Electronic – Create a circuit with 3 identical light bulbs that all have different brightnesses

Questions

• indented like this.

Answers on margin.

• I'm getting a streetlight, most likely with a HPS bulb (high-pressure sodium). I'll probably hook up some sort of sensor to turn it on only when someone is actually parking lot.

HPS is not a good choice for rapid turn on use. You'd probably be better served by a more standard bulb such as a Halogen, or something exotic like HID, or 'solid state' like LED.

• Why do some streetlights seem to take a long time to turn on? Is there anything that can be done to get parking lot lights to turn on quicker?

Long turn on is caused by the technology used. Usually this involves the vaporisation of a metal and building up temperature and pressure. There is nothing sensible that you can do except change technologies. If you have a military-type budget you may find a faster way but for mere mortals, buying what is provided is typically the only choice.

• Why do some companies say you must let the bulb cool down for a minute or so after you turn it off, before you can turn it back on? Is that a limitation of the bulb itself, or the control ballast, or what?

Most bulbs that use a gas arc or ionisation of a gas require a starting voltage which is significantly higher than the running voltage. When the tube is hot the starting voltage can be higher again - perhaps by a factor of 10 in some cases. While starters can usually be made to "strike" the tubes in such cases. Since this adds extra cost it is usually not justified in applications where only a few people occasionally need to restart lights.

In specialist cases where failure to restart is unnacceptable, the effort and expense is taken to provide the extra voltage. An excellent example is HID lighting for cars. An HID light has a very small sized arc and white light. The small arc size allows excellent optical handling for beam forming (as in cars or projectors). The white light is attractive for various reasons. When cold an HID tube requires many hundreds of Volts to strike and when hot it requires many kV - perhaps 5-10 kV. HID controller makers for lamps used for fixed lighting may decide to not provide the HV restrike capability - it adds cost and dealing with HV is a reliability issue. But in a car you must be able to turn the headlights back on instantly at any time. So all automotive HID controllers have hot strike capability.

• Can "dimmable HPS" systems reach full brightness from a nearly-off dim state quicker than traditional on/off HPS systems?

I don't know, but, almost certainly yes. Once an ionised start is obtained, if it can be stably maintained at low power, you'd expect increasing the temperature and pressure to higher levels to be much quicker than bringing them up from cold. Manufacturers' data sheets should comment on this.

• The only dimmable HPS systems I've seen so far can dim down to a minimum of 1/2, which visually seems only slightly less bright than full on. Is there some reason HPS can't dim down to 1/10 or 1/100 like other lighting technologies? (Am I just not looking in the right place for dimmable HPS systems?)

Again, an informed guess: The minimum energy required for stable ionisation is still fairly high and if you try to reduce the brightness below that, the light simply turns off. Hence my caveat above, "if [stable ionisation] can be stably maintained at lower power". I'd expect better than 2:1 to be possible but probably not vastly better.

• How do you dim a streetlight bulb -- how is the voltage/waveform applied to the bulb different at dim than at full power?

That very much depends on technology and on regulatory requirements. Where a bulb uses advanced technology and is sold as "dimmable" the manufacturer will provide very specific instructions. Technologies that accept linear current drive variation (e.g. incandescent, halogen and some LED systems) will allow very simple waveform control - either TRIAC leading edge waveform shaping or leading and falling edge or a power factor corrected reduced envelope. Other technolies may get a special waveform and conditions from a special controller and may be considered a "black box" from the entrance to the controller onwards. For example, some LED systems use 2 or 3 stage hill and valley circuits to fill the "valley space" around mains zero crossing points with energy taken from the cycle peaks. and they THEN apply a buck converter. The input to the dimmer is the start of the special magic and you then do not lightly separate the LED 'bulb' and controller. An HID controller will almost invariably be very close to its bulb. (People don't like reticulating 10 kV).

• Is there a standard protocol for sensors to tell parking lot lights that motion has been detected, so the lights need to turn on?

I dont know. But no. This is a simple task and there are liable to be various ways of implementing a fairly straight forward choice. In most cases a simple timeout will do. Complexities such as usage patterns, time of night, etc. can easily be added if a microcontroller is used in the system.

• Are there any reasonable alternatives to HPS bulbs? If Haitz's law continues, then someday LEDs will cost less -- but the current LED lamps that last 10 times as long, but cost 100 times as much to get the same brightness, don't seem reasonable in 2011.

Yes. How much light do you want?.
There are many issues other than life and cost alone.
LEDs offer almost the best to the best energy use when running, instant turn on, instant restart, wide range dimming, low temperatures (low stress on other components), low maintenance and replacement costs, etc.

HID bulbs in the 30-200 Watt range are available for automotive use and would be trivially adaptable for street lights. No doubt they are also available for streetlights formally.

Halogen bulbs have lower efficiency but offer instant on, dimmability, instant restart and excellent CRI (colour rendition).

You could build a mechanical shutter operated by a servo motor and controlled by a microcontroller easily enough, but there is another issue.

Your 400W of input power is being turned into light and heat. I don't know what the efficiency of your proposed bulb is, but even if it was as high as 25% (dubious) you'd still need to dispose safely of 300W of heat continuously. When the shutter obstructs the light the waste heat will probably ramp up to almost the full 400W because most of the 40k lumens are going to be converted back into heat.

Have you looked for high power LED's (or LED arrays) that can generate the kind of light intensity you need? LED's can usually be switched on and off pretty much as required, and they're also generally more efficient than other light sources.