Bulbs & fuse blowing

[whisky-dave]

I've wondered about brands as you can get so called value packs. Anyone know if tehse store value packs are actually a good deal or just cheaply made bulbs or designed to have a shorter life.

[Appin]

The message

from [e-mail] contains these words:

They can be VERY sensitive to vibration -- don't use them on the ground floor of a two-storey house.

I'm not sure I can figure out your arithmetic.

If I have ten bulbs with a mean time before failure of 1500 hours, I can, assuming the batch is typical, expect about 1,500x10 hours of light from them which is 15,000 hours of light.

You seem to be implying that if I put in a batch of ten new bulbs, by the time I reach 1500 hours they'll all have blown

Bulb 1 blows after 150 hours Lifespan 150 hours Bulb 2 blows after 300 hours Lifespan 300 hours Bulb 3 blows after 450 hours Lifespan 450 hours Bulb 4 blows after 600 hours Lifespan 600 hours Bulb 5 blows after 750 hours Lifespan 750 hours Bulb 6 blows after 900 hours Lifespan 900 hours Bulb 7 blows after 1050 hours Lifespan 1050 hours Bulb 8 blows after 1200 hours Lifespan 1200 hours Bulb 9 blows after 1350 hours Lifespan 1350 hours Bulb 10 blows after 1500 hours Lifespan 1500 hours

Total number of hours of illumination from the whole batch 8250 hours

Giving an average lifespan of 825 hours

For what you have indicated to be true the bulbs would need to have been replaced at genuinely random intervals so that they wouldn't be starting their life together AND their failure pattern would have to have a fairly random distribution. In point of fact though we can all think of and quote exceptions, many mass-produced lamps from a particular batch do tend to fail relatively close to each other in terms of burning time

-- if the design life is 1000 hours, a majority will fail in (say) the

800 - 2000 hour range, with relatively few failing before 500 hours and relatively few lasting more than 2500 hours.

Usage patterns distort the graph considerably.

[dennis]

The pack I bought in aldi last year are lasting well.. they are still in the packet as none of the six gu10s have failed yet.

[ARWadworth]

They last longer than the low energy equivalents. I fitted 300 low energy ones last week and had a 1 in 10 failure rate in less than 24 hours. I am up to 1 in 7 failure rate a week later.

Adam

[John Rumm]

LED or CFL?

and what brand should we avoid? ;-)

[John Rumm]

I think NT was assuming you are going to install one, then wait 150 hours, and then install the next and so on. Each could then last 1500 hours, but you would need to change one every 150 after the first 1500.

He has a bit of a "thing" about halogen lighting in case you had not guessed ;-)

[ARWadworth]

CFL and all brands:)

In my case Halolite 11W versions, but I did not supply them I only fitted them.

A complete waste of space IMHO.

And they are expected to run 24/7.

Adam

[Appin]

The message from John Rumm contains these words:

You'd need to do some very creative accounting to make the maths work out -- and we all know now what's happened with all the creative accounting of an awful lot of bankers :-)

I'm not the world's greatest enthusiast for it myself, but I do like my arithmetic to work out :-)

[John Rumm]

LOL! Indeed. Although the be fair, I did actually find a CFL that was "ok" the other day. It was a Megaman Ultra Compact Candle (which is to say, its the same shape and size as normal non compact candle filament bulb). It does actually do a passing imitation of a 40W pearl candle bulb, in both colour rendition and light output.

(spurred on by that success I also tried on of their small lightbulb shaped CFLs that was supposed to be a 60W equivalent. Needless to say the spectra is wrong, the warmup is dog slow, and the light output is the equal of at best 40W).

[Frank Erskine]

I had a problem with my bedside lamp, which uses three SES candle bulbs of ca. 25W. It has narrow fluted glass shades, which precluded most CFLs. The colour wasn't that critical, since the shades are a sort of brown/green combination, and it's only used for getting into bed and a few minutes of reading. At a local indoor 'market' I saw some tiny CFLs of, ISTR, 5 or 7 watts, which were a bit smaller than the original tungstens, and they seem to be not too bad. The warmup time is about 10 or 20 secs, the shades run reasonably (fsvo) cool and I haven't had a lamp failure in maybe a year. I'm still not a fan of CFLs generally, but I'm reasonably satisfied with these (which were quite cheap anyway!!!).

[meow2222]

what an odd thing to say. Is 10 1500hr lamps giving on average 10 failures per 1500hrs confusing?

NT

[Appin]

The message

from [e-mail] contains these words:

Yes, it's confusing.

If the lamps are purportedly rated at 1500 hours, then I should in theory expect the average life of the set of ten to be 1500 hours.

If I use them in the fashion anticipated in the testing for rating purposes and have all of them buring starting at the same time, then I would not expect ten failutres at the 1500 hour mark. I should be able to expect

[meow2222]

that much is right, though previous posts were another matter.

NT

[Appin]

The message from [e-mail] contains these words:

If you allege previous posts to be "another matter" you should justify the allegation. You haven't explained how you reached your conclusions.

[meow2222]

te:

I'm not going to sit here & teach basic stats.

NT

[meow2222]

rote:

I see there's alt.math, sci.math, alt.sci.math, net.math and uk.education.maths if you're genuinely interested.

NT

[Rick Hughes]

Thought that leccy was never distributed at 250v any more ,,,, due to EU harmonization of supply voltage.

[Appin]

The message from [e-mail] contains these words:

You don't need to -- I've done them at university level. And I can cope with spherical trig as well, for what it's worth.

What we're involved with here is not basic statistics, but *applied* statistics, so if you propound a theory with regard to a failure every

150 hours it's reasonable to ask that you at least explain the other parameters you're basing this theory on.

One thing sure, if you start your ten lamps burining at the same time and under the same conditions, then if they're failing every 150 hours, the average lifespan, as I've already suggested, would appear to be 825 hours, rather than 1500 hours. That, of course, assumes that as they aren't being replaced as they fail.

You've given no indication of the spread which you've assumed the failures to show around the design life -- and with lamps this is a significant consideration to be factored in.

I passed my university-level statistics (for what it's worth) in the

1960s. I live, however in the real world and am looking for a real-world explanation of what you've written.

[Appin]

The message from "Rick Hughes" contains these words:

It's often distributed at 250v plus. It's not always within even the specified limits and if you're near the substation you may well be over

250v. Time to get the electricity supply company to bring along a recording voltmeter.

[meow2222]

ok, fairy nuff. I'm trying to sort out where the misunderstanding lies though...

If we operate one lamp at a time (and replace it many times when it dies to give us a fair sample size) we get one failure per 1500hrs on average. If instead we light 10 rather than just 1 at once. we're going to see 10 failures per 1500hrs average, which is one per 150hrs of operation on average. Its so simple that I'm struggling to see where you think that's incorrect.

NT