Author Topic: Dummy Load for LED lights / eliminate flickering  (Read 226712 times)

Offline TimAlls

  • Hero Member
  • *****
  • Posts: 722
  • Karma: +0/-0
Dummy Load for LED lights / eliminate flickering
« on: January 27, 2012, 06:21:05 pm »
I would like to use Leviton in line Dimmers to control a group of LED lights. The lights flicker unless one normal bulb is left in place. I would like to wire a dummy load up to the circuit to smooth things out. I don't want to bury a hidden light and I don't want a heat generating resistor in line as well.
I was wondering if an inductor could be used to create the dummy load and if so how to calculate the value of such a load. Would this be less heat generating than a simple resistive circuit?
Any Ideas?
I am hoping you don't tell me to buy new dimmers!
Regards
Tim Alls
« Last Edit: January 30, 2012, 02:14:12 pm by TimAlls »

Offline kgrr

  • Full Member
  • ***
  • Posts: 155
  • Karma: +0/-0
  • My obsession: Totally automating my townhouse
Re: Inductive Dummy Load for LED lights
« Reply #1 on: January 28, 2012, 01:29:14 pm »
I had the same issue in my breakfast bar area where I wanted to replace 20 watt MR16 Halogen bulbs with 3 watt MR16s on my GE45606.   The dimmer is designed for incandescent only and works just fine on halogen bulbs, which consume half the power of incandescent bulbs and have a longer life.  But they don't work well with LED lighting because they blink at a dim setting.

I think the reason for this is due to the non-linear characteristics of the LEDs.  These dimmers (your Leviton and my GE/Jasco 45606) are dimming by controlling the current.  They require current flow in order to operate.   However, when operated much below the "knee voltage" of the LED diode, literally no current flows.  The dimmer tries to bring the current up and the light comes on.  When this happens, the dimmer senses that too much current is being drawn and turns it off again.  This is why the LEDs blink at low current.

The LED cannot operate below that knee voltage.  The load you are providing with one bulb in the circuit is simply allowing current to flow and thus preventing the blinking.  However, the LEDs just don't dim below a certain brightness level and turn off.

I think these dimmers could be tricked into not allowing the dimmer to operate at lower settings than the minimum.  I just don't understand enough yet about how to adjust the characteristics of the dimmers.  e.g. setting the minimum current setting.
« Last Edit: January 28, 2012, 01:40:22 pm by kgrr »
---------------------
Vera2 (1.5.255, UI5); ACT ZRW113W; Aeon HEM, SmartSwitch, GE/Jasco 45600, 45601, 45602, 45603, 45604, 45605, 45606, 45609, 45614; Hawking HRDS1; Intermatic CA9000; Schlage BE369GR, WCE100; SQ Blaster Plus; Squeezebox; Trane TZEMT400AB32MAA

Offline guessed

  • Master Member
  • *******
  • Posts: 5293
  • Karma: +90/-22
  • Release compat is not a bolted-on afterthought
Re: Inductive Dummy Load for LED lights
« Reply #2 on: January 28, 2012, 03:17:41 pm »
@TimAlls,
It's also LED Bulb sensitive.  I have Leviton Z-Wave  (VRI-06) dimmers throughout the house, and put in 4x LED Bulbs (Utilitech, from HomeDepot) 7.5W "40W equivalent" to replace 4x 65W Incandescent bulbs as an experiment of sorts.

See previous report here:
    http://forum.micasaverde.com/index.php/topic,6074.msg41244.html#msg41244

Longer term notes:
a) They don't soft-start as well as the original Bulbs, but that doesn't bother me
b) In practice, I haven't noticed the flickering since setting them up
c) I never run them at the lowest setting, which is where I noticed a bunch of flicker, as that level is impractical anyhow
d) The room has 1 switched (not dimmed) Lamp that I use almost all the time, with the same type of LED Bulb
e) The room has other [incandescent] bulbs that I use for more controlled lighting (eg. front dimmable lights for the TV)
f) I'm about to replace two other sets of "ganged" lighting (one 4x, one 5x) with similar LED Lighting, and have no concerns as they're mostly field-lighting, and never dimmed to the lowest settings

Offline autotoronto

  • Full Member
  • ***
  • Posts: 190
  • Karma: +0/-0
Re: Inductive Dummy Load for LED lights
« Reply #3 on: January 28, 2012, 07:17:30 pm »
In answer to the original question:

I suspect it will kill your dimmer. Inductors generate a back emf to resist the rate of change of current through them; a triac dimmer switches the load on part-way through the mains cycle, which means there will be significant interaction between your inductive load and the power output stage of the dimmer.

I bet if you look at the specification for your dimmer it will say "not for switching motors or other inductive loads".

Typically you can make an inductive load look "less inductive" by putting a low resistance (heavy load) in parallel with it; that would be like ballasting a fan motor with a lamp. But what you're seeking to do is to have the inductive part of the load much more significant than the resistive part.

Offline mikeburr

  • Newbie
  • *
  • Posts: 3
  • Karma: +0/-0
Re: Inductive Dummy Load for LED lights
« Reply #4 on: January 28, 2012, 09:20:43 pm »
Same situation as autotoronto, but with LED rope lights.  Do any of the Leviton dimmers work with low-wattage LED lighting?  If not, any suggestions for a different brand dimmer that does?

Offline TimAlls

  • Hero Member
  • *****
  • Posts: 722
  • Karma: +0/-0
Re: Inductive Dummy Load for LED lights
« Reply #5 on: January 29, 2012, 12:04:45 pm »
Thanks for all the feedback.
Autotoronto....you had interesting information and seem to understand the inductor question....here is more food for thought.
I have 10 lights in my overhead all wired in parallel.
When I use all Led replacements the lights flicker and don't dim correctly. When I leave one incandescent light in the circuit the led lights dim perfect.....all the way to nothing!
This has to be a controller issue.....not getting the operating voltage it needs to perform. I Don't agree with the"LEDs dim by current" ....the lights are in parallel so the fact that leaving one incandescent light bulb in the circuit shows that the current to each light is independent and that the controller is the culprit. In my opinion, the controller is in series with the circuit and needs a stable load to create the voltage drop across the controller and feed its circuitry. So again, there must be a simple way to stabilize the voltage that the controller needs to preform its duty of slicing up the AC wave and feeding it to the LED light bulbs. Maybe the inductor is a bad idea.....any other ideas from those of you with circuit design experience????
Thanks
Tim Alls
Edit.....I will use my scope on the controller and look at the difference with one incadecent bulb added....a picture could help figure this one out.
« Last Edit: January 29, 2012, 12:11:44 pm by TimAlls »

Offline autotoronto

  • Full Member
  • ***
  • Posts: 190
  • Karma: +0/-0
Re: Inductive Dummy Load for LED lights
« Reply #6 on: January 29, 2012, 06:34:19 pm »
Here's what I think's happening. It's guessing, from a basic knowledge of how dimmers work (significant time working on dimmers and lighting control systems for theatres and opera houses), and how microcontrollers work, so forgive me if it's not entirely accurate.

LEDs are low voltage devices which have drive modules that rectify the ac to dc and transform the voltage down to the 2-4 volt range to power the LED chips (at a fairly high current, for a semiconductor device).

A dimmable LED unit has a smart driver that's able to power its own logic from a power supply derived from a wide range of incoming wave forms (nearly off to fully on), sense the incoming wave form, and make a guess at what power to drive the LED's at, probably with a high frequency square wave using PWM. I don't think they actually work by using the chopped ac waveform to drive the LED in any direct sense.

The two-wire dimmers power their internal electronics by having a small resistance in series with the load. If current flows through the load a small voltage develops across this resistor (in the dimmer unit) which can be tapped, rectified, pumped up etc to 3 or so volts with a current of a few 10s of milliamps to drive the microprocessor in the zwave unit.

The dimmer powers up the load (the LED's). The LED drivers power up. They power on the LED chips. The z-wave dimmer now has enough current flowing that it can siphon off enough to start up the z-wave and control circuits. The control circuits now try to dim the attached load to the correct level restricting the mean current flow. Unfortunately this has the effect of starving the microcontroller of the power to operate, so it shuts down. The load defaults to full on again, so there's enough current for the microcontroller to power up, and the cycle starts again. result: flickering.

If you include the incandescent in the circuit then the controller can leak enough current through the load that it can tap some power to remain operational even when the load appears to be off.

The best way to "stabilize the voltage" is to provide the dimmer with a neutral: then the power to the microcontroller doesn't need to depend on leaking current through the load while the load is "off". This is why dimmers rated for CFL's and LED's always require that neutral. The second way is to provide a dummy load in parallel with the true load, either a resistor or an incandescent of some minimum wattage.

If there was another way to do it - you can be quite sure that you'd be seeing a lot of "CFL/LED-capable dimmers" on the market that only require the two wires.

Offline TimAlls

  • Hero Member
  • *****
  • Posts: 722
  • Karma: +0/-0
Re: Inductive Dummy Load for LED lights
« Reply #7 on: January 29, 2012, 11:13:07 pm »
I will put the scope on the dimmers Monday. And check it with and without the added incandescent bulb......the following came from another forum:

Has anyone tried a resistor in the circuit? If so, what value and what would the wattage rating be? Would there be any concerns with heat? How high of a resistance could a guy get away with and still have it eliminate the flicker? Thanks again for all your help.

Geo
Senior Member

Canada
322 Posts

 Posted - 07/31/2011 :  10:41:05 AM    
I did try it, successfully, but there may not be a universal value for the simple reason that the devices, that is CFLs and LEDs don't seem to be standardized as yet, so the values may be all over the place.
More often than not a 5W light bulb will do the trick. That represents a 2.9kOhm resistor, which will dissipate 5W. To be safe, you need to use at least a 10W resistor and now you have to worry about the safety both electrical and fire. Using a light bulb hidden somewhere solves these problems at a fraction of the cost. The added power consumption, whether you use a bulb or a resistor, is negligible, if you desire dimming.
The main problem you're trying to fix is caused by CFLs or LED lamps being reactive loads. Consequently, the current drops below the holding current of the triac too soon and the triac turns off. That causes the flicker. The holding current depends on the triac used by the dimmer manufacturer. In several cases I managed to get everything running with a 100kOhm resistor, that is 1.2mA additional current, 144mW heat dissipation. That allowed using a small 1/2W resistor placed in the junction box.
That said, DON'T DO IT if you don't know your ways around electricity! Judging from your questions, you don't. A small bulb should be your only alternative.
GJN
End quote....
Very interesting!
Regards
Tim Alls

Offline autotoronto

  • Full Member
  • ***
  • Posts: 190
  • Karma: +0/-0
Re: Inductive Dummy Load for LED lights
« Reply #8 on: January 30, 2012, 06:27:16 am »
Quote
To be safe, you need to use at least a 10W resistor and now you have to worry about the safety both electrical and fire. Using a light bulb hidden somewhere solves these problems at a fraction of the cost.
Just one thing to add. From a safety perspective, there's no real difference in the fire risk of a 5W light bulb and a resistor dissipating 5 watts. Both give off the same 5 watts of heat. You may think the resistor is worse because its surface reaches a higher temperature than the glass envelope of the lamp but that's only because the resistor is smaller. Give it the same amount of room around it as you would for a lightbulb and the temperature rise will be the same.
« Last Edit: January 30, 2012, 06:28:59 am by autotoronto »

Offline big517

  • Hero Member
  • *****
  • Posts: 576
  • Karma: +1/-3
Re: Inductive Dummy Load for LED lights
« Reply #9 on: January 30, 2012, 07:19:58 am »
I'm am very interested to see if anyone executes on this.. I have been wondering about this exact same thing I have over 30 dimmers and only 2 control a non-led load... This is a battle I'm constantly fighting.. I was able to find a (temporary) solution using luup code because I found the flicker would happen at certain ranges and this would eliminate the light to be set to that level through Vera.. if you search for my posts you'll find it.
I have a question to add to this topic.
When I turn on a light and it's ramping up per the natural ramp rate of the dimmer, I noticed that the other lights on different switches (but same circuit to fuse box) would ramp down or up as if they are receiving some signal to do so... Could it be because they use the same common? They do not use 3 way travelers wires. Would wiring this resistor solve that as well?

I'm using all GE single and 3 way Dimmers and the ecosmart 11 watt LED retrofits from home depot (cree LR6 I believe)



Offline TimAlls

  • Hero Member
  • *****
  • Posts: 722
  • Karma: +0/-0
Re: Inductive Dummy Load for LED lights
« Reply #10 on: January 30, 2012, 02:13:26 pm »
Hi All,
I just finished testing the leviton 1000 watt dimmers with Philips 7 watt 280 lumen LED lights and here is the results:
1. by adding one 10 watt 3.3K resistor to the circuit (tied between output of dimmer and neutral wire) the flickering disappears but dimming rate is still unstable.
2. by adding two 10 watt 3.3K resistors to the circuit (in parallel) the flickeing is gone and the dim rate is stabilized
3. with the two resistors I changed the number of LED lights from 10 down to no lights at all and the circuit is stable....dims correctly with any number of lights
4. with the two resistors and no LED lights and the dimmer functions fine.....it means that the ZWAVE will alway function independant of a no load situation! Never a dead node to disrupt the ZWAVE mapping which really slows down the network.

I will look at it on the scope next but I would say SUCESS! I am looking at a safe way to implement the resistors next....

More to come

Regards
Tim Alls
« Last Edit: January 30, 2012, 02:44:49 pm by TimAlls »

Offline pgrover516

  • Beta Testers
  • Hero Member
  • *****
  • Posts: 1013
  • Karma: +0/-0
Re: Inductive Dummy Load for LED lights
« Reply #11 on: January 30, 2012, 03:06:36 pm »
Hi All,
I just finished testing the leviton 1000 watt dimmers with Philips 7 watt 280 lumen LED lights and here is the results:
1. by adding one 10 watt 3.3K resistor to the circuit (tied between output of dimmer and neutral wire) the flickering disappears but dimming rate is still unstable.
2. by adding two 10 watt 3.3K resistors to the circuit (in parallel) the flickeing is gone and the dim rate is stabilized
3. with the two resistors I changed the number of LED lights from 10 down to no lights at all and the circuit is stable....dims correctly with any number of lights
4. with the two resistors and no LED lights and the dimmer functions fine.....it means that the ZWAVE will alway function independant of a no load situation! Never a dead node to disrupt the ZWAVE mapping which really slows down the network.

I will look at it on the scope next but I would say SUCESS! I am looking at a safe way to implement the resistors next....

More to come

Regards
Tim Alls
Awesome, will look forward to to hearing about your implementation
V1,V2,V3,VLite,Express Controls HSM-100,Intermatic HA20C, HA04C,HA02C,HA09, Leviton VRP15-1LW, VRS15-1LX,Home Manageables HM-TS001,Schlage FE599, Schlage BE369, Cooper RF9500, Aeon Labs Minimote, Schlage TZEMT400AB32MAA+more

Offline autotoronto

  • Full Member
  • ***
  • Posts: 190
  • Karma: +0/-0
Re: Dummy Load for LED lights / eliminate flickering
« Reply #12 on: January 30, 2012, 05:21:24 pm »
So assuming you're in North America, you've added a 1650 ohm shunt, passing a 70mA current and dissipating about 9 watts.
« Last Edit: January 30, 2012, 05:23:32 pm by autotoronto »

Offline TimAlls

  • Hero Member
  • *****
  • Posts: 722
  • Karma: +0/-0
Re: Dummy Load for LED lights / eliminate flickering
« Reply #13 on: January 30, 2012, 06:09:39 pm »
So assuming you're in North America, you've added a 1650 ohm shunt, passing a 70mA current and dissipating about 9 watts.
Yes that's correct. However, that load is only present when the lights are on......so there is little or no heat at all when dimmed down or in the off position. The easiest and safest way to install these is to place them in the lighting cans in parallel with the light socket. The cans have a thermal shutdown feature which would insure safe behavior and the metal can would easily withstand the heat from a 10 watt load. This is a job for an experienced electrician....if you are not please don't do it.....using the wrong components or leaving a bad connection could start a fire so beware!
Regards
Tim Alls
AllSeas Yachts

Offline autotoronto

  • Full Member
  • ***
  • Posts: 190
  • Karma: +0/-0
Re: Dummy Load for LED lights / eliminate flickering
« Reply #14 on: January 30, 2012, 09:52:36 pm »
Quote
However, that load is only present when the lights are on......so there is little or no heat at all when dimmed down or in the off position.
True, but also for safety you need to plan your heat budget under the scenario that the dimmer is on 100% 24/7, even if in practice it isn't.
« Last Edit: January 30, 2012, 10:24:03 pm by autotoronto »