Lighting plants with white LEDs - about efficiency and cost-effectiveness
After writing the previous article , I myself was left with the question not fully resolved - what exactly is it more profitable to buy and how much can you win in the near and far perspective. Plus, there are some uncertainties on the effectiveness of LEDs. And the question encourages the search for an answer to it, so I continued to develop this area. I will not say that the material turned out to be a full-fledged article, but as an addition to the previous one, the information contains essential data will be useful.
To begin with, let's look at the exact efficiency of the LEDs discussed in the last part. Earlier, I took the data mainly from the iva2000 article , without checking, because there was considered more about the effectiveness of photosynthesis when illuminated by light of a different spectrum. Now I decided to understand the overall efficiency.
Consider the LEDs of the company CREE, because on the one hand, they are by far the most advanced in technology and, consequently, the light output per unit of power, and on the other, all their indicators are stable and well documented (in contrast to manufacturers' nounaym). Here, this company should pay me for advertising, but alas, I do not write with their submission, but simply because it is easier and more accessible.
So, what are we going to explore the LEDs? I will not post here the whole process of studying and selecting specific series in order not to flood the material with “water”. In short, I will say that I took the most powerful and at the same time the most effective chips, provided that it is freely available and at a bargain price. According to these criteria, two types are suitable: whites will be from the XM-L series.
- These are 10-watt chips with an efficiency of 158 lm / W (but not at maximum power, but only at 1 W). Cold white (6000-6500K), neutral white (4000-4500K) and warm white (3000-3500K).
And red from the series XP-E, High Efficiency Photo Red 650-670nM.
Links to documentation on LEDs at the end of the article.
We will deal with white. Last time, the difference in efficiency of white LEDs was not taken into account and efficiency was evaluated only in relation to the McCree photosynthetic activity curve.
This time I decided to clarify this question more thoroughly. Unfortunately in the documentation for the LEDs never lead to efficiency, and write lumens per watt, so I had to do the reverse calculation. According to the spectrum of the LED and the photopic curve, it is calculated how many lumens the LED would have if its efficiency was 100%, and then the number of real lumens taken from the LED documentation is divided by this number. And that's what we got for the three types of white LEDs:
From left to right: cool white, neutral white and warm white.
It is noteworthy that in spite of the growth of lumens during the transition from the cold-white to the warm-white spectrum (with the same radiation power), the tabular values of lm / W and the overall efficiency of the LED fall very significantly from 40 to 23%. The fact is that the phosphor, of which there is much more warm white light in the LED, itself has not 100% efficiency, and, apparently, with its large amount it has a shading effect (the rays emitted by the lower layers are absorbed above lying and disappear ). At the same time, the indicator lumen per watt is used at a current of 2A (out of a maximum of three) - it can be seen that at the same time it drops from 140 at 350mA to 108 (for cold white). There is no such table in the Cree document - absolute lumens are given for a given current, and the power must be calculated using data from the graph of the current-voltage characteristic. Here are the relevant data from the datasheet:
Now let's deal with the red.
Everything is a bit easier with them, because the luminous flux is not specified in luminas, but in milliwatts. It is enough to divide milliwatts of radiation by watts of consumption and obtain efficiency with high accuracy! On all the LEDs would bring this data - 2/3 of the work could not be done!
And here we immediately make an amazing discovery - that the efficiency of these LEDs is 50%, and (one more graph, I don’t give here), unlike blue / white crystals, the luminous flux increases linearly with current and the efficiency of the chip does not fall! But when the chip overheats, the drop is much more significant than that of the blue chips. For comparison, in pure blue efficiency under the same conditions 48% (compare with this indicator for whites - higher). But the "just red" is much worse. Their efficiency turned out to be somewhere around 19%, and as the temperature rises, the luminous flux drops even faster than that of Photo red.
Interesting uses of individual LEDs and their combinations are already emerging. Now we will recalculate the efficiency table taking into account the newly obtained data.
It is seen that the red Photo-red with a large margin ahead of all. But it is impossible to cover with pure red, therefore it is necessary to combine and here there are options with white and blue. Immediately, we note (I thought everything, but threw away what was not promising) a combination of warm white and red. The low efficiency of warm white LEDs negates all the benefits of red. But cold white is very good in this combination! Themselves have a good efficiency, still enhanced by red LEDs, and the lack of the red spectrum is also covered by them. It also looks good combination of red and blue. Then they go simply cold-white and DNaT 1000, and the rest, in fact, do not pull. Well, let's see how it will look in full - with the drivers.
Further, the calculation logic went on the assumption that we want to receive more photosynthetically active radiation for the same money, therefore all figures, including the prices for LEDs and drivers, are reduced to the total phytoactive radiation of the lamp 100 µmol / s.
Color marking as in the previous table - to make it easier to understand where are the LEDs and do not take the place of duplicate headings.
But this is only the price at the start - how much money you need to invest in order to get a light bulb at 100 µmol / s. This is not enough - you need to see how much it will cost during operation. And if you count on this, and the cost of electricity in time - that's when you get the full picture, which I present for all to see!
Thanks to the close attention of the commentators, it turned out that not all the LEDs that sell on the aliexpress called CREE are in fact. The cheapest of them, about one and a half dollars per 10-watt diode or less likely, are fake chips made by the Chinese company LatticeBright, which are several times cheaper than the original ones and, unfortunately, have about 2 times the worst performance. In this regard, I conducted a search for the prices of the corresponding LEDs in the Compal company, which is the official distributor of cree in the Russian Federation. Prices there are much higher than in China, but small wholesale is quite profitable, including in comparison with foreign suppliers.
And along the way I corrected two points - I added the replacement of lamps once a year for the HPS curve. And corrected the error (my oversight), due to which the price of all the lamps was considered to be of the same power (100W), whereas the original idea was based on the unit of photoactive radiation. In the new graph, these prices are for a lamp emitting 100 µmol / s, not 100W. I apologize for the mistake.
How to understand this bundle of twigs?
On the left - the price of the lamp at the start. I remind you that all of them will produce the same amount of phytoactive radiation, but have a different spectrum. The lower the strip begins, the cheaper the set. On the X axis we have months. It is assumed that the lamp works 12 hours a day, 7 days a week, for a total of 36 months, i.e. 3 years. This is just a little more than 13 thousand hours, and 50 thousand are stated for LEDs. And if everything is done correctly with cooling, and a current of 0.7 from the maximum is applied to the LEDs (so there is more efficiency by a full third), then they will work even more i.e. more than 10 years with almost no degradation.
The more horizontally the line goes - the more efficiency the luminaire has. We see that many lines start higher (more expensive chips), but with time they turn out to be cheaper than cheaper counterparts. This is an indicative line for photo red LEDs - it has the smallest slope.
The most amazing thing is that the cheapest are now ... The most expensive photo red LEDs! This is because they have the highest efficiency and the most "easily digestible" spectrum - they need the least in the beginning and they spend the least electricity in the future! Of great interest are the combinations "Cold White + Red Photo Red". This graph shows the curve when the ratio of white: red as 2: 1 in power. And just "cold white." These three lines diverge like a fan, where the extreme ones are white and red LEDs, and the middle one is their combination. For growing plants all components of the spectrum are necessary, but in different combinations. It turns out that all options for combinations of spectra are most effectively covered by just one combination - cold-white and red LEDs (but in a different numerical ratio).
It is worth noting that the combination of blue + red, although it has a smaller inclination than white + red, but gives a significantly worse price / light output, so the combination of white + red does not catch up with even 3 years. In the 10-year term, it may be preferable, but this is an exceptional case.
Phytolamp is not so cheap. If we take into account its efficiency, it is even more expensive than cold-white LEDs, and only in the future ... Money for electricity to the wind ...
DTA and at the beginning is not very cheap (I was surprised at how much electronic ballasts cost to them, but EM PRA should not be taken - they have low efficiency, the lamp due to flicker - also, they also buzz and heat like a stove) and do not catch up with time - especially with the replacement of lamps - which will have to do at least once a year, which is displayed as steps on the chart. So go to the garden.
Here is the spectrum of the combination of white and red LEDs superimposed on the MkCree curve (4: 1 in power, 2: 1 did not alter):
Of course it is wrong to judge such things based on the beautiful charts, but given the numbers that say the same thing - in my opinion the graph is almost perfect in terms of covering the spectrum of the photosynthetically active range.
The conclusion remains the same - buy cool white LEDs and red CREE Photo red and you will have a lot of light for your plants and savings for your wallet!
Lighting with pure red LEDs is also possible; one of the commentators wrote about this experience. This will be most appropriate if the plants are partially illuminated with natural light (a vegetable garden on a windowsill, a balcony, a loggia, when direct sunlight does not fall at all or for a couple of hours a day - then the plants receive mostly blue rays from the sky, and the red ones are catastrophic It is not enough, as well as the general intensity of light. Here the red LEDs will fill the existing gap as well as possible. Only these should be highly efficient LEDs with a radiation wavelength of 660nM and better if they are CREE Photo red. Well, that's it, I went to order diodes!
To begin with, let's look at the exact efficiency of the LEDs discussed in the last part. Earlier, I took the data mainly from the iva2000 article , without checking, because there was considered more about the effectiveness of photosynthesis when illuminated by light of a different spectrum. Now I decided to understand the overall efficiency.
Consider the LEDs of the company CREE, because on the one hand, they are by far the most advanced in technology and, consequently, the light output per unit of power, and on the other, all their indicators are stable and well documented (in contrast to manufacturers' nounaym). Here, this company should pay me for advertising, but alas, I do not write with their submission, but simply because it is easier and more accessible.
So, what are we going to explore the LEDs? I will not post here the whole process of studying and selecting specific series in order not to flood the material with “water”. In short, I will say that I took the most powerful and at the same time the most effective chips, provided that it is freely available and at a bargain price. According to these criteria, two types are suitable: whites will be from the XM-L series.
- These are 10-watt chips with an efficiency of 158 lm / W (but not at maximum power, but only at 1 W). Cold white (6000-6500K), neutral white (4000-4500K) and warm white (3000-3500K).And red from the series XP-E, High Efficiency Photo Red 650-670nM.
Links to documentation on LEDs at the end of the article.
We will deal with white. Last time, the difference in efficiency of white LEDs was not taken into account and efficiency was evaluated only in relation to the McCree photosynthetic activity curve.
This time I decided to clarify this question more thoroughly. Unfortunately in the documentation for the LEDs never lead to efficiency, and write lumens per watt, so I had to do the reverse calculation. According to the spectrum of the LED and the photopic curve, it is calculated how many lumens the LED would have if its efficiency was 100%, and then the number of real lumens taken from the LED documentation is divided by this number. And that's what we got for the three types of white LEDs:
From left to right: cool white, neutral white and warm white.
It is noteworthy that in spite of the growth of lumens during the transition from the cold-white to the warm-white spectrum (with the same radiation power), the tabular values of lm / W and the overall efficiency of the LED fall very significantly from 40 to 23%. The fact is that the phosphor, of which there is much more warm white light in the LED, itself has not 100% efficiency, and, apparently, with its large amount it has a shading effect (the rays emitted by the lower layers are absorbed above lying and disappear ). At the same time, the indicator lumen per watt is used at a current of 2A (out of a maximum of three) - it can be seen that at the same time it drops from 140 at 350mA to 108 (for cold white). There is no such table in the Cree document - absolute lumens are given for a given current, and the power must be calculated using data from the graph of the current-voltage characteristic. Here are the relevant data from the datasheet:
Now let's deal with the red.
Everything is a bit easier with them, because the luminous flux is not specified in luminas, but in milliwatts. It is enough to divide milliwatts of radiation by watts of consumption and obtain efficiency with high accuracy! On all the LEDs would bring this data - 2/3 of the work could not be done!
And here we immediately make an amazing discovery - that the efficiency of these LEDs is 50%, and (one more graph, I don’t give here), unlike blue / white crystals, the luminous flux increases linearly with current and the efficiency of the chip does not fall! But when the chip overheats, the drop is much more significant than that of the blue chips. For comparison, in pure blue efficiency under the same conditions 48% (compare with this indicator for whites - higher). But the "just red" is much worse. Their efficiency turned out to be somewhere around 19%, and as the temperature rises, the luminous flux drops even faster than that of Photo red.
Interesting uses of individual LEDs and their combinations are already emerging. Now we will recalculate the efficiency table taking into account the newly obtained data.
It is seen that the red Photo-red with a large margin ahead of all. But it is impossible to cover with pure red, therefore it is necessary to combine and here there are options with white and blue. Immediately, we note (I thought everything, but threw away what was not promising) a combination of warm white and red. The low efficiency of warm white LEDs negates all the benefits of red. But cold white is very good in this combination! Themselves have a good efficiency, still enhanced by red LEDs, and the lack of the red spectrum is also covered by them. It also looks good combination of red and blue. Then they go simply cold-white and DNaT 1000, and the rest, in fact, do not pull. Well, let's see how it will look in full - with the drivers.
Further, the calculation logic went on the assumption that we want to receive more photosynthetically active radiation for the same money, therefore all figures, including the prices for LEDs and drivers, are reduced to the total phytoactive radiation of the lamp 100 µmol / s.
Color marking as in the previous table - to make it easier to understand where are the LEDs and do not take the place of duplicate headings.
But this is only the price at the start - how much money you need to invest in order to get a light bulb at 100 µmol / s. This is not enough - you need to see how much it will cost during operation. And if you count on this, and the cost of electricity in time - that's when you get the full picture, which I present for all to see!
Left for history, updated below
Thanks to the close attention of the commentators, it turned out that not all the LEDs that sell on the aliexpress called CREE are in fact. The cheapest of them, about one and a half dollars per 10-watt diode or less likely, are fake chips made by the Chinese company LatticeBright, which are several times cheaper than the original ones and, unfortunately, have about 2 times the worst performance. In this regard, I conducted a search for the prices of the corresponding LEDs in the Compal company, which is the official distributor of cree in the Russian Federation. Prices there are much higher than in China, but small wholesale is quite profitable, including in comparison with foreign suppliers.
And along the way I corrected two points - I added the replacement of lamps once a year for the HPS curve. And corrected the error (my oversight), due to which the price of all the lamps was considered to be of the same power (100W), whereas the original idea was based on the unit of photoactive radiation. In the new graph, these prices are for a lamp emitting 100 µmol / s, not 100W. I apologize for the mistake.
How to understand this bundle of twigs?
On the left - the price of the lamp at the start. I remind you that all of them will produce the same amount of phytoactive radiation, but have a different spectrum. The lower the strip begins, the cheaper the set. On the X axis we have months. It is assumed that the lamp works 12 hours a day, 7 days a week, for a total of 36 months, i.e. 3 years. This is just a little more than 13 thousand hours, and 50 thousand are stated for LEDs. And if everything is done correctly with cooling, and a current of 0.7 from the maximum is applied to the LEDs (so there is more efficiency by a full third), then they will work even more i.e. more than 10 years with almost no degradation.
The more horizontally the line goes - the more efficiency the luminaire has. We see that many lines start higher (more expensive chips), but with time they turn out to be cheaper than cheaper counterparts. This is an indicative line for photo red LEDs - it has the smallest slope.
The most amazing thing is that the cheapest are now ... The most expensive photo red LEDs! This is because they have the highest efficiency and the most "easily digestible" spectrum - they need the least in the beginning and they spend the least electricity in the future! Of great interest are the combinations "Cold White + Red Photo Red". This graph shows the curve when the ratio of white: red as 2: 1 in power. And just "cold white." These three lines diverge like a fan, where the extreme ones are white and red LEDs, and the middle one is their combination. For growing plants all components of the spectrum are necessary, but in different combinations. It turns out that all options for combinations of spectra are most effectively covered by just one combination - cold-white and red LEDs (but in a different numerical ratio).
It is worth noting that the combination of blue + red, although it has a smaller inclination than white + red, but gives a significantly worse price / light output, so the combination of white + red does not catch up with even 3 years. In the 10-year term, it may be preferable, but this is an exceptional case.
Phytolamp is not so cheap. If we take into account its efficiency, it is even more expensive than cold-white LEDs, and only in the future ... Money for electricity to the wind ...
DTA and at the beginning is not very cheap (I was surprised at how much electronic ballasts cost to them, but EM PRA should not be taken - they have low efficiency, the lamp due to flicker - also, they also buzz and heat like a stove) and do not catch up with time - especially with the replacement of lamps - which will have to do at least once a year, which is displayed as steps on the chart. So go to the garden.
Here is the spectrum of the combination of white and red LEDs superimposed on the MkCree curve (4: 1 in power, 2: 1 did not alter):
Of course it is wrong to judge such things based on the beautiful charts, but given the numbers that say the same thing - in my opinion the graph is almost perfect in terms of covering the spectrum of the photosynthetically active range.
The conclusion remains the same - buy cool white LEDs and red CREE Photo red and you will have a lot of light for your plants and savings for your wallet!
Lighting with pure red LEDs is also possible; one of the commentators wrote about this experience. This will be most appropriate if the plants are partially illuminated with natural light (a vegetable garden on a windowsill, a balcony, a loggia, when direct sunlight does not fall at all or for a couple of hours a day - then the plants receive mostly blue rays from the sky, and the red ones are catastrophic It is not enough, as well as the general intensity of light. Here the red LEDs will fill the existing gap as well as possible. Only these should be highly efficient LEDs with a radiation wavelength of 660nM and better if they are CREE Photo red. Well, that's it, I went to order diodes!
Source: https://habr.com/ru/post/410459/