I know that many problems with lamp premature deaths come from the tendency to make them as cheaply as possible - using low quality materials, reduced/lightweight constructions etc.
On the other hand, it's no secret that manufacturers don't want their products to survive too long. Phoebus cartel in 1924 started so called "planned obsolescence" with reducing the incandescent light bulb life to 1.000 hours.
But how did they do it with a tungsten filament? (Intentional impurities in material?)

And what about other light sources? LEDs compacts (lamp/s + driver + heat sink) can have their service life set quite easily - let them fry themselves by their own heat.
What about fluorescents and HIDs? (For example, Osram has HPS/SON lamps in standard /2Y(ears)?/, 4Y and 6Y versions.)

I know that many problems with lamp premature deaths come from the tendency to make them as cheaply as possible - using low quality materials, reduced/lightweight constructions etc.
On the other hand, it's no secret that manufacturers don't want their products to survive too long. Phoebus cartel in 1924 started so called "planned obsolescence" with reducing the incandescent light bulb life to 1.000 hours.
But how did they do it with a tungsten filament? (Intentional impurities in material?)

And what about other light sources? LEDs compacts (lamp/s + driver + heat sink) can have their service life set quite easily - let them fry themselves by their own heat.
What about fluorescents and HIDs? (For example, Osram has HPS/SON lamps in standard /2Y(ears)?/, 4Y and 6Y versions.)

in incandescent lamps, life is inverse proportional to the efficiency, so reducing lamp life cause an more efficien lamp.

The thing is, ordinary people can not measure the light output, but can well "measure" the life, so it becomes quite tempting to misuse that in marketing (to design the lamp with lower efficacy and then claim how "superior is that lamp compare to others, it lasts so much longer") and so making the lamp way more expensive to run for their customers.
The agreement on the 1000hour life was set as what was the compromise of the cheapest light with most makers (longer lasting bulbs were costlier on the electricity consumption, more efficient on bulb purchase) to prevent someone designing on purpose inefficient lamps in order to boost the life claims (as have happened many times in the meantime), it makes the lifetime fixed, so could not be used for such unfair marketing anymore.
Later this was lifted, the ersults were again many "superextra long life" bulbs, even gadgets like "the knob" appearing on the market, all promissing (and really giving) longer lamp life, but just "forgotting" to note the reduced efficacy.
What helped these unfair marketing was the practice to specify the light bulbs according to the input power and not the light output (and practically hide the light output information from the ordinary customers), that was (and still is) the real enabler for the swindlers...

By the way the cost optimum for an incandescent lamp was in their last decade (90's till 2000) not anymore the 1000hour life, but already something around 100..300hours...


With the LED based incandescent replacements the thing is more about finding the compromise of the lowest total cost: You may design twice as longer lasting bulb, but it would be ten times more expensive and you won't be able to use it in many places due to it's size. Not any optimum at all.
Of course the optimum is different for a screw in bulb and different for a fixed mounted lantern, but with the later you can way easier manage the heat...

Generally all it is about the money: Get as much light service with as little money as possible. The best way is the way leading to the lowest overall cost, but that does not mean having just the longest lasting lamp (when it's purchase is expensive like hell and consumes a lot of expensive electricity), nor just the most efficient one (when you have to buy many of them because they do not last and are expensive), nor the cheapest purchase price one (as that would cost way too much on the electricity and won't last long enough), all aspects have to be balanced. And on top of that it is very difficult to calculate the costcontributors, as each application is different. So what isactually happening: The market is looking for the best compromise, of course it leads to products exceeding in one way, but being bad in the others and so on. On top of that the technology evolves so, the compromise being best of yesterday isn't the best one today, because someone was able to find some way how to solve some problem by either lowering the cost of something, or improving something without an additional cost. The technology is by far notthat stable as it was in the early 20'th century with the incandescent technology...

It's hard to believe that a long life incandescent = just an underdriven incandescent (with no construction or material improvements).
How much do you need to underdrive a lamp in order to last 5 times longer? (There have been 5000 hrs. incandescents.)
And if you underdrive, aren't people going to notice the yellower and weaker light while seeing a regular and a long life lamp in the same fixture?
And... if nobody seems to notice - why not make all lamps like 5000 hrs.?
I think those 1000 hrs. have been the optimum for manufacturers, not for customers!

As for LEDs - agreed, the huge heat sinks would make them rather difficult to sell. But the shorter lamp life is also beneficial for manufacurers. They won't try to develop a lamp that will last 30 years and close their plants soon since only a few people will need to change their working lamps (provided there's no efficiency leap forward and electricity costs don't jump too high). Here is their self-frying ability quite handy.

And what about long-life HIDs (or fluorescents). Can the underdriving triple their life?

If you do not notice the difference, why not installing 40W instead of a 60W lamp in the first place (that is about the same difference)?
I guess you may calculate yourself, how much would have the lamp cost to make for the electricity cost difference. And I don't think the lamp makers are the only ones winning on that approach...

Of course to make a reliable 5000hour incandescent is not enough to just underdrive a 1000hour one (as there are other parts of the design tuned for that 1000hour life, to not increase the cost when the life is still limited by the filament evaporation), but without the lower temperature is just not possible to reach the 5000hour life. That is just the physics.
You may win a bit by choosing e.g. different gas fill, but at first these were not yet available at the time and even there, the optimum lifetime/efficacy does not shift that much.

And with HID, good example are CMH's: There the liquid salts slowly dissolve the arc tube. That means first the arc tube wall is getting thinner, plus the fill mix is getting contaminated by the dissolved alumina.
So for a lonmger life you would need thicker arctube wall. But a thicker arctube wall absorbs more light. So you may get the life longer, but it will again cost you the efficacy. So again, you need good balance, or the light would become too expensive.


Of course, there are many applications, where the cost of failing bulb is way higher (e.g. a traffic signal lamps), but those are out of the scope of general lighting service, but belong to a special applications. And there the median life (it makes sense when estimating just cost of new bulbs, not any extra expenses linked to e.g. their unexpected failure, as those are assumed as zero in general lighting) is very frequently irrelevant parameter at all. And of course, in these applications the extra power consumption linked to less efficient lamps is just a reasonable cost to pay for an acceptable failure rate (order of magnitude around 1% over the rated service life)...


So to answer the comment: Yes, manufacturers indeed did not made the lightbulbs so they would be the cheapest for the public, the lamps lasted unnecessarily long and so consumed too much power for the given light output...

They cheaply make fluorescents (low Hg, bad electrodes, etc...) so then people will buy LED replacements.

They cheaply make fluorescents (low Hg, bad electrodes, etc...) so then people will buy LED replacements.

The extra low Hg fluorescents are really the specialty of the US (and maybe Canada) markets, where when the lamp design ensures the mercury releases less than certain fixed level, it is not considered as a "dangerous waste", so no need for special, controlled (so expensive) disposal. That makes those ultra low Hg lamps so popular, even when they fail by far shorter than others. And as these lamps are intended as really a cheap lamps to operate, their manufacture and following logistic is made as cheap as possible as well, so hence the thin glass (lightweight for transport), questionable seal quality (too high speed processing), all that translates into their shorter life.

And for the other tubes the low quality is just a result of the lack of expertize on the manufacturing floor, mainly due to the desire to just lower the labor cost.

I've also seen green tipped alto HPS lamps too. And with the Mercury lamp life, I've seen a 1k and 1.5k watt merc lamp. Is the 1.5K really an overdriven version of the 1k or am I wrong? And if so, what effect does overdriving the lamp have on the life of it?

Even when the arctube is of the same size, shape and even fill (it may share the manufacturing equipment), the 1.5 vs 1kW mercs will likely differ in the electrode size (hard to notice; not that much necessary, if it is large enough) and mainly outer fill pressure or composition (not visible at all; to maintain the arctube temperature about the same for both products; and what matters is the temperature, nothing else).
They may not look so, but each of them is really different.

And for the HPS: I do not know the Alto series, but there are HPS designs without any mercury at all. They just use higher pressure Xenon gas as the buffer instead of the mercury. It has a bit more redder light, lower efficacy, shorter life and it is more expensive to buy, but it does not contain anything toxic at all. Well, some greenbrainers may consider that worth...

GE and Sylvania are now making lo Hg HPS bulbs.Have used these-they seem to produce a more yellow light than the older style bulbs.I don't know if its just me or has this been the same with everyone else thats has tried lo HG HPS bulbs here?

I found them:
http://916690855725293891.weebly.com/store/p44/1000W_HPS_Philips_Alto_Bulb.html

We have high xenon HO HPS lamps, which have low mercury and a yellower light than regular HPS lamps. They retains their original color until EOL.

Do they cycle and leak or just die at EOL.

They cycle at EOL. I've seen one that leaked.

The non cycle lamps here in the US glow till they use their sodium then they resort to burning the Hg and fill gasses.