Hi, I somewhat understand how an electromagnetic type of a photocell works. Please correct me if I'm wrong; it has a CdS photoresistor integrated in a relay(?) circuit. The photoresistor is an ambient light-controlled variable resistor that has a resistance which decreases with increasing light intensity.

But how the heck does a "thermal" PC work? How would I recognize one if I see it on sale or ontop of a fixture? Thanks!

I also don't know how they work, but I imagine the photoresistor controls a heating element, and when it gets hot enough, it switches a thermal switch and the light turns on.

But how the heck does a "thermal" PC work? How would I recognize one if I see it on sale or ontop of a fixture? Thanks!

The same, just instead of an electromagnetic relay (so an electromagnet actuating a mechanical contact) it employs a thermal relay (a heater heating up a bimetal or similar device and that then actuates the mechanical contact).
There are two reasons for that:
First the thermal relay suffices with lower control power (assume there is no need for any fast response; the AC electromagnetic relays use to consume around 0.5W, while when ambient temperature compensated, the same contact load thermal relay consumes about 0.2W or even less). That means it suffices with a smaller (so cheaper) CdS cell.
Second reason is, the intrinsic slow response of the thermal relay means it provides quite effective filtering of either shades or light flashes, so the photo control is then less susceptible to these disturbances.

Thanks Medved.

Ok, so a mere 0.3 watt difference would not entice me to use a thermal PC over an electromagnetic one, and neither would the less "wrong" switching with thermal PCs.

As an outsider (of the field electronics & of lighting) I find a few things or methodology employed sometimes strange. Which begs the question: Were the 1st photocells employed thermal type?

Thanks Medved.

Ok, so a mere 0.3 watt difference would not entice me to use a thermal PC over an electromagnetic one, and neither would the less "wrong" switching with thermal PCs.

As an outsider (of the field electronics & of lighting) I find a few things or methodology employed sometimes strange. Which begs the question: Were the 1st photocells employed thermal type?

Unless you have 10,000 fixtures, then you save 3000 Watt Hours every hour!

Which is still negligible compared to what 1E4 fixtures use when switched on, so better mind that the photocell is switching on the light exactly when needed to minimie working hours, and not about the photocell's own power use

Which is still negligible compared to what 1E4 fixtures use when switched on, so better mind that the photocell is switching on the light exactly when needed to minimie working hours, and not about the photocell's own power use

I have a weird idea to save energy, a solar powered photocell!

Hmmm, ok.

A town with 10,000 street-lights, the average of which is lets say 175 circuit watts, this yields 1750 kwhours every hour of operation. This is 7,350,000 kwhours annual consumption (10,000 x 175W x 4200hrs/yr).

Still, 26,380 kwhours in savings (3kwh x 24 x 365) is relatively insignificant in my opinion.

If you use the solar energy directly (sunlight heat to move bimetal contact), you are going to have a problem on cluody days, when there is enough light (so the lamp can be switched off) but no heat to activate the photocell

If you use solar cell to power electrical/electronic components, gotta weight in the impact (lets say environmental impact) of the solar cell manufacture and disposal vs. the electricity it would use from the line with standard cell

Lights*plus And thats why a cell that makes the lights work 3 minutes less every day saves more energy than the cell itself uses all the time

The point of saving the 0.3W dissipation is not in saving the energy directly, but having less thermal load on the CdS photoresistor itself. And that means that photoresistor could be made smaller, so cheaper and so using less of the toxic Cd.
The thing is, with an inductive nature of the relay coil, the 0.5W of the coil means about 1 to 2VA of apparent power. If you want to control that by a variable series resistor (the CdS cell itself), you end up with 1 till 2W maximum power dissipation on that resistor (that means more than a square inch plus a heatsink plate).
With a 0.2W heater, the maximum power dissipation is less than 50mW (maximum is, when the CdS resistance is the same as heater resistance, in that case there flows half of the current and half of the voltage across each resistor). That means the typical 8mm size (so 1/10'th of thatrequired by a 1VA electromagnet) is way larger than required to dissipate the power, so runs way cooler and/or needs no heatsink at all (so simpler photocontrol construction).

To really save the power, way more important is to make sure the light gets ON only when it really needs to (so not to early ON and not too late OFF). That means a bit more intelligent control than just a single threshold pair (forming the hysteresis) and better accuracy. And that is the domain for the electronic controls.