LED lamps on dimmer circuits

On Fri, 18 Jul 2008 19:24:15 -0400 John Gilmer wrote: | I tried to post before but my news server seems to have screw up. | | Anyway, I finally am seeing some LED 120 volt lamps (at Wally World.) | | They come with the warming to not use them on dimmer circuit but the best | application for now in our place is on an X-10 controlled light fixture. I | haven't tried dimming but with one "normal" lamp and one LED lamp on the | circuit is seems to work OK. | | SO: what's the problem with dimmers and LED lamps? What's the potential | for harm if you try to dim it?

The way dimmers work is by chopping the wave cycles of AC into narrower pulses. This is OK on incandescent, although it can make the filament "sing" a bit. It can do nasty bad things to electronic circuitry that is designed for clean AC. Those pulses have a lot of harmonics that can have the effect of overloading the circuitry in LED and CFL lamps that are not specifically designed to handle it (takes more circuitry and/or overrated components).

| Since I have your attention, what about CFLs? What potentially could "go | wrong" if you "dim" a CFL bulb?

Depends on the bulb.

The potential risk is it can overheat, rupture, catch the house on fire, and kill members of your family while burning down the house.

I don't know what they are doing with the 120 v LED replacements. Are they just a full wave bridge with a dropping resistor, do they use any filter caps? If no filtering they would flicker like hell on a duty cycle dimmer..

Eric

On Sat, 19 Jul 2008 10:53:43 -0400 Eric wrote: | snipped-for-privacy@ipal.net wrote: |> On Fri, 18 Jul 2008 19:24:15 -0400 John Gilmer wrote: |> | I tried to post before but my news server seems to have screw up. |> | |> | Anyway, I finally am seeing some LED 120 volt lamps (at Wally World.) |> | |> | They come with the warming to not use them on dimmer circuit but the best |> | application for now in our place is on an X-10 controlled light fixture. I |> | haven't tried dimming but with one "normal" lamp and one LED lamp on the |> | circuit is seems to work OK. |> | |> | SO: what's the problem with dimmers and LED lamps? What's the potential |> | for harm if you try to dim it? |> |> The way dimmers work is by chopping the wave cycles of AC into narrower |> pulses. This is OK on incandescent, although it can make the filament |> "sing" a bit. It can do nasty bad things to electronic circuitry that |> is designed for clean AC. Those pulses have a lot of harmonics that can |> have the effect of overloading the circuitry in LED and CFL lamps that |> are not specifically designed to handle it (takes more circuitry and/or |> overrated components). |> |> |> | Since I have your attention, what about CFLs? What potentially could "go |> | wrong" if you "dim" a CFL bulb? |> |> Depends on the bulb. |> |> The potential risk is it can overheat, rupture, catch the house on fire, and |> kill members of your family while burning down the house. |> | I don't know what they are doing with the 120 v LED replacements. Are | they just a full wave bridge with a dropping resistor, do they use any | filter caps? If no filtering they would flicker like hell on a duty | cycle dimmer..

One possibility is that instead of a ballast resistor (which wastes power as heat, negating advantages of using LED for efficiency), they might be using a solid state current chopping limiter. In effect this acts much like a dimmer, but is set at a level to prevent average overcurrent. If they operate at line frequency, like most dimmers do, the best case is that the ballast compensates for the dimmer, preventing the dimming from being effective. And this may overheat the ballast in the process in some way. Or maybe it would cause the ballast to not do its job and the LED chip would be burned out.

If an LED bulb assembly _is_ rated for dimmer use _and_ has a legitimate UL listing, then I'm sure UL has put that design through the tests in their fire-proof testing labs (what a cool job that might be to blow things up).

I never saw one of those LED bulbs with an Edison socket and I know nothing about how this stuff is made but .........................

If I could make LED lights, I would explore putting a bunch of them in series till the forward drop was the line voltage and let them rectify the line AC. This way there would be no extra parts. Or make a bridge out of two sets of LED strings one for positive and one for negative.

Advantage : Reduce the duty cycle of each set. This assumes that a bunch of LED's on a wafer are incrementally inexpensive.... they would also give or more light than a single LED Eliminate need for a dropping resistor that wastes power Would work on a dimmer if any part of the wave went higher than the forward voltage drop

Disadvantage : reliability- one fails they all go out,

Observation: The Led's in my rope light on my deck work well with a dimmer.

peace dawg

Problem is that LEDs need operating at constant current - not constant voltage. A big dropper resistor effectively turns a constant voltage supply into a constant current supply. Without it, LED current and brightness would vary greatly with ambient and supply conditions and probably destructively. So, some form of current limiting device would be needed.

-- Sue

Also if no filtering a full wave bridge is desirable as the flicker would be 120 hz instead of 60. If you used LEDs in a bridge half would be lit on one cycle, the other half on the other, at a 60 hz rate. If they were not so expensive I'd tear into one to find out what they do.. Eric

So far, I have NOT taken apart a LED lamp. I have 3 altogether. One has failed and one is "weak" (yeah, I know all about the 50k hour life). So one of these days I will open things up.

I am well underway to answering my own question.

I "accidentaly" destroyed a 120 volt LED lamp.

It has 20+ LEDs in series. It looks like from the 120 supply, there is a resistor "balast" feeding a full way bridge rectifier. There are also 3 capacitors: one is an electrolytic. The other two might be there to reduce "electronic" noise.

In times like this I wish I still had access to an EE lab with the expensive scopes, etc.

I haven't traced the wiring in detail but it's clear there just isn't any fancy electronics and that all the LEDs are in series.

It seems to me that a company could get lots of efficiency brownie points by replacing the resistor with a capacitor. It won't dissipate any power and the unit as a whole would have a leading power factor as a bonus. 20 LEDs at 3V drop each or so is 60 volts drop total, meaning about half the power is wasted in the resistor.

Of course a capacitor good for line voltage at however many mA is going to cost more than the resistor.

The capacitor is ok on a perfect sine wave, but doesn't attenuate harmonics, or narrow spikes.

A resistor (or something similar) would still be needed to limit the peak LED current on switch-on, with the capacitor fully discharged but peak mains voltage applied to the series circuit.

-- Sue

So they need to drop hmm let's see.. Probably around 150 vcd across the filter cap, about 3v each for LEDs ... about 90 volts. Seems like they need to use way more than 20 leds to be reasonably efficient... I suspected as much.. Eric

Why not an inductor instead of a capacitor? Passive ballasts for fluorescent tubes always seem to be inductive, not capacitive, and I assume there are cost reasons for that. You could also make it an autotransformer, giving more freedom in choosing the number of LEDs and thus the voltage drop across the LED string.

Dave

I haven't seen these line-voltage lamps (but I am going to visit Wally World just to do so), but here is a guess at why they say not use a dimmer: they incorporate a transformer.

Anything that has a transformer (or ballast) in it, including fluorescent lights, will run hot, and may even catch fire on a dimmer.

This is because a transformer is designed to run on a fairly clean fixed frequency. Some of the pricier fixtures will allow either 50Hz or 60Hz, but they cannot deal with harmonics. Dimmers, because they work by abruptly starting the current part-way through each half-cycle, generate a large amount of harmonic energy.

Incandescent lights have no problem with this, and it makes the dimmer cheaper, but it will overheat anything with a transformer or ballast.

Anthony Straight

I think they use a simple switcher ckt. They have been known to burn out and smoke. Feeding them with a duty cycle dimmer would probably hasten that end.. Eric