Water cooling using aquaponics tank

[frica]

I wouldnt say the epistar leds are better, but they are definitely cheaper per watt. Its a big hit to efficiency but when cooled very well and underdriven it can reach at least within arms length of the cree cobs at rated power levels. Obviously the same thing done to a cree will do a lot better, but initial cost is very high, and if i would have had a catastrophic failure and destroyed the cobs like i just did to my set i was just using, i would probably have cried a bit lol. Especially for initial testing purposes this was definitely the way to go for me. Of course not many ppl do aquaponics or have a large enough tank to support this amount of cooling, so it wont work for 99% but if someone wants to do it, it can be done.

Even when underdriven and water cooled the Epistars get destroyed in efficiency unless you cherry pick bins.

Besides water cooling will only cool the COB at most a few degrees more than a proper air cooler.

mahiulana claims the Epistar produces 138.2 lm/w at 721mA.
https://www.rollitup.org/t/pictures-of-your-diy-lights-post-your-pics.871625/page-81#post-13439413

It's quite clearly a 5000/5700k light, and Cree COB of similar CRI and colour temp go up to the DD bin.
http://pct.cree.com/dt/index.html

A DD binned CXB3590 at 50W with a junction temp of 105 degrees celcius is still 168.1 lm/w.
That makes the 3590 21.6% more efficient.

But a juntion temp of 105C is very high, even with passive cooling.
A Tj of 85C is much more realistic and puts it at 175 lm/w.
26.6% more efficient.

With a proper air cooler you should be able to get a Tj of 50C.
189 lm/w.
36.7% more efficient.

Electricity costs are much higher than the initial price of the COBs and if a properly cooled 3590 is 36% more efficient then the 30-40 dollar you pay extra is easily earned back.

And there are chips just as good/cheaper than the 3590.
The Luminus CXM22 is only 19 dollar.

 

[Ken Beck]

You are trying to sell something using shady tactics. We are not.

How would your cool mac light improve upon his current setup (it wouldnt)?

Yea i looked at it and it does have a nice mounting and is well built but i dont think it would improve what i have either. The next iteration will definitely be closed loop and i will coil the loop around the edges in the tank to spread the heat evenly through it and i should be good from here on out.

 

[Ken Beck]

Even when underdriven and water cooled the Epistars get destroyed in efficiency unless you cherry pick bins.

Besides water cooling will only cool the COB at most a few degrees more than a proper air cooler.

mahiulana claims the Epistar produces 138.2 lm/w at 721mA.
https://www.rollitup.org/t/pictures-of-your-diy-lights-post-your-pics.871625/page-81#post-13439413

It's quite clearly a 5000/5700k light, and Cree COB of similar CRI and colour temp go up to the DD bin.
http://pct.cree.com/dt/index.html

A DD binned CXB3590 at 50W with a junction temp of 105 degrees celcius is still 168.1 lm/w.
That makes the 3590 21.6% more efficient.

But a juntion temp of 105C is very high, even with passive cooling.
A Tj of 85C is much more realistic and puts it at 175 lm/w.
26.6% more efficient.

With a proper air cooler you should be able to get a Tj of 50C.
189 lm/w.
36.7% more efficient.

Electricity costs are much higher than the initial price of the COBs and if a properly cooled 3590 is 36% more efficient then the 30-40 dollar you pay extra is easily earned back.

And there are chips just as good/cheaper than the 3590.
The Luminus CXM22 is only 19 dollar.

I was running way lower than .721a to get over 150lumens per watt lol. A little over .3 amps at 29.2v got me 158lumens a watt with the water cooling...very low wattages.

 

[Ken Beck]

Cree COB of similar CRI and colour temp go up to the DD bin.
http://pct.cree.com/dt/index.html

A DD binned CXB3590 at 50W with a junction temp of 105 degrees celcius is still 168.1 lm/w.
That makes the 3590 21.6% more efficient.

But a juntion temp of 105C is very high, even with passive cooling.
A Tj of 85C is much more realistic and puts it at 175 lm/w.
26.6% more efficient.

With a proper air cooler you should be able to get a Tj of 50C.
189 lm/w.
36.7% more efficient.

Electricity costs are much higher than the initial price of the COBs and if a properly cooled 3590 is 36% more efficient then the 30-40 dollar you pay extra is easily earned back.

And there are chips just as good/cheaper than the 3590.
The Luminus CXM22 is only 19 dollar.

Does any supplier even have a DD bin of the cree cxb3590? I swear every good one is more rare than bigfoot sightings. I can find the high bins of 3500k leds all day but ask for a 4000k or 5000k and suddenly nobody has anything or ridiculously long lead times. Honestly at this point i'd pay almost anything for the high end crees, high bin luminous, or high bin citizen cobs as long as the wait was a week or less to get it.

 

[frica]

I was running way lower than .721a to get over 150lumens per watt lol. A little over .3 amps at 29.2v got me 158lumens a watt with the water cooling...very low wattages.

Using warm white COBs?

Does any supplier even have a DD bin of the cree cxb3590? I swear every good one is more rare than bigfoot sightings. I can find the high bins of 3500k leds all day but ask for a 4000k or 5000k and suddenly nobody has anything or ridiculously long lead times. Honestly at this point i'd pay almost anything for the high end crees, high bin luminous, or high bin citizen cobs as long as the wait was a week or less to get it.

https://www.cutter.com.au/proddetail.php?prod=cut2812

This should be the DD bin.

Hard to find with RIU distributors because high kelvin COBs aren't very popular.
3000/3500/4000 are by far the most popular.

 

[mahiluana]

so it wont work for 99%

I agree 101% - i have no aquaponics - but i have a bathroom and kitchen !
Like you I have a 300W light - i run it 15h and can save ~80L of hot water every day.
Enough for 2 persons. Previously i used hot water from oil central heating and in my 3 room flat
there is a hot water clock where the owner of the house can measure the anual quantity.
50% of my heating bill(EURO 120,- / month) is calculated by this amount of hot water running through the clock. So i save ~ EURO 60,- which is more than my electricity bill

300W*15h*31 days

= 139,5 KWh cost me EURO 35,- . So the owner of the house will pay in future my electricity bill - which is around 780,- EURO / year. and he doesn`t know.

OK. not everybody got the possibility to connect his grow room to the shower
but more than 1% for shure.
I connect myself with a 10m flexible garden hose from grow room to bathroom.
I plug and play in few seconds.

i am curious if it would work to the point of making the leds hit even higher lumens per watt.

Cooling down your junction temp. 30°C rise your luminous efficacy of radiation about 5,5%.

You can do a simple test to measure the heat of your chips - it always 75-80% of Vf * If

 

[frica]

Sounds like your landlord is ripping you off with warm water prices.

May as well just install an electric heater for water in that case.

Or you're lying.

 

[frica]

And again with that flawed way to measure the heat.

You don't think it's a little flawed if your test always come out between 75-80% when the difference of efficiency between chips is much greater?

Also LEDs have been sphere tested, which is going to be much more accurate and yet your findings are completely different.

 

[mahiluana]

Sounds like your landlord is ripping you off with warm water prices

You are right. But this is your only "diagnosis" in which i agree !!!

just install an electric heater for water in that case.

Thats what i did - you are to stupid to put a condom on.

You really suggest to double my electricity bill and not having savings ?????

This was my last post to you and Mr. @Shugglet//no time - no brain for hopeless cases.

 

[Ken Beck]

Using warm white COBs?


https://www.cutter.com.au/proddetail.php?prod=cut2812

This should be the DD bin.

Hard to find with RIU distributors because high kelvin COBs aren't very popular.
3000/3500/4000 are by far the most popular.

No. These are definitely 5000k cobs ive been running.
Thanks for the link. I'll be checking that out

 

[frica]

You are right. But this is your only "diagnosis" in which i agree !!!




Thats what i did - you are to stupid to put a condom on.

You really suggest to double my electricity bill and not having savings ?????

This was my last post to you and Mr. @Shugglet//no time - no brain for hopeless cases.

I am the hopeless case?

Also I've asked this a few times already, if an UV-C led puts out 200mw light while pulling 1W.
Are you going to argue it's 0% efficient because its luminous efficacy is 0?

Why do you constantly ignore most points and only respond with personal insults?

Also why would your condom test be more accurate than a sphere test?
You still haven't answered that one either.

You're just trying to sell your shitty system by spreading misinformation.

Just like you were comparing a CB binned 3590 next to a 5000/5700K epistar when 5000K 3590s go up to DD.

There are more examples, people can look at your post history.
You're just spreading lies, misinformation and insults.

 

[mahiluana]

I am the hopeless case?

you can always repeat your lessons - go back and read - till you understand the difference between luminous efficacy of radiation and light efficacy.

You still - don`t understand the main difference between the luminous efficacy of radiation and the luminous efficacy of a source. That means the latter accounts for input energy that is lost as heat or otherwise exits the source as something other than electromagnetic radiation. Luminous efficacy of radiation is a property of the radiation emitted by a source. Luminous efficacy of a source is a property of the source as a whole.

With a luminous efficacy of radiation (XTE Royal Blue 57% @ 0.35A.)
you are still around ~25% (24,51%) of luminous efficacy of your source.

If in 10 years - or so - Mr. Nichia or Cree or whoever will patent a blue light source with 100% luminous efficacy of radiation it will only have a luminous efficacy of the source ~ 43% and still produce ~57% of heat.

You still believe(d) that your (XTE Royal Blue 57% @ 0.35A.) converts
only 43% of P = Vf*If into heat and - that`s NOT CORRECT.

You dream(ed) about a led chip,that doesn`t produce heat - this is NOT POSSIBLE.

I still believe that a led will always produce heat with a min. @ 57% heat efficacy.
43% /100 *57 = 24,51% for your blue XTE and pretty exactly what i measured with my condom.

So even maybe in ten years with the highest luminous efficacy possible - i am able to double energy efficiency with a watercooled lamp.

Simply adding 1+1=2

Luminous efficacy 43% + 80% of the produced heat(57%*80%) = 88,6% energy efficient.

My Coolmac system in the moment stands where your blue chip stands:

Luminous efficacy 24,51% + 80% of the produced heat(76,49%*80%) = 85,638% energy efficient.

Hope you guy check this at the end and realize that watercooled led is not only a "little"
revolution in light tec - but also a powerfull tool to combat climate change and have enormous savings. A smart world with watercooled led light could save 10% primary energy(oil, gas, electricity...)

shady tactics

1+1=2 HUUUUUHHHH that`s really dark and hard to understand.

 

[frica]

you can always repeat your lessons - go back and read - till you understand the difference between luminous efficacy of radiation and light efficacy.

You still - don`t understand the main difference between the luminous efficacy of radiation and the luminous efficacy of a source. That means the latter accounts for input energy that is lost as heat or otherwise exits the source as something other than electromagnetic radiation. Luminous efficacy of radiation is a property of the radiation emitted by a source. Luminous efficacy of a source is a property of the source as a whole.

With a luminous efficacy of radiation (XTE Royal Blue 57% @ 0.35A.)
you are still around ~25% (24,51%) of luminous efficacy of your source.

If in 10 years - or so - Mr. Nichia or Cree or whoever will patent a blue light source with 100% luminous efficacy of radiation it will only have a luminous efficacy of the source ~ 43% and still produce ~57% of heat.

You still believe(d) that your (XTE Royal Blue 57% @ 0.35A.) converts
only 43% of P = Vf*If into heat and - that`s NOT CORRECT.

You dream(ed) about a led chip,that doesn`t produce heat - this is NOT POSSIBLE.

I still believe that a led will always produce heat with a min. @ 57% heat efficacy.
43% /100 *57 = 24,51% for your blue XTE and pretty exactly what i measured with my condom.

So even maybe in ten years with the highest luminous efficacy possible - i am able to double energy efficiency with a watercooled lamp.

Simply adding 1+1=2

Luminous efficacy 43% + 80% of the produced heat(57%*80%) = 88,6% energy efficient.

My Coolmac system in the moment stands where your blue chip stands:

Luminous efficacy 24,51% + 80% of the produced heat(76,49%*80%) = 85,638% energy efficient.

Hope you guy check this at the end and realize that watercooled led is not only a "little"
revolution in light tec - but also a powerfull tool to combat climate change and have enormous savings. A smart world with watercooled led light could save 10% primary energy(oil, gas, electricity...)



1+1=2 HUUUUUHHHH that`s really dark and hard to understand.

And again I have to tell you that the luminous efficacy isn't just a product of the efficiency of a lightsource it's also a product of how well it's detected by the human eye.

And again, answer this question.
Are you arguing that a hypothetical UV-C led that produces 200mw UV-C at 1W is 0% efficicient because it produces exactly 0 lm/w?


According to you a light that is sphere tested to produce up to 575mw radiant flux while using only 1W is not 57% efficient despite the fact that it produces 575mw radiant flux from an 1W input.

Using luminous efficacy as a measure of thermal efficiency is critically flawed as the human eye's sensivity to certain wavelengths is also a variable and that has absolutely nothing to do with thermal efficiency of the lightsource.

And again, just answer this:
Are you arguing that a hypothetical UV-C led that produces 200mw UV-C at 1W is 0% efficicient because it produces exactly 0 lm/w?

 

[frica]

http://www.cree.com/news-media/news/article/cree-launches-the-industrys-brightest-and-most-efficient-royal-blue-led

DURHAM, N.C. -- Cree, Inc. (Nasdaq: CREE) announces the new XLamp® XP-G3 Royal Blue LED, the industry’s highest performing Royal Blue LED. The new XP-G3 LED doubles the maximum light output of similar size competing LEDs and delivers breakthrough wall-plug efficiency of up to 81 percent. This superior performing Royal Blue LED expands Cree’s leading high power portfolio, enabling lighting manufacturers to deliver differentiated LED solutions for applications such as horticulture, architectural and entertainment lighting.

Using the new XP-G3 Royal Blue LED and the recently introduced XP-E High Efficiency Photo Red LED, Cree has created a new horticulture reference design that achieves a Photosynthetic Photon Flux (PPF) efficiency of up to 3.2 μmol/J at steady-state, which is over 50 percent more efficient than the traditional high pressure sodium solutions in use today. The XP-G3 Royal Blue LED delivers up to 3402 mW radiant flux, which corresponds to 13 μmol/s PPF, at its 2A maximum current and 85 C junction temperature.

 

[Shugglet]

1+1=2 HUUUUUHHHH that`s really dark and hard to understand.

Why dont you address the actual points I addressed? Did I ever claim you were being shady about 1+1=2???

Nope, but how bout your other scenarios involving Crees vs Epistars?

Or why dont you ever answer how efficient the drivers you use are?

The reason you dont address these points is because they will expose you.

With a luminous efficacy of radiation (XTE Royal Blue 57% @ 0.35A.)
you are still around ~25% (24,51%) of luminous efficacy of your source.

Can I ask you how you can claim to know the luminous efficacy of a source when you dont know anything about the source other than the LED? No driver mentioned, no cooling/recapturing method mentioned, yet somehow you feel you can remotely accurately claim a luminous efficacy of said source??? Where do you pull that from ? Your ass?

 

[Ken Beck]

http://www.cree.com/news-media/news/article/cree-launches-the-industrys-brightest-and-most-efficient-royal-blue-led

DURHAM, N.C. -- Cree, Inc. (Nasdaq: CREE) announces the new XLamp® XP-G3 Royal Blue LED, the industry’s highest performing Royal Blue LED. The new XP-G3 LED doubles the maximum light output of similar size competing LEDs and delivers breakthrough wall-plug efficiency of up to 81 percent. This superior performing Royal Blue LED expands Cree’s leading high power portfolio, enabling lighting manufacturers to deliver differentiated LED solutions for applications such as horticulture, architectural and entertainment lighting.

Using the new XP-G3 Royal Blue LED and the recently introduced XP-E High Efficiency Photo Red LED, Cree has created a new horticulture reference design that achieves a Photosynthetic Photon Flux (PPF) efficiency of up to 3.2 μmol/J at steady-state, which is over 50 percent more efficient than the traditional high pressure sodium solutions in use today. The XP-G3 Royal Blue LED delivers up to 3402 mW radiant flux, which corresponds to 13 μmol/s PPF, at its 2A maximum current and 85 C junction temperature.

Speaking of wall plug efficiency, what is the highest efficiency power supplies you have seen? Ive seen 92-93% but they are super expensive and usually only 100-200w or so. I have a 350w power supply that is good for high 80's for most wattages and was fairly cheap. I guess its cost vs savings chart time lol.

 

[wietefras]

@Ken Beck, Mean Well drivers go over 96% in some cases. Actually the bigger models are better at efficiency than the smaller ones.

 

[Ken Beck]

@wietefras id be using the 115v efficiency line at 30% to 50% load depending on the cobs. Cheapo chinese around the 50% mark and cobs like the 5000k 3590 DD bin at 30% load. So between 92.5% and 94.5% efficiency(half a percent lower than the chart for the 54v version as id be using the 36v). Not bad. but it is a bit pricey at $160+
It'll pay for itself in about 7yrs in energy savings lol. I mean they definitely would be a permanent solution. Maybe bragging rights on efficiency is the better justification as i could have over 200 lumens a watt from the outlet at 200w power with 4 cxb3590 DD 5000k cobs and that driver. Im going to order the cobs this weekend and maybe the driver soon after. Also getting a heat exchanger to put in the tank to keep the cooling loop separate and clean.

 

[ttystikk]

@wietefras id be using the 115v efficiency line at 30% to 50% load depending on the cobs. Cheapo chinese around the 50% mark and cobs like the 5000k 3590 DD bin at 30% load. So between 92.5% and 94.5% efficiency(half a percent lower than the chart for the 54v version as id be using the 36v). Not bad. but it is a bit pricey at $160+
It'll pay for itself in about 7yrs in energy savings lol. I mean they definitely would be a permanent solution. Maybe bragging rights on efficiency is the better justification as i could have over 200 lumens a watt from the outlet at 200w power with 4 cxb3590 DD 5000k cobs and that driver. Im going to order the cobs this weekend and maybe the driver soon after. Also getting a heat exchanger to put in the tank to keep the cooling loop separate and clean.

Ignore the arguments; your system is elegant, efficient and effectively utilizes all of the output from your COB LED chips, whether it's light or heat.

Brilliant!

 

[Ken Beck]

Ignore the arguments; your system is elegant, efficient and effectively utilizes all of the output from your COB LED chips, whether it's light or heat.

Brilliant!

I would say its good but it could always be better. This project has more hours invested than money lol. I wanted to make this system completely automated as well. That board mounted above the water in the tank used to have equipment but now will have an automatic fish feeder dispensing food to the fish on timed intervals and the light and again a separate pump for the cooling loop have their own timer set up now. Once the rest of the project is completed and im satisfied with its performance and reliability, i will duplicate it as i already have another ibc tote i can use. If i can get the cheap version right now perfected, i will definitely be building more for ppl to use. $80 for led related equipment and cooling, $40 for ibc tote, $46 in fish ($2 each), and $50 in hydroton. Total cost=$216 for entire system setup. Idk what its performance would be for weed but for other flowering plants it works great. Veg growth is very good and surprisingly the flowering started as soon as the plants were mature. Nitrates are still well above 160ppm with a bed that has 2 cherry tomato plants, a jalapeno plant, 5 romaine lettuce plants, banana pepper plant, cucumber plant, rhubarb plant, and 3 strawberry plants. All are producing very well and still arent consuming all of the nutrients and i have no room left to plant anything else. Should i do a parts list and step by step build with pictures for the next build for the people who arent familiar with this?