On Sun, 26 Dec 2010 12:19:56 -0600, email@example.com
Who said anything about phase control?
PWM is pulse width modulation, control of power by duty cycle.
PWM works just fine on non linear loads as its output isn't operated linearly.
The output of PWM is either on or off. What controls average power is the
percentage of the time that the device is on. LEDs do fabulously well with PWM
Dummy, common incandescent (triac/SCR) dimmers use PHASE CONTROL, not PWM.
There *is* a difference.
No, it is PHASE CONTROL. Dimming is done by controlling the firing angle of a
triac or SCR. For an incandescent (resistive load) this works out to be
similar to PWM, but it is a *lot* different when you're talking about a load
as nonlinear as an LED. PWM is often used to dim LEDs, but from a DC supply.
On Sun, 26 Dec 2010 15:55:58 -0600, " firstname.lastname@example.org"
On a civil note I will point out the problem with phase control is the
voltage will drop below the amount to bias the diodes before you get a
lot of dimming.
This is what one looks like on a scope.
As a side note you can see a transformer really doesn't do much to the
wave shape, it just gets smaller. (part of a previous investigation)
Brightness of LEDs is controlled by how much current flows through them
- same as with dimmable CFLs. All that is necessary is to make the
current control circuitry in the LED "bulbs" compatible with and making
use from the usual dimmers, similar story as with CFLs.
- Don Klipstein ( email@example.com)
I've tried all of the ones available - at least four different brands -
at my local stores. (at significant cost, I might add.) They ranged
from unacceptable (really weird colors when dimmed, made dimmer hum,
only dimmed to maybe 50% brightness) to didn't even come close to
working as advertised (e.g. instead of dimming they would flicker and
hum and eventually shut off - no perceptible difference between them and
regular non-dimmable CFLs) This with several year old standard, off the
shelf Lutron dimmers.
I've heard tell of one particular dimmable CFL that is supposed to be
used with a special dimmer, I forget who makes it, sylvania maybe?
haven't cared enough to order them, and they're not readily available in
stores. In any case, the combination would be significantly more
expensive, if it did work (ever wonder why the demo displays in the home
centers don't simply show a CFL on a dimmer so you can try them? If
someone had something like that I would consider taking the plunge) than
simply shoving an incandescent in the fixture and replacing as
necessary. Wouldn't save enough in electricity or lifespan to come
close to paying for itself.
replace "roosters" with "cox" to reply.
KRW is right that there is some difference in the voltage drop across
the junction as the current rises., It is not a big difference but it
is a difference.(1.79 - 1.96 in the ones I tested over a range of 3ma
to 15ma) It could just be heat doing it at the high end. I did notice
in my flashlight tests that have no internal esistor that a very small
series resistance starts dropping the current very fast so they may be
using this "wall" of a heated junction to regulate the current,
similar to a nichrome wire.
The only way to get an LED to work is put a resistor in series.
Are you sure the (bulb, lamp, LED) was not a 'package'? They sell
them both ways, people who want only the LED (specialized), and
people who just want it to work. The later is what I think is
causing this confusion.
They sure look like garden variety LEDs to me. It is a clear package
and all I see is the normal standoffs with the junction on top.
The drop off when I started adding resistance was virtually instant
from the full brightness then it seemed fairly linear after that but I
really needed a smaller pot to see the first few ohms going in.
I have some real low resistance pots in my junk box but I didn't try
No, an LED will light very nicely with no resistor if the voltage is
close enough to the forward drop of the led. Excede that voltage by
very much and the current goes up real fast, very quickly exceding the
maximum power dissipation of the LED, and popping it.
If I set my lab supply to 15MA and connect a standard red LED I get
1.9 volts drop across it. I can stiffen that power supply so it can
supply 20 amps of power, and as long as I do not increase the voltage
above 1.9 volts, the current does not excede 15ma
Below 1.45 volts, the LED does not light at all, and it varies in
brightness quite linearly from 1.5 to 1.9 volts
With an 82 ohm resistor in series,changing the input voltage from 1.5
to 14, the voltage drop across the LED goes from 1.5 to 1.9 volts,
with the current reaching 15ma. (roughly.02 watts) while the poor
resistor is shedding 1.875 watts.
Without the resistor,25 ma of current produces a 1.91 volt drop,30ma
produces 1.98 volt drop,35ma is 2 volts, and going to 50 ma goes to
So doubling the current from 25ma to 50ma only raises the voltage drop
by .08 volts, which is roughly 4% on this 20ma rated device (I believe
that is the spec)
On Sun, 26 Dec 2010 23:02:08 -0500, firstname.lastname@example.org wrote:
It's not heat. The I-V curve is an exponential, just like any other
(semiconductor) diode. Heat will flatten out the ideal I-V curve further but
that generally isn't seen until the device is operating outside of normal. The
effects of heat can be seen if you change current (and measure the voltage)
Experiment: Take your diode and a resistor, connected in series with a
variable voltage source, as such.
| | |
| | .-. |
| I | | | R Vr
+ | | | |
| V '-' |
/+\ | |
V ( ) +----------o -
\-/ | +
| V -> |
| - Vled
| | |
=== +----------o -
Vary the voltage source and plot Vr and Vled in a spreadsheet. Calculate
(using the spreadsheet) I (=Vr/R) and then, again using the spreadsheet, graph
I (Y-axis) against Vled (X-axis) using the "scatter" option. This graph shows
the characteristic I-V curve of this diode and can be used to set the bias
resistors for any supply voltage (if the current is within the range of the
graph). Note that each diode will be somewhat different and diodes from each
batch (or "lot") can be significantly different. You might notice that LEDs
of the same part number and manufacturer may be "grouped" by brightness (and
even color). Two LEDs with the same specs may look entirely different if put
next to each other.
The internal resistance of the battery is the ballast. Measure the unloaded
voltage (open-circuit or Voc, below) of the battery and the voltage and
current (this will be difficult) of the battery/LED combination. The internal
resistance of the battery can be calculated from (Voc-Vled)/I.
I haven't really been following, but I was in the borg last night and
they now have quite a few. I noticed, right on the box, of at least one
"Dimmable". There is little reason why not, unless flicker related.
Getting better. I was looking at bulbs priced in the teens. I seem to
recall prices 3 times higher not long ago. But I'm no expert.
Wound up buying a CFL that looked just light a regular bulb (for $4),
with glass all the way to the screw. Got one floor lamp that a regular
CFL won't screw in.
Not being able to dim them makes them even
LEDs are CONSTANT CURRENT-driven,not voltage driven.
they have to have some sort of current regulator inside that keeps the
current constant regardless of input voltate swings.
To dim a LED,you would have to vary the input current,and the typical triac
lamp dimmers don't do that,they vary input voltage.
Plus,there's a fairly narrow range of current that produces light from the
LED chip,it's not linear.Most LED dimming schemes pulse modulate the
LED,changing the duty cycle of the current,and the eye averages the output.
you get better efficiency that way,but it's more complex electronically.
One of the things that makes LEDs "harder" to dim is that they dim
linearly with current. Normally this would be a nice thing, however
present dimmers are made for traditional bulbs. Filament bulbs dim
very non linearly with voltage. LEDs typically will have some sort
of series limiting resistance, but the idea is to have as little
series resistance as possible, to cut down on wasted power in the
resistor. Pulse width dimming, as in car tail lights, is the
obvious answer. But, I think they should up the frequency of said
tail lights and future house lights dimming circuits, to prevent the
strobing effect you see when you turn your head rapidly. Some
people don't even notice it, but some do.
I hadn't thought about this, but aren't light dimmers just varying
the fraction of a half wave (pulse width)? It would seem to me that it
should work with LEDs (although not particularly well) but I don't know
what circuitry is in an LED light.
I think the key is how the current source is constructed. Limiting
the power lost there is important.
Can you shed any light on just what is in an LED lamp? I'm curious
now. I figured a wee chopper with a bit of output filtering. Add a
resistor or a constant current configured transistor. Just speculating.
Is that a household type dimmer like on the wall?
changing the duty cycle of the current,and the eye averages the output.
For the money they are, they can throw a little circuitry at it!
Chopping up DC is no big deal these days. With the low current drawn and
high voltage, I wouldn't think it would need a very large cap.
I have no practical knowledge of this, so I'll defer to those with
The problem is the clock rate. LEDs respond virtually instantly and at
60 or 50hz the flicker would be horrible.
You could either clock your chopper a lot faster or just vary the
Since the currents are so low, a rheostat may not be out of the
As soon as Hosfelt gets my box here I am going to start playing with
some high intensity LEDs to see what I can do with them,
My goal is dimmable under counter lights that run off a wall wart with
a very low profile. (perhaps even recessed into the cabinet bottom)
HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.