Is it possible for a 63W LED lamp to produce 58W of actual light?

This debate started in VerdantGreen's grow log, and to not hijack his nice thread the arguing continues here.

So here's how it started:

LEDGirl said:

For anyone wanting the "nitty gritty" as far as calculations go, here it is: The 1st generation 63W has an actual power draw of 60-63W from the wall, with 5-6W being consumed by the cooling fans. In all actuality, it creates 54-58W of actual light using 63 LED's, which makes this a 1.0 - 1.1 gram per watt grow. For such a low wattage though, I think you did great! Can't wait to see the 205W :friends:

Click to expand...

pinstripe said:

I beg to differ. Cooling the LEDs is essential to maintain a high light output. Higher temperature means noticably less light. There's a reason why those fans sit there

On the other hand, what you can adjust for is the power used up by the power supply, as HID growers never account for the power used by the ballast. If I remember this right, a 250W HPS and ballast uses about 270-290W all together.

Also I don't believe it creates 54-58W of actual light. That would mean the LEDs turn 100% of the input energy into light.
Depending on what LEDs that sit in the lamp, 15-40% of the input energy comes out as light. For this lamp that would be somewhere between 9W and 23W of actual light.


All the best

Click to expand...

LEDGirl said:

I have no idea where you seem to be gathering your information from, but Red 1W LED's consume approximately .8W of power, and blues and whites consume approximately 1.2W of power. That power is then turned into light at an exact lumen per watt ratio. So the 1W (.8W power) Red LED still emits 45 lumens, not 15-40% of that amount. Likewise, the 1W white (operating at 1.2W) produces 110 lumens, not 15-40% of that amount. So I'm sorry to say it, but you are incorrect. Try and find a MH or HPS bulb that emits 45 lumens of 640nm per watt, you won't find it.

So to correct you, our light produces 54-58W of actual light. This is after you factor in the power consumption of all LED's. The power supplies are over 90% efficient as well, so they don't "drain" as much power as you'd like to think. Lastly, regardless of what the fans are used for, they are not a part of the "light" wattage. That would be like saying that the 4" inline fan you run for your 400W HPS is included in the overall light wattage... It just doesn't work that way. It would also be like saying that an A/C unit is a part of the light wattage, because it's essential to keeping your plants from dying with HID. Again, it doesn't work that way... Anyhow, thanks for your input.

Click to expand...

The text below is also me quoting myself. Didn't quote it because quotes on quotes are hard to read.
-----------------------------------------------------
Originally Posted by pinstripe
-----------------------------------------------------

LEDGirl said:

I have no idea where you seem to be gathering your information from, but Red 1W LED's consume approximately .8W of power, and blues and whites consume approximately 1.2W of power. That power is then turned into light at an exact lumen per watt ratio. So the 1W (.8W power) Red LED still emits 45 lumens, not 15-40% of that amount. [..] So I'm sorry to say it, but you are incorrect.

Click to expand...

Think we need to hit the physics books here
Sorry, but try converting 45lm of 640nm light to watts, and you will find it is not the 0.8W the LED uses. It is somewhere between 15% and 40% of 0.8W.
The 15-40% figure is a range of common radiant efficiencies of LEDs, meaning how much of the input power that the LED converts into light power.

LEDGirl said:

So to correct you, our light produces 54-58W of actual light. This is after you factor in the power consumption of all LED's.

Click to expand...

If this unit we're talking about has a power draw of say 58W, it certainly does not emit 58W of actual light. There is no such thing as a 100% efficient conversion between electrical energy and light. I literary would love to be proven wrong on this one, but unfortunately I believe you have misunderstood something here.

Here's how it works with common LEDs: 15-40% of the power turns into light, and at the same time 85-60% of the power turns into heat.

LEDGirl said:

The power supplies are over 90% efficient as well, so they don't "drain" as much power as you'd like to think

Click to expand...

HID and T5/PLL growers don't account for the power draw of the driver, then why should LED growers do that? That would make gpw data irrelevant for comparing HIDs and LEDs.

LEDGirl said:

Lastly, regardless of what the fans are used for, they are not a part of the "light" wattage. That would be like saying that the 4" inline fan you run for your 400W HPS is included in the overall light wattage... It just doesn't work that way. It would also be like saying that an A/C unit is a part of the light wattage, because it's essential to keeping your plants from dying with HID. Again, it doesn't work that way... Anyhow, thanks for your input.

Click to expand...

I think you misunderstood me. With LED tech specs you often get a graph with light output vs. temperature. As temperature increases, light output decreases. That's why you cool the LEDs. Yes?
If you'd remove those fans, significantly less light would be emitted from the unit. The lamp would perform much worse or maybe not even at all if it weren't for those fans.
Instead of only using a big heatsink, we made the choice to mount the LEDs on a metal sheet and cool this with fans. The lamp do not work properly without the fans, and therefore IMO it is incorrect to say the fans aren't part of the light wattage.


Also, you know, no matter if it's LED or HID, lamps of equal wattage emit equal amounts of heat. A big deal of the HID heat spreads quickly as radiant heat (aka IR light.) That is one of the reasons why HPS setups are harder to cool. Physics, baby


You're welcome for the tech and physics lesson
All the best

-------------------------
End of quote
-------------------------

I didn't know light was measured in watts, so I guess I don't understand the question. Isn't watts only relevant when you are comparing the same type of light to the same type of light? I.e. a 100 watt incandescent vs. a 100 watt HPS are two very different things for our application.

So what are you talking about when you say 'Is it possible for a 63W LED lamp to produce 58W of actual light?'

What is a watt of light? I'm not being argumentative, I'm just curious.

stink.whistler said:

I didn't know light was measured in watts, so I guess I don't understand the question. Isn't watts only relevant when you are comparing the same type of light to the same type of light? I.e. a 100 watt incandescent vs. a 100 watt HPS are two very different things for our application.

So what are you talking about when you say 'Is it possible for a 63W LED lamp to produce 58W of actual light?'

What is a watt of light? I'm not being argumentative, I'm just curious.

Click to expand...

A watt is an amount of energy converted at a given moment, usually in electronics it's as heat, but when it isn't it is emitted as electromagnetic energy. Light is just electromagnetic energy that fits into your eyeball and interacts with the antennas in your retina (rods, cones, whatever they're tuned antennas for light wavelengths.)

The problem with using lumens is it isn't a linear scale, its equalized for human vision, so power at the edges of visual spectrum is distorted when given in lumens. It's not really great for comparing different color LEDs. Please read this page it does an okay job explaining:

http://ledsreview.com/blogs/Elie/17.php

The dot represents 1 watt rate of photon output.

Watts dissipated in the LED is different, it's how much power is being used by the devices, not how much is being emitted, but for objective comparison of systems for efficiency (the reason people are using LEDs in the first place), you need to know the total wattage the system is dissipating.

And to answer the main question in the OP, 58W into the LEDs from a 63W draw system is 92% efficient for the current regulation solution. If the system is a well designed switch mode current regulation circuit, it's possible.

If it's is not a switcher, it's a lie or it's a mistake or its just done wrong.

Also, where did the 63W and 58W numbers come from? What do they mean? I'm sorry if i missed the exact post but a lot of figures are thrown around in LED threads, and to me most of them seem pretty useless.

pinstripe said:

Also, you know, no matter if it's LED or HID, lamps of equal wattage emit equal amounts of heat. A big deal of the HID heat spreads quickly as radiant heat (aka IR light.) That is one of the reasons why HPS setups are harder to cool. Physics, baby

Click to expand...

They emit the same amount of absolute 'heat', but it wont affect temperature the same. The thermal resistance of the packaging, the area in which you dissipate the energy and the efficiency of the heatsink determine temperature.

While the heat dissipated may be the same, it doesn't have the same effect.

interesting, thanks for the link.

i believe LEDGirl said 54-58 watts if that helps.

V.

VerdantGreen said:

interesting, thanks for the link.

i believe LEDGirl said 54-58 watts if that helps.

V.

Click to expand...

That's 86%, vs the previous 92% with 58W, assuming the 63W is accurate. Still not bad. I think I've seen it posted that her panels use switch mode regulators, so 85%-95% is what should be expected.

pinstripe said:

HID and T5/PLL growers don't account for the power draw of the driver, then why should LED growers do that? That would make gpw data irrelevant for comparing HIDs and LEDs.

Click to expand...

Because they're doing it wrong =)

But in general I think it is a good idea to compare wattage across the LEDs vs standard HID ratings (250, 400, 600 and 1000W).

Efficiency in current regulation is a known and well researched aspect of this. 95% can be achieved right now, and slightly higher efficiency will prob become more common in the next few years.

Optimal LED spectrum is still up in the air, and once that is less of an unknown quantity, real LED vs HID efficiency comparisons can be made.

Right now I feel that LED vs LED considering spectrum is very important, which means accurate power dissipated by the LED strings (both as heat and EM energy) must be measured to compare systems.

renz said:

They emit the same amount of absolute 'heat', but it wont affect temperature the same. The thermal resistance of the packaging, the area in which you dissipate the energy and the efficiency of the heatsink determine temperature.

While the heat dissipated may be the same, it doesn't have the same effect.

Click to expand...

renz~ thanks for all the clarification.

Got a 'keepin it real' question for you...

How much heat will a 1k LED equivalent contribute to a canopy's temperature? Make whatever apples & oranges assumptions necessary to answer...


To offer my perspective... a long time ago invested in t5's as all the rage for very, very low heat profile w very efficient lumens/watts ratios. Okay, in the end... keepin it real... the t5 ballasts/bulbs still put out a lot of heat that needed to be vented/cooled. No escaping it, watts/lumens/electronics = heat.

This the deal with the LED's??? Thank you in advance!

BigTop said:

renz~ thanks for all the clarification.

Got a 'keepin it real' question for you...

How much heat will a 1k LED equivalent contribute to a canopy's temperature? Make whatever apples & oranges assumptions necessary to answer...


To offer my perspective... a long time ago invested in t5's as all the rage for very, very low heat profile w very efficient lumens/watts ratios. Okay, in the end... keepin it real... the t5 ballasts/bulbs still put out a lot of heat that needed to be vented/cooled. No escaping it, watts/lumens/electronics = heat.

This the deal with the LED's??? Thank you in advance!

Click to expand...

When I tried out an 300W LED panel it still gave of quite a bit of heat. Yes it had 6 fans on it and was cool to touch but the heat from the 300W power supply still has to be removed the panel and back into the room it goes.

pinstripe said:

Think we need to hit the physics books here
Sorry, but try converting 45lm of 640nm light to watts, and you will find it is not the 0.8W the LED uses. It is somewhere between 15% and 40% of 0.8W.
The 15-40% figure is a range of common radiant efficiencies of LEDs, meaning how much of the input power that the LED converts into light power.

If this unit we're talking about has a power draw of say 58W, it certainly does not emit 58W of actual light. There is no such thing as a 100% efficient conversion between electrical energy and light. I literary would love to be proven wrong on this one, but unfortunately I believe you have misunderstood something here.

Here's how it works with common LEDs: 15-40% of the power turns into light, and at the same time 85-60% of the power turns into heat.

Click to expand...

45lm at 640nm is almost 40% and as you stated this numbers - maybe you know some led's with this specs?? please let me know this looks promising (almost 2 µmol/s-1/watt)
if someone could provide the real spectral power distribution and lumen output of led grow lights we could tell real numbers about the output but i think no one have it :I

btw what you said is correct and when we talk about heat i can tell you that modern hid produce less heat/watt consumed than any led grow panel ive seen yet (future led panels should run cooler, cree just working on 65% efficient diodes), its just way harder to handle which gives advantage to led's

Hi there

Bit off-topic but I was wondering is there some realistic upper limit for led efficiency regarding physics? I'm dumb in that area but somehow I'd think that close to 100% efficiency would need superconductive materials etc. Feel free to correct these assumptions. Oh and that 65% efficiency sounds very promising.

unthing said:

Hi there

Bit off-topic but I was wondering is there some realistic upper limit for led efficiency regarding physics? I'm dumb in that area but somehow I'd think that close to 100% efficiency would need superconductive materials etc. Feel free to correct these assumptions. Oh and that 65% efficiency sounds very promising.

Click to expand...

Do you mean efficiency in turning electronic energy into optical energy? Because that's waaaaay low (my guess is maybe not even a few %?) but it gets better pretty consistently.

If you mean efficiency in the circuits to drive the LEDs with constant current from a constant voltage source, then yeah towards 100% efficiency is not unrealistic at all.

The world is built on switch mode regulators, your computer has a ton in the PSU and on the motherboard. It doesn't take superconductors; It just takes low on resistance FETs and maybe some schottky diodes, and some tricks with inductors. 95% is pretty typical for voltage and current regulators run in ideal circumstances, and in the next few years you can probably get towards 100% with improvements in FETs and controller IC.

Meant the first one, but I quess that one was bit too far ahead Thanks for the answer.

Edit:forget about that supeconductivity, baad example. Matter-antimatter light anyone . Bad spectrum thou..

unthing said:

Meant the first one, but I quess that one was bit too far ahead Thanks for the answer.

Click to expand...

http://ledengin.com/products/5wLZ/LZ1-00R205.pdf

Deep Red (660nm) are too far out of visible spectrum to market in lumens, so for this color (and some blue/UV LEDs) they will state energy output in mW.

So according to this datasheet: @ 1A, 25C, with a Vf of 3V, you get 525mW of optical energy for 3W electrical input.

Which is like 17.5% optical conversion efficiency. Dunno how typical this is for this frequency LED (LEDGirl would probably be the person to hit up for this info).

I've understood that 20% is pretty common for that nm, there may be better ones. And they're getting better.

How much is it for hids, around 40-50% at best? If generic 400 w (450w counting ballast losses)emits around 125-140 PAR watts so that's 30-40% efficiency in visible light, I'm a correct? Of course after that the ideal spectre and frequencies beyond 400-700nm area is another thing.

BigTop said:

renz~ thanks for all the clarification.


How much heat will a 1k LED equivalent contribute to a canopy's temperature? Make whatever apples & oranges assumptions necessary to answer...

Click to expand...

asde² said:

btw what you said is correct and when we talk about heat i can tell you that modern hid produce less heat/watt consumed than any led grow panel ive seen yet

Click to expand...

Yeah with HID, you have this single point heat source floating in the middle of a big air space, with most of the heat being reflected back at the bulb and at the canopy. Using heatsinks on the reflectors might help to move heat above the system as opposed to being focused down, but I doubt by much.

With LEDs, you just use a high efficiency heatsink, passive or active, with the bulk of the heatsink surface area pointing upwards, you can effectively move almost all the heat away from the canopy. The distributed heat sources of an LED array make this super easy to implement, however to do this at highest efficiency you need surface mount LED emitters on custom thin PCB mounted directly to heatsinks. I highly doubt standard through hole LEDs or even star mounted surface mount LEDs can approach this method's thermal transfer performance. Without fans, I'm positive this is the best way.

Heat management is basically the biggest advantage of LEDs besides tuned spectrums. You pick the temp you want them to run at then design the heatsink to achieve that. Converting power dissipation and thermal resistance numbers into temperatures given a C/W parameter for a heatsink is not difficult at all, and typically is pretty accurate with spread out heat sources like an LED array.

unthing said:

I've understood that 20% is pretty common for that nm, there may be better ones. And they're getting better.

How much is it for hids, around 40-50% at best? If generic 400 w (450w counting ballast losses)emits around 125-140 PAR watts so that's 30-40% efficiency in visible light, I'm a correct? Of course after that the ideal spectre and frequencies beyond 400-700nm area is another thing.

Click to expand...

Both those numbers sound right. As far as I understand it, in realistic conditions HID is like 3x-4x more efficient at electrical to optical conversion than current LEDs.

Unless you're using heroic status cooling on your array, LEDs will not perform at their datasheet specific 25C output rates.

Like around 75% of stated output is probably more common.

.