L.E.D. vs HALIDE/T5/VHO
 

How many watts per LED bulb is needed to grow SPS coral? Our benchmark for MH or T5 was wattage, how does this compare to LED? Is PAR the only reference available? If someone said they were runnung a 400W 14K MH bulb, you would know it would be awesome for SPS, this doesn't seem to exist in LED. All I hear is brand (mostly Cree) number of bulbs, and number of watts per bulb.

While LEDs are looking more promising these days my main issue with them is the complexity and reliability. Not too much information out there regarding long term experience and performance. The only fixture that had been out for a while was the solarus but everyone I've heard of using one has had some kind of issue with them relating to either the drivers or LED arrays. In both cases servicing them isn't exactly easy. Other fixtures may be better but their popularity and presence isn't high enough to provide much experience. They also only come with a 1 year warranty which is complete garbage for a fixture that is suppose to last 10 years. Others may disagree but as an engineer in a manufacturing company I know exactly how warranty life is determined and limiting warranty to one year for a such a product is pretty big red flag.

I prefer to spend money on a quality fixture that not only looks good but works all the time. I can't justify spending more up front for a fixture that doesn't look as good, may require constant attention and maintenance and might not even last as long as suggested. Despite the potential energy savings the chance of spending less in the end isn't exactly that good which defeats the purpose IMO.

For now I think the LEDs belong to the DIY electronic guys who can easily fix there own work or build it better with quality parts and those with nano and smaller tanks where the upfront cost is less.

It's not so much a watts per gallon type thing with LEDs as it more depends on how you set up the array. how many LEDs, what optics, how are they spaced. You can set them up to have high or low PAR depending on what you do and your budget.


The Solaris fixtures were really poorly built. They had flawed power supplies and no heatsinks so LEDs overheated and burned out. They really weren't designed well. There are some good systems such as the AI ones but also many others that make compromises that reduce output etc.

It's really not hard to design a decent LED system that will be properly cooled and durable. But the costs are still a bit high so if you want a decent LED fixture you have t pay for it. There's no free lunch. Longevity should not be an issue if it is designed well as LEDs are well understood and their lifespan and degradation properties are well known. LEDs have been around since the early 1960s. Ones with enough power for more intense lighting started to appear about 9 or 10 years ago. The technology still has room to grow in terms of output and efficiency etc. so it will only get better.

Yes I would agree the solaris fixtures where not built that great but there are still very few other fixtures that promise better results. The AI are the better looking fixtures as far as quality goes but to me the module design lacks the style I prefer to see in a fixture for open top tanks and the warranty still suggests a lack of faith in the reliability of the product.

As for LED reliability and lifespan I would strongly disagree. Yes they have been around for a long time but they are new to the hobby and the environment they will be exposed to. They will run hotter and in high humidity and also be exposed to the salty corrosive atmosphere. This will impact the lifespan of the LEDs just like it does for all other types of lights we use. Metal Halides for example are rated for 5 years before a 30% drop in standard applications but over time we have discovered that they only last one year in the hobby application. A similar result will no doubt be discovered with LEDs over time. A friend of mine also pointed out the other day that a lot of cars these days run LED tail lights as manufactures made similar promises about long life and reliability but you still often see multiple burnt out LEDs in taillights on still fairly new vehicles. At least if a regular tail light burns out it's a simple and cheap fix while the LED style simply is not.

An LED by itself is no doubt a reliable light source, however combining multiple arrays into hundreds of LEDs greatly increases the chance of failure and the fix isn't the same as simply changing a bulb.

Its true that MH are rated at 5 years before a 30% drop in standard application, that is if the standard application was not to care about the total PAR values which is often what MH bulbs are used for (parking lots and etc.). The requirement for high PAR alone is what determines how long a MH bulb can be used for in the SW hobby and not so much the environment its been run at. The humidity and salty environment would effect the longevity by a very small amount as the intensity/color shift happens inside the bulb and not outside and a MH bulb is enclosed.

The same would apply for the LEDs except with LEDs heat becomes a great cause of failure on the emitters, most high power LEDs nowadays are rated to run at a constant temperature of 50C to 80C before damage to the emitters occur and most well designed fixtures with proper heatsinks and cooling run the emitters near 40C if not a tad higher. Just for reference the spotlights we carry run at roughly 48-50C without fans and thats 9 bulbs closely package together in a 6" heatsink.

With the car scenario, the LEDs are being put through conditions much more harsh than seen above our reef tanks, they have no heatsinks, no cooling, and have to put up with absorbing the shock and vibrations from the car moving, and I'll agree that we might not get the rated 50,000hrs our of the emitters but it'll be pretty close to it, for a couple simple reasons. LEDs through there lifespan doesnt have much color shift if any at all as theres no gases and no burning seen in the MH bulbs. The intensity will drop but if the emitters are used in the suggested range of specs by the manufacture it should drop to that % claimed by the manufactures.

Eugene is correct. Unless you are running them right above your water without a splash shield then salt creep and corrosion should not be a big problem. There was one guy in the candlepower forums IIRC that posted a picture of an emitter he had been running over a salt water tank for a long time and it was covered in salt and corrosion but it still lit up.

Heat is really the enemy of LEDs. Like any lighting system it all depends on how it's designed and put together. If it is designed well and has adequate cooling you can expect the advertised life out of the LEDs. If it is not well cooled you can expect the LEDs to degrade faster. But LEDs themselves have been used on myriad different types of equipment in all sorts of environments so I would still argue that their reliability is well understood.

As Eugene pointed out the LEDs are best run at less than 80 degrees. The lower the better. I haven't measured the temperature of my heatsink exactly but what I can do is put my finger in one of the holes on top of the fixture and touch the heatsink. After running for several hours it feels just barely warm to the touch so it can't be above 40 degrees. I may add a temperature sensor hooked up to the Arduino when I get to that point.

As for cars, which ones seem unreliable? I have heard of some issues on Cadillacs but Hondas and Audis seem pretty reliable from my understanding. Again, Eugene is correct that the stresses on automotive components are far higher with greater temperature shifts and shocks and bumps and vibrations etc. Not generally a problem for aquariums.

So I think your criticism of LEDs could apply to T5s or MHs as well as there are good reliable systems and poorly built and unreliable systems. Doesn't matter what the technology is. Part of the problem you mention about difficulty in repairs is that in order to save on manufacturing costs the LED emitters are set up on modules in many systems and if one LED goes down the whole module has to be replaced as it is usually not simple to replace a single emitter once its on the PCB. I do give kudos to Maxspect as each of their LEDs is on a discrete star PCB and can be replaced individually if need be. Not so with the AI or several others where you have 3 to 6 emitters on one PCB. So again, depending on the design of the system, the fix could be as simple as replacing one emitter or having to replace an entire module. But replacing a module should not be more difficult or more onerous then replacing a ballast in a T5 fixture.

Kind of my point guys. LEDs are rated for 10 years before a 30% drop in intensity occurs just like halides are rated for 5 years. So yeah halides only last a year before par is too low so how long will the LEDs last? It would seem obvious and fair to say they won't last the full tens years before requiring replacement.

Also yes all lights have failures which was also my point but the bulbs can easily be replaced while LEDs cannot be easily replaced. Based on the design of the fixtures I've seen it appears that once the LEDs fail or become too worn the fixture is basically trash unless you have the skills to fix it which in my mind isn't fair to assume we all do. In addition the shear number of LEDs is much higher than other alternatives so the chance of failure is greater.

Like I said looking more promising these days but it's too early to say too much about the reliability and lifespan of LED fixtures which is why I wouldn't recommend the option to most people. Enough said.

The difference between LEDs and Halides and T5s etc. is that LEDs spectrum does not shift as they degrade. They simply lose output. With MH and florescent lights the spectrum shifts to undesirable wavelengths which compounds the loss of PAR. LEDs should last longer because they will only lose some output but not shift in spectrum. So 6+ years is a reasonable estimate for LEDs.

The added difficulty of changing LEDs will only be an issue if their lifespan is far less than what is estimated and I guess that's the crux of the argument. You either think they will last or you believe they won't. I suppose only time will tell.

I'd agree that six years would be a reasonable estimate at this time. That being said with the current design of LED fixtures the energy savings will likely never give a payback if you have to essentially replace the fixture every 6 years. That's without my concerns of premature failure which seems fairly likely with what I've seen with current fixtures on the market and the lack of a decent warranty period on newer fixtures.

On the topic of par vs spectrum I don't believe you're correct, just because the spectrum of the bulb changes doesn't mean the par will also change. Certain halide bulb and ballast combinations will last far past the year mark in terms of par while spectrum is a separate issue.
http://reefkeeping.com/issues/2007-04/ac/index.php

I see LED taking more of a role in supplemental lighting in combination with halides or some other type of HID lighting of the future.

PAR decreases as spectrum shift due to a few factors. PAR is a measure of Photosynthetically Active Radiation as bulbs shift in color they generally move either towards the right of the spectrum graph or the left, most bulbs we know move towards the right as can be seen in high kelvin bulbs as they go from blue to white-ish. As this happens PAR values drop as the spectrum is now shifting and you get less of both red/blue spectrum that most corals utilize. The point is even if the PAR is still high enough the corals no longer get there needed spectrum so there colors become dull or brown and none of us like that:lol:

Bulbs with a lower K rating typically produce higher PAR because, as you said, it's related to the amount of useful radiation (more of it in lower spectrum bulbs). So if PAR and spectrum wear where directly tied together, as you stated, then with certain bulbs you would see an increase in PAR as the spectrum shifts in the lower K direction. We of course know this isn't the case and the article I linked suggests that the change in par and spectrum overtime are independent not dependent. The decrease in PAR is related more to the decrease in intensity not the shift in spectrum. Of course certain bulbs preform differently so some may be more effected by the spectrum shift but avoiding such bulbs is pretty easy these days with all the information we now have.

So just because an LED doesn't suffer the same spectrum shift as other bulbs doesn't mean the PAR rating will be unaffected, it simply means you won't notice a change in color overtime, only intensity.

your forgetting those spectrums that were already in the two extreme zones of red and blue, where do they go? Those gets pushed further right into the 800nm (reds) range that we cant see, nor can corals utilize so you lose a good portion of the spectrum that you would if the bulbs were new, so you see a decrease in PAR output. The decrease in PAR is not just about the intensity decrease, losing red spectrum and blue spectrum decrease par values quite a bit as well

The way spectrum works is, too far right, the spectrum isnt used, and same goes for to far let. if the bulbs we had originally were spiked at 450nm and spectrum shift it to 500nm in a year, all that blue light turns into green. Corals dont utilize green light very well and the and useful PAR goes down. Then everything in the 650nm-700nm range literally just disappears from the usefull radiation spectrum and moves into IR range.

It's actually more complicated then that. The spectral shift of bulbs actually makes a big difference in PAR. First have a look at this figure that shows the peak absorption wavelengths for chlorophyll:

http://upload.wikimedia.org/wikipedi...n_spectrum.gif

You can see that chlorophyll likes blue and red but not much in between. Now here is a good article on the spectral output of various MH lamps:

http://www.personal.psu.edu/sbj4/aqu...omparison.html

Here is the output of a 10,000K Ushio bulb as an example:

http://www.personal.psu.edu/sbj4/aqu...son/Image3.gif

So as that big blue peak not only decreases but shifts over the the right you get less real useable PAR since the green part of the spectrum is less efficient then the blue and red in terms of photosynthesis. The situation is probably even worse with 14,00K and 20,000L bulbs that have more blue and less red in their spectrum. Here's a 20,000K bulb:

http://www.personal.psu.edu/sbj4/aqu...son/Image4.gif

So yes, the lack of spectral shift in LEDs is a big advantage in that you stay in the peak output frequencies for photosynthesis. I imagine the T5s etc. put out similar spectrums as some of the MH lamps so the situation is similar.

Eugene beat me to it but I had diagrams :razz:

Nice summary with good data and purty pictures. LED FTW

:biggrin: your diagram helped explain it better then what i can with words.

While I applaud your skills of posting diagrams exactly how is posting diagrams of bulbs running with different ballasts relevant to bulb life? I also obtain the skills to post diagrams from articles however I will post ones more relevant to the discussion.

First we can look at the same bulb at 0 months and 17 months:
http://www.advancedaquarist.com/2008...variant/medium
While the AVERAGE spectrum or CCT will shift in one direction the actual plot does not actually shift but rather decrease in certain areas while increasing in others resulting in a new CCT. Typically the intensity of the blue will decrease but depending on the ballast you use the other areas of the spectrum will actually increase maintaining your total PAR but changing the overall color of the bulb.

Here we can see PAR readings of the same type of bulb over time
http://i11.photobucket.com/albums/a1...c/lampdata.jpg
Clearly shown is a steady par reading but a decrease in CCT

So we can pretty clearly conclude
So yes the CCT of halide bulbs will change from a result of a change in intensity of different wavelengths but this doesn't mean lower par. Results will vary for different bulbs but proper selection will prevent overly negative results.

The reason you don't see the change in CCT with LEDs is simple. The spectrum is much more defined and concentrated over a smaller area. The intensity simply decreases over time and since it lacks the other wavelengths there is no increase in other sections keeping CCT constant.
http://www.advancedaquarist.com/2006...um/Figure5.JPG

So once again the change in PAR is more directly connected to the change of intensity of each wavelength that makes up the Photosynthetically Active Radiation (PAR) and not necessarily the overall change of the Correlated Color Temperature (CCT).

Also there is more to PAR than just red and blue
http://www.advancedaquarist.com/2009/3/review
http://www.advancedaquarist.com/2008/7/aafeature1
http://www.advancedaquarist.com/2006/8/review2

The posting of a graph is not to show different different ballasts used but the spikes in individual ranges.

I'll agree with you that PARs values change according to the change in intensity of different wavelengths but how is that different with a shift in the spectrum? (might just be my way of reading it that i see them as the same). So would you say that the PAR value decrease is correlated to a decrease in intensity and color spectrum shift not intensity alone? now that you've mentioned that its % of intensity drop in certain spectrums.

Its true that theres more to red and blue in PAR but those are the 2 main spectrums that plants/corals utilize. There is a need for green/yellow but the 2 most utilized spetrums are red and blue thats why i only mentioned the 2.

This discussion is interesting!

I would say that PAR is obviously related to CCT, different bulbs produce different PAR based on this, but not so much for the change of PAR overtime from bulb wear. A decrease in PAR would be more a result from a decrease in intensity alone as changes in CCT could either increase or decrease PAR or even remain constant.

So I guess to sum up what I'm suggesting is a decrease in par is a result of an overall decrease in intensity while a change in CCT is a result in a shift or varying fluctuation in intensity over different wavelengths. CCT and PAR are not directly related when we're looking at bulb wear.

So when comparing LEDs to halides, LEDs will still suffer from a decrease in intensity and therefore PAR but the rate of reduction of intensity in LEDs is less than halides. The narrow spectrum in LEDs makes the change in CCT is basically unnoticeable while halides suffer greater changes due to the more broad spectrum. So LEDs obviously last longer but halides are cheaper and easier to replace and by selecting bulbs with a more concentrated spectrum in the right areas you can extend you bulb life dramatically.

Eugene is correct. I showed those charts to show how the bulbs we choose for reef applications spike in the same wavelengths as the most efficient ones for coral photosynthesis.

PAR meters measure all light between 400nm and 700nm. However, 500 to 600nm light does not produce as much photosynthesis as light around 450nm or 650nm to 700 nm. So while overall PAR may seem to stay the same or drop somewhat overall PUR (Photosynthetically Usable Radiation as shown in some of the articles you linked to) may drop more than PAR since the decrease in blue is probably more than the decrease in the green to red part of the spectrum. So we should probably look more at PUR than PAR but few people have spectrophotometers that can do the more complex analysis. What is intersting is that if you measure PAR of a cool white LED and compare to a blue LED, they put out very similar PAR numbers. I would be really interested to see what a PAR meter measures for a green LED of similar wattage. I suspect that PAR would be very close to the others but that's not PUR and corals would not do well under pure green LEDs.

MH and florescent bulbs degrade relatively quickly (within 1 to possibly 2 years with the better bulbs) such that PUR decreases as the overall spectral output changes (or shifts depending on your frame of reference) to be more warm. It is fairy well accepted that nuisance algae grow better at warmer colour temperatures and old bulbs can promote their growth, as I have seen first hand with my old T5 bulbs. If things did not change in that way why change bulbs so often? So to summarize, blue decreases a fair bit but green to red increases somewhat so the loss of blue in the PAR is offset somewhat by the increase in green and red. But that does not mean PUR stays the same, it will decrease and the corals will not be photosynthesizing as efficiently.

I still contend that LEDs have a great advantage. A 30% decrease in 50,000 hours means that with a 10 hour a day lighting schedule you could theoretically get 13.5 years of use and PUR will drop 30% in that time. But the CCT will stay the same and you should not have problems with nuisance algae as the LEDs degrade. Realistically, how often do you change your MH bulbs and what do they cost each time? My T5s would realistically last no more than a year and replacement costs were $250 to $300. Let's say the LEDs last 10 years then I am saving $2500 to $3000 in bulb costs. Then add the electricity savings as well as the reduced likelihood of needing a chiller (more an issue with MH than T5) along with the stability of the CCT and LEDs look pretty good.

This whole debate reminds me of the T5s vs. Metal Halide debate. There was great resistance to T5s with people strongly believing that you could not have a thriving SPS tank under T5s and that T5s were inferior to MH (heck, that debate is probably still ongoing). But we now know that is not the case. The same accusations are being leveled at LEDs but I do believe that time will show that LEDs are a very good option for lighting reef tanks.

Well first realistically you won't get 13.5 or 10 years out of your LEDs as we already agreed. The same unrealistic numbers are given to both halides and T5s and previously discussed. Second how does your bulb and energy savings compare if you have to replace your fixture every six years compared to simply changing bulbs every year?

The other advantage to T5s that people tend to forget is the ability to setup a kind of bulb replacement rotation. Basically only replacing 2 out of 8 bulbs every six months, every time replacing a different set. This keeps your light levels more consistent eliminating the need to lift and lower fixtures and shocking corals. It also means you're only placing half your bulbs every year which saves you significant money in bulb changes.

The most common argument to go with LEDs is to actually save money in the long run which I think you LED guys should stop using and push more on the lower heat and more defined spectrum (if that really is an advantage).

If you're LED fixture costs $2000 and lasts say 6 years and uses around 250W which means it'll cost around $110 per year for power. If you have to replace your fixture every six years then after 10 years (if you keep it that long) it'll have cost you around $5100.

Now if you compare that to say a 500W halide system which can be purchased for around $1000 (equal quality) and will cost around $220 per year to run for power and around $140 per year for bulbs. After 10 years (again if you keep it that long) that adds up to $4600.

Both realistically are comparable in basic cost however neither include premature failure which is possible for both options however the halide system would offer a cheaper fix. Also after 10 years you only have 2 years left on your second LED fixture and if at some point you decided to sell your fixture for whatever reason the halide system will no doubt return a larger percentage of your investment. And I know that you can argue that you build your own fixture so it was cheaper and you can replace LEDs easily and blah blah blah but what about the rest of us who don't want to or can't build such things, I call this the real world :wink: as at least 90% of people in the hobby don't want to build there own light fixture.

Well you make a few assumptions in your financial analysis. I do believe 10 years is a reasonable lifespan even though you disagree. Since nobody has had a decent LED fixture for 10 years yet we can't really tell. So let's look at the 6 year point as an example. The costs you laid out for the LEDs was $2000 purchase price with $110 per year in running costs for a total after 6 years of $2660. The MH was $1000 to purchase + $220 per year in electricity and $140 per year in bulbs for a total of $3160. That is also assuming you don't need to buy and run a chiller which will add a fair bit to the MH costs.

You have also made the assumption that in 6 years if the LEDs need replacing then a fixture will again be $2000 but the costs are likely to come down dramatically by then and you will get a comparable setup for far less money. But you might also be able to simply replace the LEDs by that point and the costs for comparable emitters would likely be less than $200 for say 80 LEDs by that time. There is a big push to incorporate LED technology into mainstream lighting so in 6 years the economies of scale will bring the prices way down. I can already buy generic 3W LEDs with maybe 70% the performance of Crees for $1.50 each (I have some I plan to test soon). In 6 years they will exceed the performance of the current Cree LEDs at or below that cost. You are also assuming that users of MH or florescent lighting use the same fixtures indefinitely. I venture that most people in this hobby change their fixtures at least every 5 or 6 years. Of course some won't and you will find the outliers who have been using the same MH setup for 10 or 15 years but they would be the exception rather than the rule.

I also think it is a false to assume that repairing an LED fixture is more difficult or costly than repairing a faulty MH or T5 fixture. It may or may not be depending on what fails. Replacing an LED is not any more difficult then replacing an end cap or moonlight and replacing a driver should not be any harder then replacing a ballast. If you would feel comfortable repairing a MH or T5 fixture you should be able to do the same with an LED one. If you aren't comfortable then it has to go out for repairs no matter what lighting technology it uses. And MH and florescent ballasts aren't cheap to replace and they do fail over time. So we should leave the failure/repair estimates out of the analysis because there are so many variables there.

The other day you seemed to agree that 10 years wasn't reasonable, what's changed??
Perhaps changing a ballast is as easy as changing a driver, that's fair enough, but how many drivers are used to drive 250W of LED compared to the number of ballasts for 500W of halide?
How is changing an LED as easy as changing a Halide or T5 bulb? Even the shear number of LEDs to replace in comparison, it wouldn't take much longer to start over. Sorry but if you have to break out a soldering iron it's not that easy. Plus I'd like to talk about actual fixtures not DIY approaches. Like I said let's stay in the real world.

My assumptions were very little, you're assuming way more including price drop when more realistically the LEDs you need to replace down the road will be extinct and difficult to source. You'll have to adapt new types of LEDs into the fixture which will add more complication. Maybe perhaps someone could provide a manual from a supplier of an LED fixture that shows the user how to replace the LEDs? I imagine such thing doesn't exist, for obvious reason.

I say we design LEDs that are easy to replace and make a killing :)

I think it all comes down to the design of any commercial fixture on how easy it is going to fix. Having quick disconnects on parts would allow the plug and play feature you see in most MH systems now. I still remember before when a MH system had multiple components to it before it would work properly. The same would apply here. until a set standard comes out for LED units and connections we are going to see a ton of ways to do it until a "user friendly" way is discovered. We are already seeing this in some of the fixtures coming out, Ron had mentioned the Maxspect fixtures using quick disconnects for there individual emitters (no soldering there). AI has 3 emitter modules (not as efficient but easy to change out)

Regarding the assumption of LED bulb prices droping is not all that off. LED lighting is moving at an incredibly fast pace. In 2000 the cost of a 5mm high intensity white/blue LED would cost $2.00 each from manufactures and now they cost a few cents. No doubt will the same thing happen with high power 3W + LEDs , look at when the luxeons first came out we were looking at manufacture cost of $10 and look at where is it now. LEDs are unique that they are current driven, so even if the specs on newer LEDs change it would be the total forward V and the amount of current taken thats not going to effect the way you pair up LEDs to drivers and should only take a few minutes to figure out how many LEDs can be in an array from simple math.

Watch someone is going to put a patent on quick disconnects now that we've mentioned it for LED purposes.

Already being done :smile: the Maxspect fixtures are nicely designed in that sense. The individual LED stars are fixed to the heatsink by a couple of screws and the electrical connections are by connectors like computer fans. So to replace them you unscrew the screws, disconnect the wire. screw down the new emitter and reconnect the wiring. A couple of minutes and one screwdriver is probably all you need. Also Bridgelux is developing an interchangeable LED module geared towards home lighting etc. but it could probably be used for aquariums too if it works out.

I never said that 10 years wasn't reasonable but rather than argue the point I conceded at least 6 years for our modeling of scenarios. Only time will tell what the real lifespan is but given what is known with 40 years of LEDs existence, 10 years is quite possible.

If you keep the junction temperature below 80 degrees C you will have 30% decrease over 50,000 hours. Keeping those temperatures is not hard. The OC spotlights run with heatsinks below 50 degrees and my heatsink is barely warm to the touch (I will try to get accurate temperature readings later). So if we assume degradation is linear then I will have 20% decrease in output/PAR/PUR after 33,333 hours. That is just over 9 years with 10 hours per day of use. Adjust that accordingly depending on your lighting schedule. Light for 8 hours per day and you could see over 11 years with only a 20% decrease. At 12 hours that is over 7.5 years. But if you want to call it 6 years we can do that for now. Let's talk again in six years when I can do some PAR measurements on my fixture :lol:

As for changing the LEDs it is all down to the design. See my comments above regarding the MAXspect fixtures. As Eugene mentions, the AI units have 3 emitter modules that are easy to change out and shouldn't break the bank. Here's a teardown of the AI so you can see it:

http://reefbuilders.com/2010/03/31/i...aillumination/

And here's something on the Maxspects where you can see one image of how their LEDs are hooked up:

http://reefbuilders.com/2010/01/22/m...ium-led-light/

Both the above would require the use of a screwdriver to replace the LEDs. Real world enough for you :lol: Is it more work then replacing a bulb? Maybe a little but not by much. And using a soldering iron to swap wiring is really not that difficult. If you don't want to deal with that I am sure a repair shop could do it for you fairly inexpensively.

With regards to drivers, they should use multiple drivers. Mine has 8 separate drivers. One failing means replacing one driver at $30 rather than a single ballast at $60 to $150 depending on what type of ballast is used.

That's good news about fixtures already having ways to easily replace the LEDs, I didn't know they where being made like that yet. Sounds like the MAX fixture is going to be somewhat more economical than the AI as it sounds like it takes a more universal LED replacement where as the AI requires replacement sets that can only be purchased from the manufacture. Ideally a standard should be introduced to all these fixtures for an easy screw in replacement or something so various lighting companies can make replacement LEDs just like metal halide and T5 bulbs.

Basically it seems now my only real concerns are reliability and lifespan. I have extreme doubts that these things will do more than half what they are rated for but I guess time will tell, it's just an awfully long time to wait. I would definitely consider a LED fixture if

• The LEDs could be easily replaced with a somewhat universal LED design (open up competition in the future)
• The fixture was visually appealing (something like the giesemann fixtures)
• Made from quality material like aluminum and did not require fans
• Remote drivers making the fixture light enough to be hung easily from two single points
• Controllable by 0-10V interface
• Minimum 3 year warranty

With all that I would expect to pay between $2000-$3000 and would be happy to do so.

cool, I go away for 3 days and stuff been flying.

the problem is like any other lighting there re going to be good ones and bad ones. good ones will run for 50000 hours with out a problem as they have done there home work and designed a system that runs under 80 C , exotic ones may last longer as they will have super cooled systems that run under 40 or 30C which in all likly hood will give a longer life span as it is basicly heat that is the determanaing factor for the life and intensity drop in LED, they are not effected by water, salt, vibration, ect.. unless the water, salt, ect affects the contact between the heat sink and die, but then heat will go up. but yes there will be the junky brands out there that will last under 6 years.

as for the PAR shift in MH bulbs, there is most definatly a shift and it can be large or it can be small. depends on the bulb its self and the ballast you are driving with. a 10000K bulb will generaly have a lot smaller of a shift than a 20000K bulb, but the that in its self realy means nothing. I have measured a 20000K bulb with a 40% reduction in PAR at 14 months with only a 32% reduction in LUX, so right there it is telling you there is much more than just the intensity lowering the PAR value as if that were the case the PAR and lux values would decrease at the same rate. I have had some 10K bulbs where the LUX decreased faster than the PAR and some the otherway. (an theres were all on the same ballast) so it realy depends on the initial make up and quality of the bulb including which halides and the ratios they use to make there color spectrum, as there are hundereds of different combanations that will give you any givin color.

so back to LED :mrgreen: you want to look at longjevity, we can all agree the solaris is a lower end LED fixture, but asside for the odd problems they are going on is it 5th year now (they first came out in 2005 I believe?) with only ballast replacments (the ballasts were garbage ) and the odd LED replacment which is in all likely hood a result of the ballast failing.

Oh and for the record, T5 are infieror to MH :mrgreen:

Steve

thats going to be hard as die's are propritary, but all that being said I guess you could solder any LED to almost any die so who knows.. the easy answer would be to design a holder for the die that attaches to the heat sink. at any rate I understand this one as it is easy on any fixture if you know ho to use a solder gun right now but if you don't..

• Quote:

  Originally Posted by sphelps (Post 511398) The fixture was visually appealing (something like the giesemann fixtures).

like any other lighting I can see nice ones and ugly retro kits.. you have the ability to have a more sleak looking fixture with these maybe two inches thick

• Quote:

  Originally Posted by sphelps (Post 511398) Made from quality material like aluminum and did not require fans.

I think they will all still require fans, at least the high end ones, but the difference will be in the size of the fans. a realy good heat sink design will have small almost totaly silent fans that just insure air movment over the fins, where cheaper designed units will have larger fanse and depend on bulk air movment ro remove heat.

• Quote:

  Originally Posted by sphelps (Post 511398) Remote drivers making the fixture light enough to be hung easily from two single points.

already being done.

• Quote:

  Originally Posted by sphelps (Post 511398) Controllable by 0-10V interface.

same here, it all depends on the company designer and which ballast he picks. I think some will have intagraded controlers also so they are independant from your other equipment.

• Quote:

  Originally Posted by sphelps (Post 511398) Minimum 3 year warranty

hmm.. If I was making a high end product like this I may concider this but 3 years, as a bumper to bumper wouln't make sence.. how about something like 1 year on ballast wiring, electronics, 4 year on LEDs ?

Steve

Steve, the fact that you measured a 40% drop in par and 32% drop in LUX would suggest the two are directly related. The difference of 8% is minor considering the likely test conditions and equipment used . Anyways the point I was making as a shift in CCT (not PAR) does not necessarily mean your par has dropped, depending on the bulb and ballast it could mean an increase or a decrease in PAR meaning the two are not directly related. A decrease in PAR is simply a result of an overall decrease in intensity which could be measured in LUX. The argument was that LEDs last longer than halides because they don't suffer the CCT shift meaning a more stable PAR reading over time, I suggested this isn't the case.

Producing a standard LED replacement for fixtures is no different than any other lighting type, they've already adapted LEDs to fit into almost all standard residential applications. This is no different, first the standard is introduced and then manufacturers build fixtures to accommodate them.

Retro kits are different than fixtures but yes there are good looking and bad looking fixtures for almost everything. Currently the LED fixtures I've seen are more in the cheaper plastic look like cheaper halide and T5 fixtures, nothing really high end yet with the exception of a few module based fixtures such as the AI but the module design itself isn't visually appealing IMO. Realistically this is a very small constraint and can be easily overcome.

Most dimable or controllable lighting fixtures are based on a standard 0-10V system. While not everyone would require the ability to externally control their light fixture it should certainly be an option. One would think with the current patent on LED fixtures with built in controllers having a 0-10V plug in for an external controller would be a good thing.

Proper design of a heat sink could maximize the inductive heat transfer eliminating the need for fans. Fans are needed in the current fixtures because the heat sinks are simply standard units not designed specifically for LEDs and therefore requiring addition convective cooling. Designing the heat sink to concentrate additional surface area at the LED locations would be the first step. There are also different grades of aluminum or other materials which offer higher inductive properties than what is probably being used. Fans simply cut the cost of the heat sink dramatically which is why they are used. Eliminating the fans would actually increase quality and reliability since fans often fail or become restricted over time which could result on over heating. Also using fans will result in an uneven heat distribution meaning some LEDs will run warmer than others. These two problems could explain why fixtures have suffered from LED failure, I can't imagine any fan lasting for 10 years over a SW aquarium. Eliminating the need for fans would definitely increase reliability.

Perhaps the warranty isn't as big of an issue. Most other fixtures and equipment we use is usually limited at 1 year as well. The difference with these LED fixtures is they are new and marketed as something that will last 10 years and save you money in the long run. Manufacturers could back up these claims by offering better warranty and with a higher end fixture I would expect it to be that way but perhaps such a fixture doesn't exist yet. I wouldn't expect the warranty to cover every component but if I look at my RD pump it was marketed as an efficient quality pump designed to last longer and therefore actually save you money in the long run. It comes with a lifetime warranty on the bearings. I've had bearings fail on other pumps and while it's not the only component that can fail it still says the manufacturer is standing behind there claims and product.

If you look closely at well designed heatsinks they all incorporate fans to dissipate the heat and thats because most components work better at lower temperatures. One of the things that you see heatsinks on mostly are computer parts, there are heatsinks that are "silent" and require no fan but run the processor at 10-20C higher than the same heatsink with fans. The design of the heatsink itself is definitely going to help with the heat dispersion but if these units are run without a fan you/we can all expect them to never meet the 50,000hr that the manufactures claim. The main reason is heat, the lower the heat you can keep on the LEDs the more efficient they run, and the slower the intensity drop on the LEDs as well. If you want to see an example, wire some 5mm leds and put them in isolated environments of room temp, 50C, 80C you will no doubt see the 80C intensity drop along with a major color shift as well.

If manufactures actually source out reliable fans for there fixtures, like those found into computer cases, low RPM high CFM fans the units will be near silent yet still provide the cooling needed for the LEDs. Some companies will still try to build fanless LED fixtures, elos is one that comes to mind, but they dont drive the LEDs to its full capacity in those fixtures. But again it all comes down to what the company wants to do with the units, so i think what needs to be understood is how to channel the air flow inside the fixture itself to provide efficient cooling so that all emitters get cooling thats somewhat the same.

I think it is not only heatsink design but the overall fixture design as well. Convection cooling can work if the heatsink is open to the air and visible. if you want to enclose it in a housing then you need fans to help move air. And even if open to the air fans will help keep the LEDs cooler than convection cooling alone.

My fixture is enclosed in a housing and I am running 3 low RPM computer fans. I cannot hear the fans at all (well maybe if I put my ear right next to them) and my heatsink is very cool. So I expect very long life out of my array with no noise. The fans should last as they are low rpm thus putting low stress on the bearings and are on top of a closed housing. Unless the location of your aquarium is excessively humid or you frequently splash water up and over your light fixtures I don't see fan reliability being a huge problem.

If you look at the Solaris fixtures that suffered LED failures it was not likely due to the fans but to the fact that there were no real heatsinks to help draw heat away. The heatsink setup of the Solaris units was completely inadequate and I do not think it is representative of what should and could be done with LEDs.

wohwoh man, I never said they wern't related only not the only factor. the simmple fact is if the big blue spike shifts to the left, PAR will drip off as it shifts compound this with a decrease in intensity and it will be compounded.

you can not gage PAR out put with LUX, two totaly diferant and unrelated measurments. you are in one case measuring the amount of the emitted light with in a specific wave length, in the other you are measuring the amount of visiable light that falls on a given area. kind of a mass vs PSI but of different componants. Like I said in the previous post, no two MH or T5, ect age identicaly. this is due to various factors, including gas concentrations/types inside and so on. LED eliminates this by using a solid state chip encapsulated with no air/gas/ect surounding the chip, but they do have a problem as mentioned... heat. if they are allowed to warm up they will go through a reduction in PPF, which rises again once cooled. this can be as much as a 30% reduction in as little as 1/2 an hour with out any cooling and no drop in intensity. so using a heat sink and fans properly is a twofold benifit.. maintains the PPF output and the life on the LED. I think you would need the fans personaly, other wise the fixtures would be huge inorder to give each LED enough heatsink to make passive cooling possible. by having fans and preperly using them you will be able to produce lighter, more sleak fixtures. for what it is worth I had 4 computer fans running over my sump for 8 years starting and stopping 4 times a day. although they were coated with salt spray they still worked..

Steve

ya warenty is a tough one.. I can build you a light fixture that is bullit proof, but I can guarentee the ballasts or bulbs will last more that what the warentee I get when I buy them.. now pumps and berrings is a different story, there is no reason any manufacture can't offer life time warenty on berrings as you'll find that is any simular sized pump designed to do the same thing the berrings will be very simular if not the same. I have been rebuilding pumps as part of my former job for over 20 years and it was very very rare I came across a berring problem. most of the time the pump wore out befor the berring did. but we did change berrings on some just becaue they are very cheep and we were rebuilding the rest of the pump, but I did save the old berrings for emergency spares..

Steve

Agreed that you can't measure PAR from lux directly, obviously the PAR of a bulb depends on the spectrum distribution. However that was never my point. I'm purely talking about bulb wear and the simple fact is a decrease in PAR over time is the result in an overall decrease in intensity which could be measured by LUX. For example mass and psi are separate measurements but if you have a mass over a given area and the area and gravitation force remains constant then a decrease in PSI means a decrease in mass.

Your "big blue spike" does not shift, or move to the left or right, it simply decreases in intensity and the overall CCT shifts to a lower K rating. Depending on the bulb you can also see an increase in other areas of the spectrum like red and green which is why some bulbs will actually increase in PAR overtime or stay relatively constant. In order for PAR to decrease the overall intensity of the bulb must decrease.

As for fans, they are not needed, there are already plenty of LED lights for other applications which use high powered LEDs with properly designed heat sinks and no fans. It's a simple mater of cutting costs and the temperature gradient and heat transfer rate of LEDs is high and low enough respectively to eliminate the need for convective cooling. Computer components like CPUs run much hotter, their heat flux is way higher than an LED and require convective cooling. Computers are also built as cheap as possible, it's a huge commodity market.

:lol: So you want a fixture that looks stylish, has a long warranty but doesn't run any fans. I'm not sure all those are compatible criteria :mrgreen:

Well the warranty isn't that big of a thing if the quality is obviously there. But yeah for sure on stylish with no fans, otherwise I'll stick with what actually works, looks good, is reliable and quiet. I call them Halides and like it or not they are the best all around type of lighting for the average tank, just saying :lol:

Im not so sure about that, give me an example of high power LEDs thats out there that doesn't utilize some form of active cooling, the ones found in the new Audi cars utilize passing air from inside the engine bay is a quick one that comes to mind. Its not so much cutting cost, im sure that has something to do with it, But if the heatsinks were made adequate enough for the LEDs to start the addition of fans only adds to the overall benefit of the fixture by keeping the LEDs at a much cooler temperature than without. You keep pointing out that fans have reliability issues. If you rip open one of the fixtures made for our hobby you'll quickly see that the fans they use inside to reduce cost....only costs $1-2. A Prop fan that is made by a good company usually run in the range of $10-12 and are 10x more quite than the cheap ones and push more air as well.

Also i would differ to say that computers are build as cheap as possible, at my previous job i was custom building PCs to fit the need of individual people and if you dont get into the watercooling side of things, the heatsink design + fan is really focused upon in the design of a system. The cooling of a system sometimes allows you to push that much more "juice" out of the system without getting lag. Theres only so much a heatsink can do, in other words you'll never see a top of the line computer system run without some form of active cooling on its components.

The other thing to consider is weight. Using fans lets you use smaller heatsinks. I have big heatsinks in my unit that probably would have done the job without fans and the bloody fixture weights at least 50 pounds. That is heavy. If I could have found thinner or lighter heatsinks in my price range I would have gone with them.

As you say, LEDs generate less heat than a CPU but you don't have 80 CPUs on one heatsink either. Either way, if you want to run the LEDs with higher current (more output) and keep the size and weight reasonable then fans are likely necessary. if you look at the units without fans they are either using 1 watt LEDs or driving 3 watt LEDs at closer to 1 watt to minimize heat.