This shouldn't be happening.

June 26, 2008 1:20 PM

Help me diagnose repeated failure of my lighting dimmers.

When I remodeled my kitchen about a year and a half ago, I installed nine line-voltage 4" recessed lights which are controlled by two Leviton Illlumatech preset dimmers, with three cans/150W on one dimmer and six cans/300W on the other (these are 600W dimmers). Everything works fine, except that when a bulb burns out there's a better-than-even chance it will take the dimmer with it. A dimmer has failed three of the four times that a bulb has burned out since the initial installation. Only the slide dimmer fails -- the on/off button continues to work normally.

Leviton has been replacing the dimmers as I mail them in, but their tech support people are, so far, unable to imagine how a bulb's burning out could cause the dimmer to fail. They've speculated that it could be a problem with a fixture or a loose connection, neither of which ideas holds water; the failures have all been simultaneous with burn-outs in different fixtures, and there are no symptoms of loose connections anywhere.

I'm hoping someone here has a brilliant insight to offer. The bulbs, if it matters, are Sylvania 50W PAR20s.

posted by jon1270 to technology (20 comments total)

how do they fail? have you cracked one open and had a look-see?
posted by stubby phillips at 1:27 PM on June 26, 2008

have you cracked one open and had a look-see?

No I have not. If I were to do so, what should I look for? I have a multimeter and understand basic electrical ideas but have zero expertise with solid-state electronics.
posted by jon1270 at 1:31 PM on June 26, 2008

" unable to imagine how a bulb's burning out could cause the dimmer to fail."

That's because a bulb burning out shouldn't cause the dimmer to fail. But what's causing the bulb to burn out?

It's reasonable that the dimmer would fail at the same time as another device on the circuit (the bulb). It's not reasonable that the failure of the bulb itself is the cause of the problem. It's more likely that poor line conditions are causing both problems.

"When I remodeled my kitchen about a year and a half ago... A dimmer has failed three of the four times that a bulb has burned out since the initial installation."

You've had four bulb failures and three dimmer failures on this circuit in a year and a half? There's definitely something wrong. Most bulbs on a circuit with decently clean power will last a couple of years, not months.
posted by majick at 1:31 PM on June 26, 2008

sorry, i didn't google the dimmer until after i posted. i was assuming the big rheostat kind of thing.
posted by stubby phillips at 1:39 PM on June 26, 2008

majick, the four burn-outs were not consecutive failures in the same fixture; they were all failures of the original bulbs in different fixtures. The first failed when all nine were about a year old, the next when the bulbs were perhaps 14 months old, etc. The average rated life of these bulbs is 2500 hours and these fixtures, being in the kitchen and on the north side of the house, are often on for several hours a day.

Might be worth mentioning that the two dimmers are on different breakers, and supply different brands/models of fixtures.
posted by jon1270 at 1:43 PM on June 26, 2008

This is a problem of underdesigned or poorly designed lamp dimmers. The cheap lamp dimmers I've seen never have fuses, though.

Triacs and thyristors are sensitive to overcurrents. When dimming normal light bulbs, short circuits caused when filament burns are quite probable. For this reason, light dimmers must have their own fuse which protect it against failures in this kind of situation.

Thyristors have a defined overcurrent handling capacity and the fuse must be selected so that it burns before the thyristor in overcurrent situation. This typically means that the thyristor/triac must have a current rating of 2..5 times bigger that the rating of the fuse in order to be sure that the fuse burns before thyristor/triac in case of short circuit. The fuse type must be also fast enough to burn in this case before the thyristor/triac. In some cases it might be necessary to use special fuses to be able to protect the components effectively.

The thyristor must have a high enough surge current rating also for normal operation. For example in case of normal light bulb dimming of a light bulb with cold filament is turned on at 90 degrees after zero crossing (means at maximum line voltage peak), the peak current can be 20 times bigger than the nominal current of the lamp. http://www.epanorama.net/documents/lights/lightdimmer.html#componentselection

What usually happens is that when one of the lamps blows, the element in the lamp momentarily creates a plasma arc before it completely breaks, and will often blow the 3 Amp fuse. However, fuses are usually too slow reacting to stop the THYRISTOR from being damaged by the overcurrent. I have never known the little IC (chip) get damaged, or the diode, it is nearly always the THYRISTOR/TRIAC.

If you are confident with soldering and general electrical work, this CAN be fixed, provided the electronic unit that contains this part is not sealed with a potting compound. They are usually just a 2 part plastic box, that unclips to reveal a small circuit board, with a small chip, and the THYRISTOR on it. The first problem is getting into the base, to take the box apart. These are sometimes rivetted on, so you will have to drill out the rivets, and glue the base back on later ( a glue-gun is good for this).http://www.letsfixit.co.uk/html/lamps_faq.html

posted by jepler at 2:02 PM on June 26, 2008

(the slide dimmer is the part controlled by a thing called a "triac". In the off position the lamp is not part of an electrical circuit, and in the full-on position the triac is bypassed just as if the lamp was not on a dimmer circuit at all)
posted by jepler at 2:03 PM on June 26, 2008

Long wiring run combined with grounding problems at the dimmer fixtures?
posted by the Real Dan at 2:13 PM on June 26, 2008

It's going to take me a little while to digest jepler's post, but I wanted to respond to Dan. No, the wiring runs aren't very long (tiny house, 12G wiring and 15A breakers). Can you say more about "grounding problems at the dimmer fixtures?" Do you mean problems at the dimmer itself or at the luminaire (lighting fixture)? How would this explain the failures?
posted by jon1270 at 2:24 PM on June 26, 2008

It's going to take me a little while to digest jepler's post

In summary, a light blowing can create a very brief short circuit, leading to too much current flowing through the dimmer switch. This can damage the dimmer switch. You could try to prevent this problem with a fuse or miniature circuit breaker - if we have identified the problem correctly.

When you say "Only the slide dimmer fails -- the on/off button continues to work normally" do you mean the light comes on at full brightness all the time?
posted by Mike1024 at 3:03 PM on June 26, 2008

Seconding jepler.

Translation: This is a known phenomenon. When a lamp burns out, that doesn't mean it always goes directly from its normal state (conducting the proper current, giving off the normal amount of light) to an open circuit (no current, no light). In the fraction of a second while it is burning out, it can temporarily go into a state that is something like a dead short (excessive current, flash of light, surge of current through the dimmer). The science behind this is irrelevant to your needs — the bottom line for you is, when a lamp burns out in this way, it takes your dimmer with it. This is consistent with the dimmer no longer working, but the on/off switch working normally.

Have you tried a different brand of light bulbs? Different brand of dimmer?
posted by exphysicist345 at 3:14 PM on June 26, 2008

Jepler covered it well. I have repaired dimmers by replacing the triacs with a higher current rating device. Triacs only cost a dollar or two and it is too bad Levitron doesn't just make their switches with a higher rating. Replacing the triac is not too difficult if you are handy with a soldering iron. You can get replacements at Digikey. Get the triac part number off your current switch then look for a similar one in the family with a higher rating.

Another thing you might try is using a different brand of bulb in your fixtures. The bulbs are supposed to have a fusible link in the leads that opens the circuit if a plasma arc forms when the filament blows, but don't always work fast enough to protect your triac. A different brand bulb might have a more sensitive fuse.
posted by JackFlash at 3:32 PM on June 26, 2008

If would try a different brand of dimmer the fact Leviton keeps replaces them no questions ask makes me think they know they have a problem
You can also put your hand on them to see if there running warm/hot
Or it could be your installing them incorrectly
posted by SatansCabanaboy at 4:19 PM on June 26, 2008

Halogen bulbs are really not designed for use with dimmers:

Line voltage (120V) halogen bulbs can be dimmed by regular incandescent dimmers. Using a dimmer with halogen bulbs actually has negative effects. When dimmed, the halogen filaments do not reach the 250 C needed for the halogen cycle to take place. This could cause the inside wall of the bulb to blacken reducing light quantity and life. Running the lamp at full brightness will help restore and clean the bulb.


Although this paragraph is wrong about the temperature of the filaments-- they are over 1000C if they look yellow at all (perhaps the walls of the inner quartz envelope do need to be >250C)-- the overall thrust is accurate.

I have a feeling the (conductive) black deposit on the wall could strongly worsen the arcing effect mentioned by your other answerers, too.

Try replacing one bank with ordinary incandescents.
posted by jamjam at 4:46 PM on June 26, 2008

When you say "Only the slide dimmer fails -- the on/off button continues to work normally" do you mean the light comes on at full brightness all the time?

Exactly. A preset dimmer separates the on/off control (in this case, a button) from the brightness control (the slide). The button continues to work like a normal on/off switch , but the slide no longer dims the lights; they remain on at full brightness.


Have you tried a different brand of light bulbs? Different brand of dimmer?

No, I have not. The thought had occurred to me, but I've been feeling cheap. Maybe time to consider a change in that strategy...


You can also put your hand on them to see if there running warm/hot

They don't run hot. They can feel slightly warm, but that's normal for any dimmer.


Or it could be your installing them incorrectly

Not likely. There are only three connections to make, and the dimmer won't work at all if any of them are wrong.


Halogen bulbs are really not designed for use with dimmers...

I'm familiar with this line of reasoning. It's true that the life of halogen bulbs is shortened if they are always dimmed, but the typical rule of thumb I've read is that running them at full power for as little as 10% of the time is sufficient to prevent problems. My usage habits easily clear that hurdle.


Jepler's line of thinking makes sense to me generally, but the implication that this is somehow a normal/common problem seems like a bit of a stretch. I mean, why would Leviton produce a product that, for many customers, demands professional installation, and that fails so frequently? It seems like doing so would be brand suicide.
posted by jon1270 at 5:31 PM on June 26, 2008

By the way, what is the model number of the lamps you're using?
posted by exphysicist345 at 7:00 PM on June 26, 2008

Jepler & everybody else is right; it's a result of the usual failure mode of halogens to be a brief s/c before going open, and the triac dying in that instant.

It's worth noting 2 things though:

1. higher-rated triacs often have a reduced tolerance to s/c loads - that is, the larger the continuous rating, the shorter time they can stand a short. This is a consequence of the semiconductor die size & construction - often it's very similar size & construction to lower rated devices, just with a bigger heatsink/mounting tab.
2. A dimmer, fully on, is still only providing ~90% of the power that wiring via a normal on/off switch would provide, and it's noisy as hell. The other part of the reason dimmed lights fail, apart from not running at the optimal temperature, is that even at "full" on the light is suddenly being switched on (or off) after the start (or before the end) of the normal AC power's sine wave - rather than gradually ramping up from (or down to) 0 volts, it gets a sudden transition. This causes thermal & mechanical shock to the filament/element, reducing its life.

Short story - lights on a dimmer, even if the dimmer is always run full-brightness, will always have shorter lives. This goes doubly for lower wattage bulbs of any type, as their filaments/elements are thinner (thin = higher resistance = lower wattage, but also = less tolerance of thermal/mechanical stress).
posted by Pinback at 10:00 PM on June 26, 2008

By the way, what is the model number of the lamps you're using?

50PAR20/CAP/SPL/WFL40
posted by jon1270 at 3:35 AM on June 27, 2008

I just opened up a blown dimmer. It was easy enough to get into, but the part I imagine is the triac is riveted to the aluminum plate on the front of the dimmer (which acts as a heat sink) so that's a little more work to remove. Nothing looks visibly damaged, and there is no fuse.

Since I can get these dimmers replaced for the cost of mailing them in, I don't see the sense in trying to find the correct part and repair them myself. I would, however, be interested in protecting them with a fuse or miniature breaker if that were feasible.

Is there a specific type of breaker or fuse that I should look into, that would trip/blow quickly enough to protect the triac?
posted by jon1270 at 6:15 AM on June 27, 2008

I don't think you are going to be able to protect these with a fuse. The fuse would have to be sized exactly right and have the right delay to handle the large inrush current that occurs when you turn on a light normally but still be able to quench an arc when a bulb burns out before the triac blows. This could vary depending on the dimmer setting at the time. You are better off depending on the fuse built into the bulb if you can find a brand that works properly with your dimmer.

Another way of looking at it is that the triac is your fuse. You just replace the dimmer when it blows the same as if it had a fuse. Just keep a spare or two around and keep swapping them for free replacements from Leviton. You just have to accept the fact that Leviton doesn't have a very robust design. It probably works fine for the majority of their customers. It just doesn't work for your particular configuration of lights.
posted by JackFlash at 12:21 PM on June 27, 2008