Can't Change a Motherfucking Lightbulb

[Lisa Jesusandmarychain]

Job applicant rethinks reply to the question " What would you say are your weaknesses?"

[sick as a pike]

Certainly don't do what I did before I realised the correct process (what Buelligan described up there- using a lot more press/twist force than you expect). 
Do not get frustrated and pull the entire light fitting down and out, not realising that it is held in place by two incredibly powerful spring clips which snap onto your fingers like a rat trap, making you think you've been electrocuted, scream and fall off the ladder. Do not do this.

[Bobtoo]

I've never seen a GU10 fitting that wasn't fucked, and I don't think I've ever seen a room with GU10 lights that were all working.

On a related note, possibly one for Busby.

How come Halogen headlight bulbs in a car can withstand all the bumps, voltage surges and heat cycles for years or even decades, yet Halogen bulbs in the house have a lifespan ranging from a few minutes to a few weeks?

[buzby]

How come Halogen headlight bulbs in a car can withstand all the bumps, voltage surges and heat cycles for years or even decades, yet Halogen bulbs in the house have a lifespan ranging from a few minutes to a few weeks?

Halogen bulbs used in cars have a very well regulated supply voltage (normally about 14.1V with the engine running) courtesy of the regulator pack on the alternator. They also are typicaly low wattage (5W for most of the constantly-on bulbs like side and tail lights) and have plenty of free air cooling around the bulb itself. They also aren't usually on for all that long - only a small percentage of journeys are at night and even then will only be on for a couple of hours.

Since the mid-80s cars have also been fitted with bulb failure monitoring systems that apply a low current to the bulb even when it's off to measure it's resistance (these systems are what cause problems when people try to replace bulbs with LEDs - the so-called 'CAN BUS' compatibility problem that has nothing to do with the car's CANBus network at all). This low-current voltage barely causes the filament to glow, but also has the benefit of stabilising the bulb so when it's turned on it isn't exposed to as much thermal shock or current inrush.

Halogen bulbs used in houses are either fed direct from the mains or a cheap, poorly regulated switch-mode PSU for low voltage types, so are exposed to surges much more than automotive bulbs. They are usually higher power units - they start at 10W, but 20 or 50W is typical (the latter is the same as wattage as the headlamp bulb in a car) and 75W is also pretty common. The biggest problem however, particularly for downlighters, is heat. A halogen bulb runs at about 200 degrees C, and in downlighters there is no free air cooling them and they have to have firestop hoods fitted above them, which basically eliminates any convection cooling. They are on constantly for hours at a time every day, and eventually that heat stress will cause the filament to fail prematurely.

For low power types, a poorly regulated power supply can also mean that once one bulb fails, the others in the chain will be getting slightly more current which can lead to a 'cascade' failure of the other bulbs.

Replacing mains-powered halogen bulbs with LEDs won't always mean they will last any longer either - the mains GU10 LED bulbs each have their own small buck regulator PSU in the base of the bulb. LEDs running at current levels required for ambient lighting have similar problems with heat generation, and the heat rising up from the LED's heatsink will prematurely kill the electronic components that are sitting directly above it.

TL:DR - downlighters are a really shit idea, and are very much a triumph of style over function from an engineering or fire safety standpoint.

[touchingcloth]

For low power types, a poorly regulated power supply can also mean that once one bulb fails, the others in the chain will be getting slightly more current which can lead to a 'cascade' failure of the other bulbs.

Does the same thing apply to filament and CFL bulbs? It's a common observation that bulbs are like buses in that they work for ages and then all start failing at once, which I've always chalked up to bulbs replaced at the same time and being on for the same number of hours are likely to burn out at around the same time, which must be true to an extent, but could it be that with bulb types other than halogen that the cascade is triggered in a similar way, and that the small surge combined with older bulbs leads to a greater likelihood of a failure?

We've honoured mook's memory by talking about bulbs, so let's move onto rings, or hoops if you will.

I understand now that heat is the death knell of a halogen bulb and more of a problem in a downlighter than a headlight, but what about halogen stove rings? They're designed to get hot as fuck, yet I've never seen a broken one.

[buzby]

Does the same thing apply to filament and CFL bulbs? It's a common observation that bulbs are like buses in that they work for ages and then all start failing at once, which I've always chalked up to bulbs replaced at the same time and being on for the same number of hours are likely to burn out at around the same time, which must be true to an extent, but could it be that with bulb types other than halogen that the cascade is triggered in a similar way, and that the small surge combined with older bulbs leads to a greater likelihood of a failure?

It can do, but usually only on low voltage systems with bulbs sharing a common supply (such as cars), where you are relying on the voltage regulation to cope with the reduction in load once one bulb fails. On most mains systems the loss of load from one bulb going is not going to have any noticable effect on the output of your local substation. Failures of multiple mains-powered bulbs in your house are usually down to the MBTF of the bulb's design/reaching the far end of the bathtub curve.

CFLs have similar problems to LEDs regarding failure of the cxomponents of their built-in electronic balllasts. In their case the failulres are because they have to boost the mains voltage up to around 1000V to strike the arc in the tube, which in the space available in the back of the bulb means the components are quite highly stressed (and also generate heat, with not much in the way of cooling available).

I understand now that heat is the death knell of a halogen bulb and more of a problem in a downlighter than a headlight, but what about halogen stove rings? They're designed to get hot as fuck, yet I've never seen a broken one.

The tubes in a halogen stove ring (which were patented by Thorn EMI's lighting division in the mid-80s, and introduced on Thorn's Tricity Sovereign cooker) are designed to radiate most of their power as infrared light, which passes out of the quartz glass envelope of the tube with very little loss (about 75-80% of the input power to the tube is emitted as infrared light). The actual tube itself doesn't get any hotter than a normal halogen bulb. They also usually have a big metal reflective backing plate that acts as a heatsink.

Pure halogen hobs are a pretty rare thing nowadays. Induction hobs stole their market, really. The technology does live on in some 'ceramic' hobs though, which have a hybrid of a halogen tube around the edge with a conventional resistance element in the centre.

The halogen tube is used to provide the initial rapid heating effect before the resistance element comes up to temperature.