend a day or two to learn enough Visual
Basic to write the program to control the printer ports. Use 1 bit for
on/off, leaving 7 bits to control 128 different light levels. Could
even use the 2nd printer port to control water valves, and simulate
mid-day T-storms, even annual variations.

This sounds like a job for Linux! Or a microcontroller with a cheap
wireless transciever (probably about $20 for the whole package). And
blue LEDs for moonlight... But now I'm getting ahead of myself. Getting
light without smoke, fire, or that horrible **POP** is step one.

One other advantage to the dimming ballast is getting longer tube
life. Rather than changing the bulbs every 6 months as is sometimes
recommended, simply add in one more tube than you need, say 5 instead
of 4, and set your max intensity at 80%. As the tubes age, you
gradually increase the intensity to make up for the drop in intensity
as they age. This way you can use them for their full life span of 1-3
years; though with the newer T8 and T5 tubes (at least for CW, WW,
full Daylight tubes. Not sure if this applies to the specialty bulbs
for plants though), it seems they hold their intensity throughout
their life span, after an initial drop in the 1st 100 hours, so this
may not be very useful.

I hadn't thought of it from that angle. It's an interesting side benefit.
I think that this would be well worth playing with. After getting a
workable design, these would be pretty easy to make. If I get to that
point, I'll let everybody know.

Thanks for your help.

-Matt