Naked Science Forum
Non Life Sciences => Technology => Topic started by: graham.d on 30/11/2008 11:49:55
-
A neighbour asked me to help change the draw cord switch in their bathroom. They had a mechanical problem with the fitting as it happens. The reason they wanted to change it, however, was that they heard arcing inside the switch which correlated with intermittent operation of the energy saving light bulb it was switching on/off. Eventually this led to a catastrophic failure of the light (a quite expensive Philips bulb which comes on immediately and brightly from the start). The bulb is the usual fluorescent type (maybe 11W or 15W) but seems like it has a fast start system built in. The bulb had heat damage at one end of one of the tubes and this has cracked the glass.
After fitting the new switch and a new bulb of the same type, the light occasionally fails to come on. Pulling the draw cord does reveal the sound of some arcing in the switch as the light tries to activate. Switching off and on again usually works but the failure mechanism still occurs about 1 in 10 times.
I have a suspician that there is some mechanism in the bulb that is applying the heater to the cathode but then failing to generate the initial arc, so leaving the bulb heating but not striking. Now I intend to go and have a look to make sure the socket in which the bulb is fitted is OK, but I feel it maybe a fault in the bulb's electronics - either a design problem or a batch failure mechanism that has not been recognised. It is not good, because it costs money, maybe a fire risk and, by cracking the glass, potentially releasing noxious chemicals.
Any ideas?
-
Ahem! I'm embarrassed to say that I had inadvertently refitted the old cord switch back. My excuse was that I was doing it under torch light and the parts were being handed to me by my neighbourn but I should have checked. The new switch now fitted has fixed the problem.
It is still perhaps curious that the old switch had failed in an odd way and that the bulb got damaged so as to cause the glass to crack. Which caused which is debatable so I won't delete the post as it maybe still of some interest.
-
Arcing in the switch contacts may have produced an embarrassing RF component in the supply which confused the poor control circuit.
You brute! [:0]
-
You may be right SC, but all switches will arc a bit and they also bounce for several milliseconds. I wonder if it was not that there was just some series impedance from the switch that could cause this problem. The failure mechanism seemed to be that the switch could be turned on but the light fail to operate (the arcing only occurred when pulling the cord a bit to get it to spring into life). It is likely that my neighbour could have left it in this state for a long period. Could it be that the high series impedance (with the cathode heaters being applied) dropped the voltage sufficiently to stop the attempt to strike an arc in the tube but just kept the heaters on for a long period. This could cause the damage observed. It is also a failure mechanism that maybe Philips should be made aware of.
-
I should have thought that a large series resistance would only produce low input volts and that would have been dealt with adequately by the control circuit. A series inductance could produce strange effects, if the reactance were high enough to cause a high voltage due to resonance with some internal C. That was why I suggested the RFI from an arc could produce a problem.
-
It would not have to be a large impedance as I am guessing the heater must take a fair bit of current and so will drop the voltage. If it drops the voltage low enough it may be insufficient to trigger the circuit that strikes the arc in the tube. This would mean the tube sitting there indefinitely with its heater on. This would fit with observation of the damage but does depend on a design flaw. The reason I doubt the RFI from an arc in the switch is because you always get this and, additionally, a variable amount of switch bounce. The tube would have to cope with this normally. It could be that it can only cope with so much, but I don't know enough about the circuitry that is used in these tubes to assess this. However, I am not sure. Perhaps I should get the old tube (if not yet disposed of) and, out of interest, take it apart to reverse engineer the circuitry. Can I be bothered though? (Rhetorical).
-
My point is that the design (any half arsed designer) would take care of a simple high series resistance. Also, arcing at switch off (which is when it will happen) is normally short lived (and afterwards there's no power supplied) but sustained arcing - with power being available all the time - is an extra stress and the RFI might not be dealt with by the existing decoupling / filtering. It's obviously a cumulative effect or the thing would have blown up instantly.
-
Sustained arcing was not happening or I would agree that it would be a likely cause. Having been in electronic engineering for 38 years, I can tell you there are plenty of half arsed designers about :-)
-
But your first post stated that they heard arcing. ??
-
The arcing was only during switching, and as I found out by experiment, it occurred when pulling slightly on the cord when the lamp had failed to light though the faulty switch was in the on-position, and this could then cause the lamp to strike. The fault was in the switch in that the contacts were clearly poorly operating and replacing the switch worked. This was with a new lamp of identical type. The failed lamp looked overheated close to the cathode.
Anyway, the whole problem could be a cascade of failure mechanisms. I only posted to see if the problem had been seen elsewhere to know whether it could be any fundamental design problem that was worth reporting to Philips.
-
It could be any one of several cascaded failure mechanisms. That said, there's many well-known failure modes of CFL lightbulbs which involving charring of the lamp base, noxious smoke, and cracked glass.
Also, although not explicitly stated on the packet, CFL lamps are not designed to be installed in enclosed fittings as the heat build-up (though much less than from a conventional bulb) can damage or shorten the life of the electronics. It would also not be unreasonable to suggest that the likely humidity (and possible condensation) in a bathroom is not good for electronics. Finally, CFLs are known not to strike in cold conditions (eg below 10C) and people whose bathrooms are less heated may expect problems.
Even setting aside issues of ghastly colour-rendering, CFLs are not the universal drop-in replacements for proper bulbs that their proponents proclaim them to be.
http://sound.westhost.com/articles/incandescent.htm (long (and I have some minor tech quibbles with some of the content), but see the mid-point of the page)
http://www.hpa.org.uk/webw/HPAweb&HPAwebStandard/HPAweb_C/1223534061375
http://www.myhamilton.ca/NR/rdonlyres/87F25B8E-4ACA-4EC4-B55A-E97A9D28D1E3/28575/Safety20Advisory20CFL20Issues.pdf
Apparently above link does not work (I retrieved it from old records).
Try this one: www.bluewaterpower.com/pdf/Safety%2520Advisory%2520CFL%2520Issues.pdf
-
They are interesting websites and much food for thought (the last site came up blank btw). Out of interest I will see if I can retrieve the dead CFL from my neighbour, if it hasn't been chucked, and take it apart. Mains voltages on a PCB an subjection to steam is probably not a good idea. I rather assumed there would be a degree of sealing of the electronics but I guess not! The enclosure it was in was fairly open, but downward facing so that the temperature would tend to be higher at the socket (and electronic) end.
-
Also, although not explicitly stated on the packet, CFL lamps are not designed to be installed in enclosed fittings
It stated "not for use in an enclosed fitting" on the ones I bought (made by GEC) Techmind.
I have a battery powered lantern with a five watt fluorescent u-tube. When the batteries start getting weaker, it takes to coming on only 50% of the time, in other words, switching off, then on again will do it, at least until the batteries get even weaker, in which case replacing them is the only option, of course.
-
The one I like which they put on CFL packages is where they say it will last for eight years- then you read the small print, which says "based on an average use of two hours a day." I don't know about you, but the lights in the more commonly used parts of my house are on for a lot more than two hours each night.
What about LEDs? I've heard the main barrier to using them as house lighting is that they cost too much to produce. Would they be better than CFLs? You seem to be fairly opposed to CFLs, Techmind.
-
The reason why switches arc is mainly physical, that being friction between the surfaces making contact as they pass from coroded to better conductor contact.
When this occurs the principals of inductance occur in the breakage and change of current values, meaning that a (falsely) expressed voltage-amperage rise and drop occurs(alike oscillation frequency) on the surfacing of the two conductors "because of the uneven scraping contact" resulting in a "field breakage(from poor surface to conductive efficient surface) that operates alike an amplifier inductance"(massive voltage and tiny amperage).
Moreover, the initial effect of that change over the scrape distance e.g. 0.3mm only result in a frequency of
0.3/time BUT
the side-band harmonics (of the e.g. 0.3mm/time frequency) produced BECAUSE a base frequency is produced(by frequency change) results in various harmful unfilterable frequencies and voltages of those massive voltages(alike an amplifier) allowed to pass inside the light bulb for operational levels.
SEE: "SSB transmitter Harmonics frequency production for this method".
AND amplification coil inductance values.
-
The one I like which they put on CFL packages is where they say it will last for eight years- then you read the small print, which says "based on an average use of two hours a day." I don't know about you, but the lights in the more commonly used parts of my house are on for a lot more than two hours each night.
my packaging read,"bulb life is 10,000 hrs. subtract 2hrs. for each time the bulb is tuned off/on.
-
I would like to caution readers against using devices like fluorescent bulbs via a dimmer switch, I a had a small power measuring device with a transformer less PSU which for curiosity I ran via a dimmer.
To my surprise it went up in smoke presumably due to the dimmer putting out a high frequency that reduced the impedance of the input capacitor on the PSU.
-
If you want to make a dimmer system that does not arc , you need to use an amplifier like resistor system.
Just a pair of coils on a ferrous rod that is mounted in a screw thread shaft that exposes more of one side to one coil when wound and loses exposure to the other coil.
Go browse an electronics shoppe.