An Inherent Flaw in LED lighting-- Why all LED fade, and what you can do.

GARRIGA

2500 Club Member
View Badges
Joined
Oct 12, 2021
Messages
3,692
Reaction score
2,952
Location
South Florida
Rating - 0%
0   0   0
Buying more than you need and using less then it produces always applicable when it comes to prolonging the life of most things.
 
Last edited:

luxdium

Community Member
View Badges
Joined
Apr 17, 2022
Messages
87
Reaction score
102
Location
SF Bay Area, California, United States
Rating - 0%
0   0   0
Until the new gen light comes out and they tell you they no longer supply parts for yours.
SuckerGIF.gif

Our pucks are actually designed specifically for the previous-gen lights and can support fixtures going as far back as 2014. We've been beta testing them since mid-2020 and where possible, have tried to maintain compatibility with a broad swath of lights.
 
Last edited:

RedoubtReef

Valuable Member
View Badges
Joined
Jul 4, 2016
Messages
1,179
Reaction score
1,253
Location
The Redoubt
Rating - 0%
0   0   0
Our pucks are actually designed specifically for the previous-gen lights and can support fixtures going as far back as 2014. We've been beta testing them since mid-2020 and where possible, have tried to maintain compatibility with a broad swath of lights.
Your website only shows them for AI. Is that correct?
 

radfly

Active Member
View Badges
Joined
Apr 29, 2022
Messages
222
Reaction score
158
Location
Charleston
Rating - 0%
0   0   0
Long post but, here we go. I'll start at the basics.

In layman's terms, essentially when a LED is created, there's a substrate that is laid down, usually a crystal lattice. This substrate sits between your LED and your wafer from the chip. This substrate sits on top of that wafer and unfortunately the process of LED Wafer chip manufacturing is pretty complex, and ultimately the substrate almost always develops micro cracks to some extent as a result of this, or most likely, micro cracks in the die used for making the semiconductor.

These cracks are called Threading Dislocations. Ultimately, they get bigger over time and multiply when the chip on the LED itself is exposed to high temperatures.

This ultimately means that in layman's terms, for a LED to work, you take a board, you coat it with a net negative impurity, then a net positive impurity over that, then when a charge is applied, the negative electrons travel between the two layers to fill a hole in the positive layer. This process is supposed to emits a photon, if all the electrons have a pathway to meet a proton partner.


If they don't find this partner however, (let's call this electron, "Barnabus"
) It's because instead of Barnabus having to travel a mile to their friend with benefits, Carol's house, now the city installed a huge tower apartment complex across the street with lots of hot beautiful women. Because Barnabus is actually a lazy vampire and takes the path of least resistance, and they view love as a binary event, a reason to feed, They don't actually care about Carol, they just want the closest food possible. So they go across the street.

Because Barnabus goes across the street now, he has no reason to walk back down that road and meet up with Carol, and Carol no longer has any reason to keep the outside lights on anymore.

When they do this, eventually everyone else in the area gets the same idea over time and more and more of these buildings open up across from eachother.

As this happens, these people in the country begin to develop mental health problems they didn't have before when the previous system was working. Other parts of their life may begin to crumble or fall apart when they begin to compare themselves to what they too, could have..... (The chip begins to lose it's ability to do the job properly and begins creating a lower threshold for a breakdown voltage event to occur, eventually conducting current in reverse.) Once this reaches about 5V, most manufacturers of LED will short circuit and or failsafe. This is virtually never going to happen, but this is why your LED dim over time, and because this threshold is lowered, it means it's more likely that additional reverse bias from a voltage spike or ESD would cause a complete premature failure.


That said, basically after about 3000 hours of burn in on most led they won't get any dimmer. There's not really an efficient way to test that either however, it looks like the average usable life sites somewhere around 35-50 k hours and your typical CFL is 8-10k, and metal halide is 7.5k-20k.

So at most, led change every 4 cfl replacements, and at least, led change every 6 cfl replacements or 7 mh replacements. To maintain peak. But this assumes you run 100% capacity. Most of us do not. So I think over time, and some might already be experiencing this, there is a meaningful shift in PAR even on nicer lights, if in consistent use. But with a value at this level, and 12 hours a day of use, this could still be 3-4 years you are looking at before it becomes a problem, in theory. So the key is to purchase the biggest light possible, run it at the lowest power to reduce stress and maximize your cooling ability relative to the entire fixture, and you'll get a lot more use out of them because there's less energy turning into heat overall.

Some sources :







Thanks for coming to my LED Ted talk

Tl;Dr

If you want to save a lot of money on this hobby, replace your LED on your fixture once a year and make sure your lights have enough heatsink or cooling so as to slow down microcracking aka thread dislocation process.

Running your lights at lower power and having more of them, can also extend the life of your LED.

If you opt not to replace your LED, increasing your current to the LED to compensate will accelerate the thread dislocation process, but if you plan on buying new lights in the next 1-2 years, it's not worth getting upset over.

If you replace your LED, you can ultimately control your spectrum as well if you wish to make some changes to it.

If you are electrically inclined, making sure your light's fans are always on either by rewiring or by selecting the option is always a good decision, as is adding a heatsink such as aluminum and doing things to improve the airflow around your tank. This will not only help your LED, but ultimately most tanks are oxygen deficient at some level, especially if you aren't using a skimmer, or you have the tanks in a room that doesn't get open windows, etc.

I'm not suggesting you liquid cool your lights, as that too, will likely cause dimming. But be aware that the likely issue with your lights dimming resides exclusively within the individual LED wafers, assuming you don't let heat become such an issue that it damages your actual board.


So yep, that's what I spent the last day reading about on depth.


In a weird twisted way, a dimmer light is a better light, and the best designed light is going to be designed to limit things it can actually mitigate, like output and heat.

I'm not sure that's a $800++ price tag, but you as a customer need to do your due diligence here and read the articles above to come to your own conclusions and begin asking questions to manufacturers that will
A: Improve the hobby

and

B: Give you a reliable product that matches your expectations as a consumer.



You know, the problem with LED is the technology moves so quickly, most of these degradation arguments are pointless for everyone except people like us who seek extreme stability long term for something like light output.

I can say with confidence that whether you buy the most expensive LED or the cheapest, there is little to no way to do quality control manually at the level that is required to determine each LED without the use of Machine Learning, AI centric, Camera-Based algorithms that are designed to identify them based off of regression models.

I will tell you with confidence that virtually nobody will do this because the length of time for these to fail is such that most protocols replace them before it becomes a relevant problem that could cause life or death or massive loss situations.

Ask anyone here growing coral and or running a website how often they replace their lights? Their pumps? Heaters?


I'm not telling you to go buy GHL or go buy Radion, or buy a black box. I'm telling you the product that you all want doesn't exist because there's a inherent requirement to replace LED regardless of manufacturer because of the above design flaws present in all LED.

So, if you really want to keep customers as a lighting manufacturer, it might make more sense to design the LED to be as easily replaceable as possible, and sell those so that if a customer thinks their blues aren't all there, and you as a manufacturer can sell them the new kit, then great.


But don't go around changing key spectrum or coverage when people already trust your product and enjoy it.

Personally I think maximizing your surface area for heatsink is just about the smartest thing you could do to help mitigate dimming.

Light manufacturers with good fans or huge heat sinks, or both? Major props to you for thinking of longevity for the customer instead of putting a spotlight on a metal plate of circuits with a small cooling fan that is ready to heat up in a hurry or maybe even too hot to touch sometimes.

Anyway, I figured this community deserves to have this information presented to them and I feel I can make a much more informed decision about LED and LEd lighting deaign as a whole after reading the above articles.

Maybe some other people heavily involved with the process of designing LED can chime in here, but I just want people to be aware that there's a difference between radion and say, Maxspect, but the most important thing to look at isn't necessarily something you can test by turning it on or putting in a spectrometer. I the case of LED, assume all will eventually fade and die if given enough time. So look for ways the manufacturer has found solutions to successfully mitigate the problems that exacerbate that process, so you get as much life out of your lights as possible, regardless of what you paid for them, and regardless of what you have right now. You can take steps to extend your LED life, but as I went to great lengths above, I hope now you will be able to identify them.


What a beautiful world to live in and have the extremely privileged opportunity this hobby affords us.


Peace and love,

Kyle
LED’s will likely last the lifetime of your reef hobby.
 
OP
OP
AKG

AKG

Active Member
View Badges
Joined
May 1, 2022
Messages
255
Reaction score
308
Location
Longmont
Rating - 0%
0   0   0
I'll humor you though. Lots of people leave the hobby within a year or two.
LED’s will likely last the lifetime of your reef hobby.
 

Attachments

  • Screenshot_20221024-235455.png
    Screenshot_20221024-235455.png
    190.3 KB · Views: 107

HomebroodExotics

Well-Known Member
View Badges
Joined
Nov 30, 2020
Messages
932
Reaction score
1,070
Location
United States
Rating - 0%
0   0   0
That’s cool and all but telling people to replace their leds every 4 years is kind of ridiculous in my opinion. Can you provide some proof that there’s a significant spectrum shift and does it even matter?

And if anyone needs to toss out their 4 year old leds hit me up. I’ll even pay for the shipping to send it to me. I’m a nice guy like that.
 

oreo54

5000 Club Member
View Badges
Joined
Sep 18, 2017
Messages
5,970
Reaction score
3,673
Rating - 0%
0   0   0
I'll humor you though. Lots of people leave the hobby within a year or two.
Think that was their point.

Not sure the methodology used but it lines up pretty well with the numbers I heard back in the day from one of the reps who frequented the LFS I worked at.

I remember him telling me 50% gone by the first year and 85% gone by year 3

 

RedoubtReef

Valuable Member
View Badges
Joined
Jul 4, 2016
Messages
1,179
Reaction score
1,253
Location
The Redoubt
Rating - 0%
0   0   0
Yeah, LUX Engine was designed for AI lights. Expansion to others is a possibility in the distant future.
Well, that's not exactly a 'broad swath' of lights but, it's a start. I think what you are trying to do is great. Will you, or have you, run into any issues with the manufacturers in regards to duplicating parts for their proprietary products and do they maintain functionality with the software? If you are successful, you are cutting into their profit margin.
 

luxdium

Community Member
View Badges
Joined
Apr 17, 2022
Messages
87
Reaction score
102
Location
SF Bay Area, California, United States
Rating - 0%
0   0   0
Well, that's not exactly a 'broad swath' of lights but, it's a start. I think what you are trying to do is great. Will you, or have you, run into any issues with the manufacturers in regards to duplicating parts for their proprietary products and do they maintain functionality with the software? If you are successful, you are cutting into their profit margin.
Broad, with regard to AI, and that's what I was referring to when I mentioned a "broad swath of lights," as presently, we make for those. They work with Prime (non-HD), Hydra 26 (non-HD), Hydra 52 (non-HD), Prime HD, Hydra 26HD, and Hydra 52HD. So that's 6 lights, some of which are almost 8 years old.

Our goal is to make spectral quality accessible. For the fraction of the cost of a new light, you can upgrade what you have and take it to levels of output that surpass the newest flagships, and even the historical benchmarks.

By focussing on the clusters, we pack custom-binned chips made by some of the leading chip manufacturers, building on the foundation of their reliability and longevity. Since we also make the thermal substrate they go on which almost doubles the heat dissipation, it's not a mere duplication by any means.

Yes, there are/were challenges. Even in the same model generation, there are internal hardware revisions, which may affect performance. So an extensive beta program spanning 8 countries had to be conducted to gather insights on all design variations and the clusters were designed accordingly to preserve compatibility with existing hardware, and existing control interfaces like myAI, AI Cloud, and IoTa.

Testing them before shipping also gets interesting. Since LEDs have a degree of variability and we strive for consistency, every cluster we make is tested for 24 hours at full blast in the representative fixtures they upgrade. Each cluster gets PAR and spectrum tested and we keep track of them with unique serials, so you can rest assured that if you were to buy clusters months apart, their colors would still match. We go through these extensive processes because it's no fun when you have lights with colors that don't match.


SPD.png
Test Array.jpg
Variability.png
 

Treefer32

Valuable Member
View Badges
Joined
Apr 1, 2013
Messages
1,466
Reaction score
1,033
Location
Fargo, ND
Rating - 0%
0   0   0
I was just wondering this today from a budget planning perspective, I have 3 AI Hydra 52 LEDs as well as an Aquatic life T5 4 bulb, 80 watts per bulb, fixture that I replace the T5 bulbs on annually (religiously). However, if I have to replace 3 LED fixtures annually I'd have to sell out of the hobby now. I couldn't afford $2k on lights alone a year. That said, I'm starting year 5 that my LEDs have been operational. Other than an occasional cleaning, I haven't messed with them since setting them up.

I was hoping to run my LEDs another 5 years? Fans are on most of the time, I have 2 hour ramp up and 2 hour ramp down and around 6 hours at 90%.

Should one plan to replace after 10 years of use?
 

Nano sapiens

Valuable Member
View Badges
Joined
Apr 25, 2010
Messages
2,496
Reaction score
3,693
Location
East Bay, CA
Rating - 0%
0   0   0
I was just wondering this today from a budget planning perspective, I have 3 AI Hydra 52 LEDs as well as an Aquatic life T5 4 bulb, 80 watts per bulb, fixture that I replace the T5 bulbs on annually (religiously). However, if I have to replace 3 LED fixtures annually I'd have to sell out of the hobby now. I couldn't afford $2k on lights alone a year. That said, I'm starting year 5 that my LEDs have been operational. Other than an occasional cleaning, I haven't messed with them since setting them up.

I was hoping to run my LEDs another 5 years? Fans are on most of the time, I have 2 hour ramp up and 2 hour ramp down and around 6 hours at 90%.

Should one plan to replace after 10 years of use?

With quality arrays (especially those that are running presumably with less intensity due to T5s being present), I wouldn't expect much drop off in emitter intensity over many, many years of use.

Arguably, the #1 killer of the emitters in these arrays is when the fans fail, and the aquarist doesn't catch the issue in time before the emitter's junction temperature exceeds critical thresholds. That's why I built my old DIY array with a proportionately large heatsink so that no fans would be required and some high quality commercial units employ the same strategy today.
 

ca1ore

10K Club member
View Badges
Joined
Oct 28, 2014
Messages
14,091
Reaction score
20,005
Location
Stamford, CT
Rating - 0%
0   0   0
When did the Radion G4 come out ….. early 2017? OK, so been running them for six ish years. I don’t doubt they’ve lost some oomph, but corals still grow fine. I’ll hate the day when they start to go …. I’m a Mobius free reefer and plan to stay that way.
 

69cam

Community Member
View Badges
Joined
Nov 20, 2011
Messages
46
Reaction score
15
Location
Tracy ca
Rating - 0%
0   0   0
I was just wondering this today from a budget planning perspective, I have 3 AI Hydra 52 LEDs as well as an Aquatic life T5 4 bulb, 80 watts per bulb, fixture that I replace the T5 bulbs on annually (religiously). However, if I have to replace 3 LED fixtures annually I'd have to sell out of the hobby now. I couldn't afford $2k on lights alone a year. That said, I'm starting year 5 that my LEDs have been operational. Other than an occasional cleaning, I haven't messed with them since setting them up.

I was hoping to run my LEDs another 5 years? Fans are on most of the time, I have 2 hour ramp up and 2 hour ramp down and around 6 hours at 90%.

Should one plan to replace after 10 years of use?
I am on year 7 with my 26hd's and was wondering the same thing...
 

92Miata

Valuable Member
View Badges
Joined
Feb 26, 2020
Messages
1,523
Reaction score
2,490
Location
Richmond, VA
Rating - 0%
0   0   0
The general gist here - that LEDs wear out - is true - but all the math here is based on a number of misunderstandings and poor assumptions.

Firstly, the 35-50k hour lifespan isn't the amount of usage until the led fails - it's the L90 lifetime - or the amount of time before the emitter drops to 90% of it's original output. For Cree Xp-E LEDs, this is 50k hours.

50k hours, at 12 hours a day, is 11 years, and that's not even getting into the fact that most people aren't running them at full. 10% degradation at 10+ years is way out into the "it doesn't matter" range and probably similar is seen in MH and LED due to corrosion on the reflectors.

I'm running gen3 Radions on most of my tanks. They're about 8 years old. They're capable of producing way more par than I can use, and the spectrum is fine.

The vast majority of reef LED fixture failures are due to failures in the drivers, control boards, and cooling. In a properly designed fixture, LED lifespan really isn't a worry.
 

BeanAnimal

5000 Club Member
View Badges
Joined
Jul 16, 2009
Messages
5,071
Reaction score
8,108
Rating - 0%
0   0   0
So we have established that LEDs wear out over time. Should I just go back to fluorescent and Metal halides? Are there any other considerations that should be taken into account?
Everything wears out over time... even the universe (entropy)

You will easily get 5-10 years out of your LEDs. Will output drop a bit? Sure. Does it matter that much? Not likely and after 5 years you will still easily be at 90% or more. Chances are you will be either out of the hobby or looking at the next generation of lighting by that time anyway.
 

HAVE YOU EVER KEPT A RARE/UNCOMMON FISH, CORAL, OR INVERT? SHOW IT OFF IN THE THREAD!

  • Yes!

    Votes: 32 45.7%
  • Not yet, but I have one that I want to buy in mind!

    Votes: 9 12.9%
  • No.

    Votes: 26 37.1%
  • Other (please explain).

    Votes: 3 4.3%
Back
Top