Erratic LED Lighting

Status
Not open for further replies.

Jim M426

Member
Location
SW Florida
I recently purchased an old Florida house that was in dire need of lighting (among other things). After replacing the 200A, 20 space panel with a 200A, 42 space panel I started doing electrical upgrades in the house. Thinking I'd save some electrical bill $$$ going with LEDs, I bought some LED can lights and dimmers that the lighting distributor approved as compatible. I settled on 3" and 4" cans with trims rated at 9 watts each.

For the kitchen and lanai lighting I ran a single 15A circuit fed by 14/2 NC cable. On that circuit are:
  • (6) 3" cans controlled by (1) Lutron Maestro MACL-153M and (1) Lutron Maestro MA-R (kitchen ceiling)
  • (4) 3" cans controlled by (1) Lutron Maestro MACL-153M and (1) Lutron Maestro MA-R (soffit)
  • (3) 3" cans controlled by (1) Lutron Maestro MACL-153M and (1) Lutron Maestro MA-R (island)
  • (6) 4" cans controlled by (1) Lutron Maestro MACL-153M (lanai)
  • (2) 3" cans controlled by (1) Lutron Maestro MACL-153M (above sink)
  • (2) 4" cans controlled by (1) Lutron Maestro MACL-153M (above bar)

When the lights in the lanai are on and I turn on the kitchen ceiling lights, the lanai lights brighten then dim back to where they were.

When the lanai lights are turned off (regardless of the state of other lights on the circuit) 5 of the 6 lights remain lit but very dim so you can only see it when it's dark. The only one that is completely out is the one first fed from the dimmer and through which the remainder are fed.

I did the installation a little at a time (I don't move so fast anymore) and the lanai & bar were the last sets of lights I installed. Everything seemed to work fine up to adding the bar & lanai lights.

Any thoughts on why the LEDs are acting up?

Thanks,
Jim
IBEW (retired)
 

Electric-Light

Senior Member
I recently purchased an old Florida house that was in dire need of lighting (among other things). After replacing the 200A, 20 space panel with a 200A, 42 space panel I started doing electrical upgrades in the house. Thinking I'd save some electrical bill $$$ going with LEDs, I bought some LED can lights and dimmers that the lighting distributor approved as compatible. I settled on 3" and 4" cans with trims rated at 9 watts each.

You thought wrong. Lighting energy consumption is not a significant percentage of residential energy profile and making LED conversions(not upgrade) with high expectations means feeding quarters into a vending machine that dispenses the same quantity of nickels. Changing from 20 space to 42 space don't really resonate with saving power usage either. LED lights are usually the most expensive option as they have the most problems, require the most time and effort to commission properly when combined with dimming. How far are you up to just in parts expense for all that Lutron gear? I'd bet it takes many years if ever for savings from power savings from those lights to even pay off the cost for all that.

When the lights in the lanai are on and I turn on the kitchen ceiling lights, the lanai lights brighten then dim back to where they were.
You've given us the details about the controls, but what about the specific details of the load? Compatible doesn't mean absolute freedom from noticeable annoying behaviors.

Phase cut LED dimming is touchy and susceptible to problems for technical reasons. It is the type you're using, and it is the most common type in residential applications. When you turn on the kitchen light, it draws an inrush current which causes a very brief voltage dip. The L. E. D. ballasts in your lanai lights are tripping out. Does it only happen when the lanai lights are deeply dimmed?

When the lanai lights are turned off (regardless of the state of other lights on the circuit) 5 of the 6 lights remain lit but very dim so you can only see it when it's dark. The only one that is completely out is the one first fed from the dimmer and through which the remainder are fed.

Microprocessor dimmers have a bleed current and the particular design of LED ballast that happens to be used in your fixtures/lamps allows the tiny amount of current flowing through the fixtures to impose current on the LEDs. You only need a couple microamps to make an LED give off visible light. If you throw a regular incandescent lamp into the string, I bet they'd extinguish.


Any thoughts on why the LEDs are acting up?

Thanks,
Jim
IBEW (retired)

You're using a combination of modern fads that makes the system prone to shakes and rattles like you're experiencing that is expensive to commission as the final fine tuning requires hand fitting by trial and error.
 

tom baker

First Chief Moderator
Staff member
I disagree on the savings from LED lighting. I heat, cook, dry and heat water with Nat Gas so my electrical loads are minimal. In my basement I had square recessed cans, they were 100W, I went to 26 W CFL and now have 13 W LED, 5 each that's a considerable savings.
I have no problems with my LEDs, they are far better than the CFLs. I have not had a screw in A base LED fail over 3-4 years, in the same time I had many CFLs fail.
 

b1miller

Member
Location
Washington
Occupation
Electrician/Electrical Administrator/SCADA Engineer
LED residential lighting

LED residential lighting

I would also agree that there can be substantial savings when installing LED lights. I have replaced 15 of the 65W floods in my house with replacement LED floods. I have also swapped out another 12-15 or so incandescent bulbs with the LED equivalent bulbs inside and outside. Over the last 3 years I have not had to replace a single one. Got most from Costco or other big box store on sale or with manufacturers rebates. Even my wife has noticed a reduction in our electric bill. Looks like a 10-20% savings so far. No CFL's in my house any more!! Even went to the 4 foot LED fixtures in my shop.
 

Electric-Light

Senior Member
Because, LEDs are very labor intensive to commission

Because, LEDs are very labor intensive to commission

I disagree on the savings from LED lighting. I heat, cook, dry and heat water with Nat Gas so my electrical loads are minimal. In my basement I had square recessed cans, they were 100W, I went to 26 W CFL and now have 13 W LED, 5 each that's a considerable savings.

Savings depends on annual hours. I've found vacancy sensors great for limiting the duration lights can be accidentally left on. It works as a normal manual on/off switch. Count down starts from the last motion detected, which can be set from about a minute to 30 minutes. Many have auto-restart if movement is detected within 30 to 60 seconds of shut off before reverting to manual on mode. It's much less burdensome than twist timer since count down resets whenever there's motion.

I have no problems with my LEDs, they are far better than the CFLs. I have not had a screw in A base LED fail over 3-4 years, in the same time I had many CFLs fail.

That is compared to CFLs, which do not dim well except for externally ballasted 4-pin type.

Bunch of expensive dimmers he bought suggests he's quite demanding on aesthetic performance.
It requires 1% dimming to "seem like 10%" because of non-linear response of human vision and that's the expectation held by common people. Flicker free, full range, stable, silk smooth fade to dark is very technically challenging to do with LEDs especially with phase cut dimming and presents an enormous burden.

Dimming to 10% lumen output feels like 32% as bright and this draws quite a few complaints about LED lamps with comments such as "it only seems to dim to about a 1/3 and it says 10%....". Going below 10% ("feels like 1/3") have negligible energy advantage while making it exponentially increasing commissioning difficulties, stability issues and other performance problems for technical reasons.


Hand fitting to make ever fluid products work correctly and getting it stabilized requires time, which must be added to the bill, thus making it an very expensive technology to implement.

Getting LEDs to work flawlessly, consistently and reliably over the entire range is usually EXTREMELY difficult. Usually, the trouble is on the lower end setting.

If you dim to the lowest level it will remain on, it may not come back on if shut off at this level and turned back on without bumping it up a significant level, then bringing it back down. Whether done manually or automatically, this will cause a pronounced flash.

Dimmable LED ballasts are far more prone to flipping out at line noise than incandescent. For example, other loads switched (as threadstarter complained). I know that the ballast in some CREE bulb gets confused when a fan is switched off on the same circuit and causes the lamp to completely shut off and restart even when it is not on a dimmer.

Some combination have a sore band. When you're inside the sore band, the lamp exhibits shimmer.

So I still conclude that getting LEDs to satisfy high end decorative dimming comes at a very high cost that it blows away any savings.

Finally, flicker, color and brightness unevenness becomes more pronounced under deep dimming and presently, LEDs exhibit very unnatural looking color as they're dimmed. Color remains pretty much the same but turns bluer on the extreme low end,which is the opposite of incandescent lamps which leads to "just don't like the color" dissatisfaction.

Residential grade LED lamps have "years of life" assuming 3 hrs per day. 65W saving (100w bulbs on dimmer, likely not used at full level all the time, so if dimmed power use is 75W and L. E. D. substitute uses 10W) at 3 hrs per day is an annual saving of 71 kWh, which is <$10 per year per lamp saved.

Even if added cost for the fixture is small, the increased cost in using high cost specialized controls and added labor cost for commissioning with time consuming adjustments and likely multiple visits may make it impossible to save anything. This thread is a very example of very common LED commissioning frustrations.
 
Last edited:

Fulthrotl

~Autocorrect is My Worst Enema.~
Any thoughts on why the LEDs are acting up?

Thanks,
Jim
IBEW (retired)

any chance you have a loose neutral somewhere
in the mix? it's a long shot, but....

those dimmers are particularly tolerant, but sometimes stuff
just has teething problems... i'd try this....

take one of those dimmer out of the mix, and substitute
a lutron diva c-l, and see how it works. DVCL-153P

if it doesn't work with that, you've got lights that aren't
ever gonna dim well, imho.

i've got above 50 DMF led can lights in my house, i'm
running those dimmers, and while you can dim them
almost to black, they won't fire at that low level with those
dimmers, but really nothing that isn't 0-10v. controls will.

and yeah, you can save some money on your electric bill.
if i turn on every led in my house, and dim them all the
way down, they draw total, about 40 watts, measured with
a fluke 345.

and electric-light has some strong opinions about LED lighting
in general... but you probably already got that.

let us know how it turns out.
 

K8MHZ

Senior Member
Location
Michigan. It's a beautiful peninsula, I've looked
Occupation
Electrician
I had the same thoughts with one circuit effecting another.

OP says all the lights are on the same circuit.

For the kitchen and lanai lighting I ran a single 15A circuit fed by 14/2 NC cable. On that circuit are:

  • (6) 3" cans controlled by (1) Lutron Maestro MACL-153M and (1) Lutron Maestro MA-R (kitchen ceiling)
  • (4) 3" cans controlled by (1) Lutron Maestro MACL-153M and (1) Lutron Maestro MA-R (soffit)
  • (3) 3" cans controlled by (1) Lutron Maestro MACL-153M and (1) Lutron Maestro MA-R (island)
  • (6) 4" cans controlled by (1) Lutron Maestro MACL-153M (lanai)
  • (2) 3" cans controlled by (1) Lutron Maestro MACL-153M (above sink)
  • (2) 4" cans controlled by (1) Lutron Maestro MACL-153M (above bar)

So, I doubt it's a problem with the neutral.

Everything seemed to work fine up to adding the bar & lanai lights.

What happens when you 'back up time' and disconnect the bar and lanai lights? Does the problem go away?
 

Jim M426

Member
Location
SW Florida
Thanks for the replies. I've had many of the same thoughts about possible problems. The wiring is good, I've checked it several times to see if there was any installation problems I missed. Everything tells me it boils down to the dimmers and LEDs.

So everything works fine until I install the lights in the lanai and at the bar. Both the latter are 4" trims, as opposed to all 3" trims previously installed. Something quirky in manufacturing the 4" trims? I don't know. I know little about troubleshooting LED lighting because I never ran into it on the job.

The other difference is the lanai and bar are each single pole switched. Could the companion dimmers on the 3-ways be working to stabilize the current? That would be something I'd have to take to Lutron but I wouldn't expect any eureka response.

In my last house I turned to the Maestro dimmers because I wanted dimming control at all locations and became hooked. All the lights there were incandescent. I had over (70) 6" cans with 65W lamps in that house and my electric bill showed it. After installing the LEDs here, I've barely noticed the bill increase. But I was skeptical about dimming LEDs to the low level I like in the evening. The Maestro dimmers allow me to program the dimmers to get to that level. All looked good until I added the last lights.

Disconnecting the bar and lanai lights, and no problems. But the problems have all been in the lanai lights. Those are the ones that, when dimmed, brighten up when adjusting dimming for the kitchen ceiling lights. They are the only switched sets that have 6 lights on them. But the lanai lights (all 4") is the only set that never turns fully off. The dimmer has a tab on the bottom of it that is supposed to completely cut the power to the fixtures, designed for changing out a lamp. When I pull the tab out, the lanai lights remain in the same dimmed state. So there is still a trickle current there. That tab is designed to completely open the circuit, cutting the trickle completely. That's not happening.

I'm going to have measure the current at each fixture through all states of dimming and once the tab is pulled and see what comes up. I think the dimmers are affecting one another through the trickle current. I don't think there's a bleed over through the wiring but that tab issue has me wondering.

One other thing that is probably completely unrelated is all the switch boxes are tied together by plastic flex. Then I pulled the travellers and whatever wires I needed through the flex. Feeds to the lights are done in 14/2 NM. I come from the land of conduit where NM is sacrilege but I can't see how the NM drops to the switch box could be causing any cross wire bleed over. It's probably just my conduit conditioning causing me to feel guilty for using NM. :roll:

I've got a couple of other thoughts I have to check out and see how that goes. Thanks again for the responses. My enthusiasm to solve this has been rejuvenated.
 

K8MHZ

Senior Member
Location
Michigan. It's a beautiful peninsula, I've looked
Occupation
Electrician
The dimmer has a tab on the bottom of it that is supposed to completely cut the power to the fixtures, designed for changing out a lamp. When I pull the tab out, the lanai lights remain in the same dimmed state. So there is still a trickle current there. That tab is designed to completely open the circuit, cutting the trickle completely. That's not happening.

Suspecting a bad dimmer, I would check for voltage across the load terminals after the tab was pulled out. If present, which I think will be the case, the next thing I would do would be to disconnect the hot load connection and see what happens. If the lights go out, check for voltage at the hot load terminal. If voltage is present and you KNOW the line and load hots are wired correctly, I would replace (or swap) the dimmer and re-check.
 

Electric-Light

Senior Member
any chance you have a loose neutral somewhere
in the mix? it's a long shot, but....
The thread starter mentioned all the lights are on the same branch and the interaction originates from devices on the same branch. Dimmers fire twice per AC line cycle, 120 times per second. Neutral influence from other phases can cause misfiring but in this case it doesn't involve influence on neutral from another pole.

Dimmers that require neutral takes all the timing reference from which to advance or retard the timing but those designed for highest wiring compatibility are designed to function in switch loop configuration.

those dimmers are particularly tolerant
Your perception of my opinions of LEDs don't change the fact that this is a technical challenge.
www.apec-conf.org/Portals/0/Industry Session Presentations/2016/1053.pdf

The relationship between input Vrms and output Vrms present at incandescent lamp as well as its dimming behavior is absolutely predictable.
The only function of a phase cut dimmer is advance/retard timing referenced to zero crossing. Different LED ballasts behave differently with different waveform even if RMS voltage is the same.

Notch, spike that's present in power line that occurs during the on cycle passes right onto the load even if the dimmer is unaffected, but what happens to the light output depends on each LED ballast design. When an inductive load is switched off, it produces a spike. Just as I mentioned earlier, the LED ballast used in some LED lamp will flip out, shut off and restart like the original "CREE LED light bulb" or flash even if not used with a dimmer.

The effectiveness of notch and spike dampening is entirely up to the LED ballast. The time constant of household incandescent lamps are extremely well known and very short transients get dampened out to avoid illumination quality issues. Most electronic ballasts have a capacitor for the DC link. The size and inrush current mitigation depends on each design and this is just as applicable to ballasts for the LED type load too.

take one of those dimmer out of the mix, and substitute
a lutron diva c-l, and see how it works. DVCL-153P
This is a DIY project, but you're echoing my point. This takes time and projects known for high chance of complication adds time and therefore cost. The cost of products also goes up. Obviously, LEDs are much needy products and require vastly more technical support resources to support installers.

and yeah, you can save some money on your electric bill.
if i turn on every led in my house, and dim them all the
way down, they draw total, about 40 watts, measured with
a fluke 345.

Hypothetical situation assuming whatever non-LED solution would be left on all the time. You don't know the nor does the customer so you have to guess and BS some but LED sales advisers tend to fill the blanks with ridiculous numbers to project positive ROI in a single digit years.

It's important to distinguish kWh, $ and service type. Favorable words about savings from PG&E and SCE users on residential kWh tiered structure is totally meaningless. One of them seems to be in PSE territory which is about as low as power gets.

electric-light has some strong opinions about LED lighting
I don't think anything I said in this post is an opinion. ;)

Thanks for the replies. I've had many of the same thoughts about possible problems. The wiring is good, I've checked it several times to see if there was any installation problems I missed. Everything tells me it boils down to the dimmers and LEDs.

But what about the model(s) of the lamp/fixtures?
Does the flashing occur regardless of the dimmer setting on the load that flashes or only when it's very low?


So everything works fine until I install the lights in the lanai and at the bar. Both the latter are 4" trims, as opposed to all 3" trims previously installed. Something quirky in manufacturing the 4" trims?
Not about 3" vs 4". Just whatever LED ballast used in that engineering revision of whatever product you have.

The other difference is the lanai and bar are each single pole switched. Could the companion dimmers on the 3-ways be working to stabilize the current? That would be something I'd have to take to Lutron but I wouldn't expect any eureka response.
Where is the dimmer getting power from? If it doesn't tie into a neutral, the only way is through through the load. The "companion dimmer" is an electronic device, so it too requires power.

In my last house I turned to the Maestro dimmers because I wanted dimming control at all locations and became hooked. All the lights there were incandescent. I had over (70) 6" cans with 65W lamps in that house and my electric bill showed it. After installing the LEDs here, I've barely noticed the bill increase. But I was skeptical about dimming LEDs to the low level I like in the evening. The Maestro dimmers allow me to program the dimmers to get to that level. All looked good until I added the last lights.

The low-end limit isn't determined exclusively by the dimmer. The circuitry of the ballast used in the lamp/fixture has a greater influence but they both affect it. The lowest level you can reach may not start up if you were to shut off at that level.

Disconnecting the bar and lanai lights, and no problems. But the problems have all been in the lanai lights. Those are the ones that, when dimmed, brighten up when adjusting dimming for the kitchen ceiling lights. They are the only switched sets that have 6 lights on them. But the lanai lights (all 4") is the only set that never turns fully off. The dimmer has a tab on the bottom of it that is supposed to completely cut the power to the fixtures, designed for changing out a lamp. When I pull the tab out, the lanai lights remain in the same dimmed state. So there is still a trickle current there. That tab is designed to completely open the circuit, cutting the trickle completely. That's not happening.

Indefinitely or under a minute? It's normal for some LED based lights to glow at extremely low level for some time after powering down as the capacitors in the ballast discharge.

I'm going to have measure the current at each fixture through all states of dimming and once the tab is pulled and see what comes up. I think the dimmers are affecting one another through the trickle current. I don't think there's a bleed over through the wiring but that tab issue has me wondering.
If grounding is not proper, voltage can appear between frame and neutral.
 

iwire

Moderator
Staff member
Location
Massachusetts
What would be your reasoning? How would a bad neutral cause lights to get brighter and never turn off?

Experience tells me to do my own research. When I go to troubleshoot something I listen to what the users or other troubleshooters have to say but I take it with a grain of salt and start from scratch.

If you have one thing affecting another that typically means something is in series instead of parallel. That being the case mis-connections of a neutral would be a good place to start and easy to spot.
 

K8MHZ

Senior Member
Location
Michigan. It's a beautiful peninsula, I've looked
Occupation
Electrician
Experience tells me to do my own research. When I go to troubleshoot something I listen to what the users or other troubleshooters have to say but I take it with a grain of salt and start from scratch.

If you have one thing affecting another that typically means something is in series instead of parallel. That being the case mis-connections of a neutral would be a good place to start and easy to spot.

So it's methodology, instead of having knowledge of a bad neutral causing problems like this on a single circuit. That's fine, I was just wondering.
 

Jim M426

Member
Location
SW Florida
I haven't been able to do any troubleshooting but while I was installing the two 4" trims over the bar (just repainted the ceiling) I noticed that when I plugged them in they were unlit. As I installed them in the can (held by 3 spring tension clips) the lights lit. When I moved one a bit, the light went off. Same with the other. So it looks like the ground is completing a trickle circuit.

I could easily isolate the spring clips and be done with it but I want to know what's going on here. Problem is the wife has me on kitchen cabinet duty - I suggested we dump the old doors and drawer fronts and I'd make new ones. I got as far as removing the doors and building new cabinets for the oven, cooktop and frig when the alarm was sounded announcing the newly applied ceiling paint (living room, dinette and kitchen) was wrong. :weeping:

Anyway, at least I know it has something to do with the ground. The LED trims each have their own driver built in to the trim. Who knows what is going on inside that? Probably someone in China.
 

Jim M426

Member
Location
SW Florida
If grounding is not proper, voltage can appear between frame and neutral.
As I mentioned before, I come from the land of conduit where we had a metallic ground throughout. In this instance, there is no metallic conduit beyond the panel but there are some metallic boxes previously installed. Everything I have installed with this lighting circuit is plastic boxes, plastic conduit and NM cabling. Because of this, I was anal with the grounding.

When I replaced the panel, I installed a grounding bushing on the 2" EMT connector because I was using concentric KOs in the panel tub. From there I ran ground wire to ground bus that were screwed into the panel tub. I even scratched off the paint behind the ground bus. Probably overkill but from all I had seen here I wasn't taking any chances. What I don't know is the integrity of the 2" EMT as it leaves the panel and travels about 50' across the attic to a 200A disconnect breaker. There is no ground wire in that conduit. I did check the locknuts and they are tight.

But...

At the time I replaced the panel, I opened the 200A breaker in the disconnect. It didn't "snap" into the open position but rather slid to the open position. I thought the breaker was bad but I put a wiggy on the load side and both legs showed open. But because the hold was non-existent, I installed a lockout on it just in case.

What I'm saying is this is an old house and from everything I have seen there have been a lot of fly-by-night "electricians" who have left their mark. I may need to check the ground integrity all the way back to the ground rod. No copper plumbing feeding the house here.
 
Status
Not open for further replies.
Top