The hidden, extremely costly LED lighting maintenance

[Electric-Light]

Here's the thing. Oil capacitors and magnetic ballasts can handle temps. electronic ballasts can only dream of and they can tolerate getting coated in a thin layer of moisture as it typically happens when warm moist air hits the lighting components. There are a lot of components and solder joints on transistorized induction as well as LED ballasts so hard potting can place enormous stress on LED ballasts used for outdoor dust to dawn light.

Soft potting can pass enough moisture to cause insulation break down and fry out the ballast or worse wreck sensing circuitry and make the LED ballast spaz blink. LED ballasts are notorious for their inability to directly fit to 480v service or experiencing abnormal failure rates.

Incan, HIDs and fluoro continue producing a fair deal of light through the zero crossing to significantly reduce flicker compared to clear neon tube or LEDs without an electrolytic capacitor and this makes LEDs less than ideal for passive ballasting with a magnetic ballast.

Here's a very typical low power transistorized LED ballast design used for disposable lamps, "ballast bypass" TLEDs and many lower lumen fixtures.
Circled on top is a transistor. Circled in bottom is Diodes Inc AL1696 LED ballast controller.

LED fail video from Detroit.
https://www.youtube.com/watch?v=CAOZxP-ufTY

Only light fixtures that use a combination of LED and transistorized LED ballasts are known to suffer this failure pattern. When there are vulnerability in design, a great number of them can develop symptoms in a short span, or you could have one develop a symptom due to degradation that leads to LED ballast failing into spaz blink. It's in the end user's best interest to specify ballasts protected against spaz blinking. Even if labor is not part of negotiated warranty, it's just a very good idea to make exception to labor costs due to spaz blinking both on emergency and non-emergency basis, so the LED vendor is held entirely responsible for the added cost, because only LEDs can suffer from this. Street lights are obviously higher powered, but not ridiculously high.

Without this requirement, a series of spaz blink related calls can easily wipe out all the savings and even end up costing far more than continuing to use or installing and maintaining new HIDs. What's the big deal? It would just be a filler page of useless paragraphs if they never experience spaz blink. It should also disallow routine activation of thermal throttling to cover up design flaws.

The cost to address can be quite a bit, but what's the big deal if you're confident enough this doesn't happen?

 

[Electric-Light]

OK, good to know. So in the DC-DC case, are the manufacturers just being lazy by using a PWM rate that is low enough to be visible to some?

Cheers, Wane

Many LED back lit monitors as well as LED lighting products are known to suffer from indirectly noticeable flicker that becomes noticeable the deeper you dim unlike CCFL based. They do so because it is cheaper to dim LED this way without significantly increasing ballast loss. LED dimming is often done by slamming the switch on and off rapidly and this produces a lot of interference. (Hint: go to youtube and listen to the difference between sine wave and square wave). Many battery powered devices and home electronics use around 400 Hz. Too low, it flickers, a bit too high, the ballast whine coincides with our hearing sensitivity, go even higher and you can eliminate the flicker but the ballast efficiency is lowered and can start spewing interference into some licensed radio spectrum.
PWM based 12v and 24v rope or tape lights usually do not pass FCC requirements for consumer use except for 60Hz AC flicker rich line frequency phase cut dimming. Same with chopper dimming modules that go on the load side of 12v/24v supply you can find on the internet.

 

[hbiss]

What I don't get is why the failure mode in all these things seems to be the same- flashing. The drivers are all of completely different designs so I would expect different things to happen when they fail, like not working, maybe low light, even burning up. :?

-Hal

 

[tom baker]

I got a message from the moderator, seems we are running short on server storage due to this long post...

 

[gadfly56]

I got a message from the moderator, seems we are running short on server storage due to this long post...

Noooo, plenty of room. Based on a certain other thread, we have at least 6,700 to go.

 

[Besoeker]

Many LED back lit monitors as well as LED lighting products are known to suffer from indirectly noticeable flicker that becomes noticeable the deeper you dim unlike CCFL based. They do so because it is cheaper to dim LED this way without significantly increasing ballast loss. LED dimming is often done by slamming the switch on and off rapidly and this produces a lot of interference. (Hint: go to youtube and listen to the difference between sine wave and square wave).

Now how exactly did we get into dimming LEDs?

Many battery powered devices and home electronics use around 400 Hz. Too low, it flickers, a bit too high, the ballast whine coincides with our hearing sensitivity, go even higher and you can eliminate the flicker but the ballast efficiency is lowered and can start spewing interference into some licensed radio spectrum.

Well, 440 Hz is nicely within the human audible spectrum. It is about the same frequency as the G below middle A - which is 440 Hz.
So I'm not sure what point you you are trying to make.
Check out Bach's air on a G string. And, please don't infer anything rude. It is a lovely piece of music.

 

[mivey]

Now how exactly did we get into dimming LEDs?.

Not end-user dimming but rather the behind-the-scenes dimming that takes place without end user knowledge and that can impact performance. Mentioned before but I don't feel like finding the posts.

 

[Fulthrotl]

Noooo, plenty of room. Based on a certain other thread, we have at least 6,700 to go.

sigh. you aren't helping here.

 

[Besoeker]

Not end-user dimming but rather the behind-the-scenes dimming that takes place without end user knowledge

Do you know, or have any information on how that dimming is implemented and what it is intended to achieve?

 

[GoldDigger]

Do you know, or have any information on how that dimming is implemented and what it is intended to achieve?

One implementation is output maintenance dimming that starts out dimmed to just deliver rated output and increases input over time as the emitters age.

Sent from my XT1585 using Tapatalk

 

[Fulthrotl]

One implementation is output maintenance dimming that starts out dimmed to just deliver rated output and increases input over time as the emitters age.

Sent from my XT1585 using Tapatalk

yep. there's a distinction made between lumen maintenance (TTL)
fixtures with regards to daylight harvesting.

 

[Besoeker]

One implementation is output maintenance dimming that starts out dimmed to just deliver rated output and increases input over time as the emitters age.

Semiconductors age?
And how is dimming implemented?

 

[GoldDigger]

Semiconductors age?
And how is dimming implemented?

When current runs through a diode and heat is generated, the doping materials in the semiconductor can slowly diffuse across the junction, decreasing its effectiveness. There are other aging factors, all pretty much related to thermal stress during use rather than chronological age.

No [authoritative] idea. Probably current control in the driver?

 

[Besoeker]

When current runs through a diode and heat is generated, the doping materials in the semiconductor can slowly diffuse across the junction, decreasing its effectiveness.

Not a phenomenon I have encountered. I'm not saying it doesn't happen.
My background, as you probably know, is variable speed drives and high current rectifiers and has been for around 50 years.
Yes, heat is generated and is a big element of the design process is in calculating heat dissipation and the required cooling. Be it convection, forced air flow, or liquid cooling.
We have VSDs in service that have been operating 24/7 for decades with no semiconductor failures. And these are with major clients like the manufacturers of Kleenex, oil refineries, water companies etc.

The cost of lost production in any of these continuous process industries would make your eyes water. Degradation may happen, but in my experience, if it happens, it is a very slow process. And, as I have mentioned previously, we have used LEDs for control panel and PCB indicators since Adam was a boy. Zero failures have been reported.
Zero.

 

[GoldDigger]

See https://www.digikey.com/en/articles...g-the-cause-of-fading-in-high-brightness-leds
For information. The mechanism is actually growing crystal dislocations.
You do not see this in power semiconductors because the materials used in LEDs are completely different from silicon and germanium, and dislocations do not reduce performance in the same way. You do not see it in indicator LEDs where luminous efficiency is not a big issue and temperatures are lower.

Sent from my XT1585 using Tapatalk

 

[mivey]

Not a phenomenon I have encountered. I'm not saying it doesn't happen.
My background, as you probably know, is variable speed drives and high current rectifiers and has been for around 50 years.
Yes, heat is generated and is a big element of the design process is in calculating heat dissipation and the required cooling. Be it convection, forced air flow, or liquid cooling.
We have VSDs in service that have been operating 24/7 for decades with no semiconductor failures. And these are with major clients like the manufacturers of Kleenex, oil refineries, water companies etc.

The cost of lost production in any of these continuous process industries would make your eyes water. Degradation may happen, but in my experience, if it happens, it is a very slow process. And, as I have mentioned previously, we have used LEDs for control panel and PCB indicators since Adam was a boy. Zero failures have been reported.
Zero.

Something you have yet come to terms with is that the use of LEDs for panel indicators is different than the use in area lighting. This was covered more than once but you did not follow.

Apparently, and I'm no expert on scene here, the circuitry and the stress experienced with area lighting apps is much more complex than indicator apps. You may very well have indicator lights that last practically forever but they are not being driven to a performance edge like with some area lighting. Different stress levels.

You also mentioned the LED lights in your home lasting a long time. EL asked about specs and you esentially said the only thing that mattered was that they had lasted a long time. But the specs do matter if you are on a performance edge like with commercial lighting. You can't just say it lasts. Kind of like comparing a race car with a sedan and saying the sedan is better because it endures longer.

In competing light designs performance and cost both come into play. If you get long life but the performance falls below specs then it matters. Initial and long term performance matter.

With indicator lights who cares if they dim a little or the color is slightly off? But if I spend extra to get extra performance over the fixture life and only get that performance at the start I should be asking questions. If I spend more on one technology to meet performance requirements that can be matched or exceeded in different tech at a cheaper price I should be asking questions.

You must make an apples to apples comparison and you are not doing that. I'm sure you get those lazy comparisons with your most excellent VFD designs compared to cheaper competition and it probably drives you nuts.

If you want to make a fair comparison on LED lighting, quit comparing your indicating LEDs to high performance LED area lighting.

 

[Besoeker]

Something you have yet come to terms with is that the use of LEDs for panel indicators is different than the use in area lighting. This was covered more than once but you did not follow.

All the lights in our house are 230V LEDs.

 

[mivey]

All the lights in our house are 230V LEDs.

You have mentioned that before as well and I will repeat with perhaps better clarity: the performance of some non-descript lights in your house are no comparison to an application where high performance is expected and time and money really matter.

In my house, I want a long life at a reasonable price (reasonable being a pretty broad term here). If I spend twice as much as I should because I don't research the specs and performance then no biggie.

Not the same as a performance based commercial lighting design involving more than just me buying a box of whatever and installing with no one else involved and no real consequence if time and money is wasted. I am mostly happy with the LEDs in my house even though there are some performance issues I solved by simply buying different brands and trying them out.

That is different from how a commercial install should go. You have many times referenced your careful VFD designs made to meet requirements at work. Why do you think lighting design should be carefree and void of careful planning?

If I really want to evaluate my choice of mostly random boxed stuff I put in my house, then I have to dig into some details about cost and performance. It would be more than just: "I just stuck some stuff I bought in at my house and it still works so it must be the best choice for high performance commercial designs as well". Please stop that kind of nonsense comparison.

 

[Besoeker]

You have mentioned that before as well and I will repeat with perhaps better clarity: the performance of some non-descript lights in your house are no comparison to an application where high performance is expected and time and money really matter.

In my house, I want a long life at a reasonable price (reasonable being a pretty broad term here). If I spend twice as much as I should because I don't research the specs and performance then no biggie.

Not the same as a performance based commercial lighting design involving more than just me buying a box of whatever and installing with no one else involved and no real consequence if time and money is wasted. I am mostly happy with the LEDs in my house even though there are some performance issues I solved by simply buying different brands and trying them out.

That is different from how a commercial install should go. You have many times referenced your careful VFD designs made to meet requirements at work. Why do you think lighting design should be carefree and void of careful planning?

If I really want to evaluate my choice of mostly random boxed stuff I put in my house, then I have to dig into some details about cost and performance. It would be more than just: "I just stuck some stuff I bought in at my house and it still works so it must be the best choice for high performance commercial designs as well". Please stop that kind of nonsense comparison.

The topic is about the hidden, extremely costly LED lighting maintenance.
What exactly has to be maintained?
What are the costs compared to HID, LP sodium etc?
What is the life expectancy of LEDs compared to other technologies for, say, street lighting?

 

[Besoeker]

That is different from how a commercial install should go. You have many times referenced your careful VFD designs made to meet requirements at work. Why do you think lighting design should be carefree and void of careful planning?

I don't think that at all. Why would I?
I would like to see some figures to back up the OP's "hidden extremely costly maintenance" claim.