Dimming LED

Status
Not open for further replies.

George Stolz

Moderator
Staff member
Location
Windsor, CO NEC: 2017
Occupation
Service Manager
I am looking to dim 650W of LED lamps on one switchleg. From what I can discern, the best i can get for Lutron is 250W.

Leviton claims a 1000W dimmer for LED. Is there fine print I am missing, or is Lutron sucking wind on this particular front?
 

Dennis Alwon

Moderator
Staff member
Location
Chapel Hill, NC
Occupation
Retired Electrical Contractor
I am looking to dim 650W of LED lamps on one switchleg. From what I can discern, the best i can get for Lutron is 250W.

Leviton claims a 1000W dimmer for LED. Is there fine print I am missing, or is Lutron sucking wind on this particular front?

Lutron makes a power booster - model PHPM-WBX120-WH - that fits in a 2 gang box that works with LED. You may need 2 but I would call them with the make and model of the fixture or bulb and see what they can do. I think the booster are good for 16 amps incandescent so 1/4 of that is 4 amps-- probably need 2

1-800-523-9466
 

Dennis Alwon

Moderator
Staff member
Location
Chapel Hill, NC
Occupation
Retired Electrical Contractor
Leviton claims a 1000W dimmer for LED. Is there fine print I am missing, or is Lutron sucking wind on this particular front?

Not sure I believe this because one of the issues with LED is that the inrush on them is 4 times that of incandescent lights. Not sure how they over come that but they might.

I just look up one 1000 watt LED dimmer and here is what it stated

[h=2]Item Description[/h]Z-Wave 450-Watt LED/CFL or 1,000-Watt Incandescent Scene Capable Dimmer with LED Locator - White/Ivory/Light Almond
 

Dennis Alwon

Moderator
Staff member
Location
Chapel Hill, NC
Occupation
Retired Electrical Contractor
Edison base A19 LED replacements.

Dennis, educate me: power booster...?


The wrong man to do that....

The power booster somehow increases the capacity of a 600 watt dimmer to almost 1920 watts. Remember LED now must be calculated at 1/4. 1920/4 = 480 watts so that is why you need 2.

The info on this page does not say LED but I am certain it will---http://www.lutron.com/TechnicalDocumentLibrary/369-358.pdf

51nZWwtGLXL._SY355_.jpg
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
160306-2257 EST

George Stolz:

What is the meaning of 650 W of LED? Is it actual wattage or incandescent equivalent? Have you run any tests on the particular LEDs to see what their characteristics are?


Dennis Alwon:

Not sure I believe this because one of the issues with LED is that the inrush on them is 4 times that of incandescent lights. Not sure how they over come that but they might.
What does this mean and where did the information come from?


A CREE 9.5 W (60 W equivalent) is rated at 0.079 A continuous RMS. At 4 A this would be 50 bulbs. The 0.079 A is the continuous repetitive current. I see no repetitive peaking. This bulb has a peak inrush of about 1.7 A, or a ratio of Peak Inrush to Steady State RMS of about 21. A 60 W incandescent has a peak inrush of about 8 A, or a Peak Inrush to Steady State RMS of about 16. But if you compare on a single bulb basis a 60 W equivalent CREE with a 60 W Incandescent, then the CREE is 1/4 the peak inrush of the incandescent. Steady state the CREE is 1/6 the RMS current of the incandescent.

Controlling the CREE with a phase shift dimmer I do not believe produces any particular repetitive current peaking except for possibly some high frequency oscillation. this I need to test.

Each different bulb is likely to create different problems for electronic controls. Some Feit bulbs have a lot of high frequency oscillation every AC cycle.

.
 

Dennis Alwon

Moderator
Staff member
Location
Chapel Hill, NC
Occupation
Retired Electrical Contractor
What does this mean and where did the information come from?

Not sure I believe this because one of the issues with LED is that the inrush on them is 4 times that of incandescent lights. Not sure how they over come that but they might.

I talked with Lutron Tech support. He told me that they were caught off guard when they first dealt with the LED bulbs. They assumed the 600 watt dimmer could handle 600 watts of LED but they found out that the inrush (his words not mine) current when you first turn them on is 4 times greater and it was burning out the dimmer. If you buy a 600 watt C/L dimmer it will do 600 watts of incandescent wattage but it is only rated 150 watts.

George mention 250 watts for LED was what he found because Lutron has one but it's their 1000 watt C/L which again will do 1000 watts of incandescent wattage but only 250 watts for the LED.

When I read the stuff at Leviton that George claimed they said would work on a 1000 watt LED all I found was a 1000 watt dimmer that did 400 watts of led-- not sure how they got it to work with 400 watts but that is what they claim.

I can tell you no more than that and you know I am techie inept....
 

George Stolz

Moderator
Staff member
Location
Windsor, CO NEC: 2017
Occupation
Service Manager
After further review, I think Leviton's spec sheets are simply incomplete. "These work with LED" and "This is a 1000W dimmer" do not equal "This dims 1000W of LED".

Not sure how to tackle this one. I think that for the moment, the room is not dimmable.
 

Dennis Alwon

Moderator
Staff member
Location
Chapel Hill, NC
Occupation
Retired Electrical Contractor
After further review, I think Leviton's spec sheets are simply incomplete. "These work with LED" and "This is a 1000W dimmer" do not equal "This dims 1000W of LED".

Not sure how to tackle this one. I think that for the moment, the room is not dimmable.


Lutron thinks it can be done but they need more info. You got the number if your want to pursue the answer.
 

chris kennedy

Senior Member
Location
Miami Fla.
Occupation
60 yr old tool twisting electrician
Dennis is close, looks just like a power booster but is actually a phase adaptive interface controlled by a dimmer.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
160307-1215 EST

Dennis:

Thanks.

I made a 1 ohm shunt from multiple carbon comp resistors to reduce the inductance of my current measuring shunt this morning. This reduces some ot the 5 to 10 MHz oscillation I was getting.

Tested 3 bulbs for inrush current. The source is 120 V 60 Hz house power. The switch is a 50 A mercury relay. In the past I have switched upwards of 10,000 A for a moment with this relay.

This morning's test results follow:

60 W incandescent
View attachment 14576


The peak inrush is about 8 A. Peak to RMS steady-state is 8/0.5 = 16. Total duration about 3 mS. An estimate of I^2*T = 8*8*0.001 = 0.064 WS.




9.5 W CREE (60 W equivalent)

View attachment 14574
.

.
View attachment 14575


The peak inrush is about 5 A. Peak to RMS steady-state is 5/0.079 = 63. Peak current is at about 100 microseconds, and total duration is about 400 microseconds. An estimate of I^2*T is 5*5*0.0003 = 0.0075 WS. Thus, the energy per turn on per bulb of the CREE is about 1/10 of that of the incandescent. The steady-state wattage ratio is 60/9.5 = 6.3 . Thus, from a steady-state current basis you could use 6.3 CREEs in place of 1 incandescent. From my estimated I^2*T ratio of 0.064/0.0075 = 8.5 you could use 8.5 bulbs..

Unless I am way off in my estimates equal actual wattages should work between CREE and incandescent.
.

.
9.5 W Feit with high frequency oscillation near 1 MHz (60 W equivalent)

View attachment 14577

I have not seen any inrush current on this Feit bulb. The AC voltage before the turn on time is from the capacitive coupling thru the mercury relay to what appears to be a very high impedance FEIT bulb when it is off. This is not a dimmable FEIT.


.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
160307-1556 EST

This forum has real problems with photo attachements, and restore of auto-save..

A redo of my last post.

160307-1215 EST

Dennis thanls.

I made a 1 ohm shunt from multiple carbon comp resistors to reduce the inductance of my current measuring shunt this morning. This reduces some ot the 5 to 10 MHz oscillation I was getting.

Three bulbs were tested for inrush current. The source was 120 V 60 Hz house power. The switch was a 50 A mercury relay. In the past I have switched upwards of 10,000 A for a moment with this relay.

The test results follow:

Red is voltage, and blue is current.

60 W incandescent.

DS2_QuickPrint71-I.JPG
.
The peak inrush is about 8 A for one bulb. The Peak to RMS steady-state current ratio is 8/0.5 = 16. Total duration of inrush is about 3 mS. An estimate of I^2*T = 8*8*0.001 = 0.064 A^2*S.


9.5 W CREE (60 W equivalent).

DS2_QuickPrint70-I.JPG




DS2_QuickPrint69-I.JPG


Both plots are exactly the same measurement. The first is at 2 mS/div, and the second at 100 microseconds/div.

The peak inrush is about 5 A for one bulb. Peak to RMS steady-state is 5/0.079 = 63. Peak current occurs at about 100 microseconds, and total duration is about 400 microseconds. An estimate of I^2*T is 5*5*0.0003 = 0.0075 A^2*S.

Thus, the energy per turn on per bulb of the CREE is about 1/10 of that of the incandescent. This inrush only occurs at turn-on, not every cycle. The steady-state wattage ratio between incandescent and CREE is 60/9.5 = 6.3 . Thus, from a steady-state current basis you could use 6.3 CREEs in place of 1 incandescent, and have the same total current. From my estimated I^2*T ratio of 0.064/0.0075 = 8.5 you could use 8.5 CREE bulbs for the same transient energy as the incandescent.

Unless I am way off in my estimates equal total actual wattages should work between CREE and incandescent.



9.5 W Feit with high frequency oscillation near 1 MHz (60 W equivalent)

DS2_QuickPrint72-I.JPG


I have not seen any inrush current on this Feit bulb. The AC voltage before the turn on time is from the capacitive coupling thru the mercury relay to what appears to be a very high impedance FEIT bulb when it is off. This is not a dimmable FEIT.


.
 

ELA

Senior Member
Occupation
Electrical Test Engineer
Peak Repetitive Inrush current

Peak Repetitive Inrush current

The issue is that LEDs do experience a peak repetitive inrush current when dimmed. That is why you have to base their load calc. on their equiv. incandescent rating rather than their actual wattage.

Not really 4x the inrush.

It is that the dimmer experiences a current spike on each half cycle continuously when dimmed. These current spikes stress the semiconductor switch with local heating of the die.
Incandescent filaments to do cool off between cycles so do not have repetitive inrush currents.

Have plots but you can just google it.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
160307-2247 EST


A reference for a mercury relay. Probably banned in California. The brand I am using is Ebert.

http://www.wolfautomation.com/assets/15/100NO-120AH-6A.pdf


ELA:

Plots of the CREE repetitive inrush current using 1 ohm shunt for current measurement. There is extremely low I^2*T energy in those repetitive pulse currents. The majproty of the energy per cycle comes from the turn-on to the next zero crossing. This current is barely visible..


DS2_QuickPrint73-I.JPG

.

.

DS2_QuickPrint74-I.JPG



Estimated I^2*T = 2*2*0.00003 = 0.00012 .

.
 

Dennis Alwon

Moderator
Staff member
Location
Chapel Hill, NC
Occupation
Retired Electrical Contractor
Status
Not open for further replies.
Top