posted by [identity profile] keirf.livejournal.com at 12:07pm on 19/10/2005
I assume it's a modern dimmer that uses a TRIAC circuit to switch the light on and off rapidly. I think they contain capacitors to smooth out the choppiness of the current too.

1. So up to the point that you come to turn the light off, energy use is the same for both cases. If you then turn the switch off, energy use stops, whereas when you turn the dial down a small amount of energy is used until you turn off the light. Therefore dimming and turning off uses fractionally more energy. However, when you come to turn it on again you save the same amount of energy in the dimmed down state (provided you dim up and down at the same rate). But the circuit will use a very small amount of energy to drive the transistors in the dimmer. So the answer is turning it off immediately saves a tiny fraction of energy, namely the energy consumed by the during the period of dimming down and up. Dimming the switch to minimum before turning it on and off uses a minimal amount of extra energy.

2. Light bulbs burn out when the tungsten in a section of the filament evaporates away so much that there's a break. Switching on from cold means more current is drawn because cold filaments have more resistance, but the warming up happens very quickly. This surge can cause the bulb to blow (it's why bulbs often blow when you turn the light on). You would think that turning up the dimmer slowly is like a soft-start device such as a thermister, which will extend the life of the bulb, but we're chopping the current, rather than reducing the current draw, so a dimmer is a very poor soft-start device. And good slow-start thermistors only minimally extend the life of bulbs anyway. So dimming the switch to minimum before turning it on and off will result in a minimal life extention for the bulb.

That's what I think, anyway.
 
posted by [identity profile] cartesiandaemon.livejournal.com at 12:16pm on 19/10/2005
Quick googling gave me

http://en.wikipedia.org/wiki/Dimmer
http://freespace.virgin.net/tom.baldwin/bulbguide.html#mll

but support what keirf said.

My dimmer fantasy was presumably drawn from the days of rheostats: until I looked it up, I always felt leaving a dimmer on half was wasting something.

I wouldn't wager money on these, but it seems fairly clear that if you want to worry, worry about replacing them with something else, the dimmer isn't very important.

ETA: dumb_soprano claims anecdotal evidence that dimming *does* help not blow bulbs. OK, maybe we need to go beyond lj, go beyond google, and actually try something!
 
posted by [identity profile] davefish.livejournal.com at 12:23pm on 19/10/2005
That mostly covers what I was going to say.

Since you are worried about energy efficiency and using the spots, I'm fairly sure that higher wattage bulbs tend to be more efficient than the lower wattage equivalents.

gerald_duck: (duckling frontal)
posted by [personal profile] gerald_duck at 12:47pm on 19/10/2005
Compared with mains periodicity, the temperature of a light bulb's filament reacts quite slowly, and the heavily chopped mains waveform does approximate a lower voltage pretty accurately.

Agreed, for normal domestic use the improvement in bulb life from dimming it on rather than turning it on abruptly is way less than in stage or disco lighting, but I've heard guesstimates of around 20%. Whether or not that's a useful improvement is subjective.

Yes, the dimmer itself will be consuming a little power when dimmed, but it's also consuming a little power when the light is fully on. But I suspect resistance in the cable from switch to bulb loses more power than that once the house wiring's a decade old. (-8
 
posted by [identity profile] adqam.livejournal.com at 03:42pm on 19/10/2005
Excuse the pedant, but cold filaments have lower resistance, surely? Then, as the wire heats up the metal "ions" vibrate more, colliding with more electrons and increasing resistance. It is this low resistance when cold that leads to the surge in current you speak of:

I = V/R

This is, of course, entirely inaccurate if tungsten is a semiconductor :P
 
posted by [identity profile] keirf.livejournal.com at 01:24pm on 20/10/2005
Yes. I typed the wrong word.

October

SunMonTueWedThuFriSat
      1
 
2
 
3
 
4
 
5
 
6
 
7
 
8
 
9
 
10
 
11
 
12
 
13
 
14
 
15
 
16
 
17
 
18
 
19
 
20
 
21
 
22
 
23
 
24
 
25
26
 
27
 
28
 
29
 
30
 
31