LED vs. HPS: Truth about photosynthetic eff. & uniformity

[DarkGreenGoo..]

Except:

a.) They are likely not going to give you a power savings vs. best-of-breed HID, maybe in a few years when the technology develops more, as this thread as pointed out many times:
https://www.icmag.com/ic/showpost.php?p=6679113&postcount=72

b.) The poor irradiance uniformity (footprint) of LEDs means they're not well suited for commercial grows with large canopies.

Did you take into account the $300-400 ballasts needed for the HID systems as well as all the extra fire protection needed? HID lights need lots of different fire protection devices or systems installed or you could get shut down by the fire-marshal.

It also doesn't seem to account for the cost of HVAC/cooling as well as Humidity in the room (As the heat produced by HID's often lowers humidity recuring a humidifier.

 

[Beta Test Team]

the reality is led aint happening for commercial growers only small scale or those with money to waste..end of story

There are a number of commercial grows running LED's now. As well as medical collectives.

There are also people jumping off of bridges. Shall we join them? (Data is what one should use, not other people's opinion, to make decisions.)

I'm not trying to be a jerk, just making a stark point. Just because someone else is doing it doesn't mean it's wise to do.

Right now large scale commercial (that is, very large grows) use of commercial LEDs offer little to no benefit over best-of-breed HID (as it related to this thread).

 

[Beta Test Team]

I hope this makes some of the mad people happy. I am going to ask the mod in this forum to rename this thread to: LED vs. HPS: Truth about photosynthetic eff. & uniformity

Here is what I added to the first post:

[NOTE #1: The original thread title ("LED vs. HPS: A reality check") was from the first PDF below. But a few LED growers got mad at the title, so to try to keep the thread on-topic I changed the title - but not the content]

[NOTE #2: I added the following sentence to this post to try and keep this thread on-topic: "This thread is about the photosynthetic efficiency of luminaries, and their irradiance uniformity, that is, converting watts into photosynthetic photons and the uniformity of the footprint."]

 

[NorPene Solutions]

Interesting info if you can weed through the fluff to get to the meat.

 

[Payaso]

I have 'cleaned' this thread of the off-topic posts and trolling by request of the original thread starter.

Off-topic posting will be removed, from wherever it exists on ICMag. It's just sensible, like running the vacuum cleaner around the house to get rid of the dirt and debris.

 

[Beta Test Team]

Thank you! (And sorry for having to ask, I tried to avoid asking.)

 

[flat9]

Ahhhhhhhhh. Pissing match be gone.

 

[flat9]

Did you take into account the $300-400 ballasts needed for the HID systems as well as all the extra fire protection needed? HID lights need lots of different fire protection devices or systems installed or you could get shut down by the fire-marshal.

It also doesn't seem to account for the cost of HVAC/cooling as well as Humidity in the room (As the heat produced by HID's often lowers humidity recuring a humidifier.

These are valid points, especially the humidity, but I think if you're running sealed rooms with CO2 you'll likely need to remove humidity rather than add it as the plants transpire. The A/C will lower it somewhat, but it all depends. Of those I know even in WA they have to remove humidity for a sealed room, not add it.

If you're running active exhaust and passive intake you'll definitely have to raise the humidity substantially w/ HPS, and LEDs are indeed easier to control humidity.

As for the ballasts, well I can find some nice double-ended HPS lamps, reflectors, and ballasts built in for about 450 USD. Similarly capable LED costs me about 850 (from China).

 

[SupraSPL]

Even if commercial COB or LED luminaires turn out to be more efficient in turning joules into photosynthetic umol than best-of-breed HPS (which isn't true right now), if the price doesn’t come down relative to umol output they will not be adopted by businesses with bottom-lines (ROI) as effected by how much it costs to produce each gram. Which includes things like yield efficiency, i.e. gram per kWh, which accounts for all manner of electrical appliances.

My point and your point seem to agree: the technology isn't commercially available yet, as a luminaire one can buy (rather than build).

The best performing LED lamp in the test was red/blue. Apparently there are some varieties of cannabis that can flower under R/B but it seems that white is very helpful for most varieties and it avoids the issue of color mixing. The downside is, white LEDs are made from blue LEDs, so whites can never be as efficient as blues. The upside is, blue LEDs can be amazingly efficient (70%+) and even higher in the lab, so some cool white COBs currently on the market are as high as 60% efficient at usable light levels. I am using a pair in my vegging space that are 5000K, 18W ea, 60% efficient (4.5umol/J), zero lumen depreciation.

We are starting to see a few commercial COB lamps coming to market. I have a sample from Optic Lighting using 4000K Bridgelux Vero 18s and their larger models are using the Vero 29. Unfortunately the sample I have is running too hot to be optimal, but they have the right idea and I estimate it is about 30% efficient even when running smokin hot (still less efficient than HPS). Also, they are using a glass lens that is surprisingly effective at directing the photons to the canopy.

 

[SupraSPL]

Pic of the glass lens cutoff line on my shed from a distance of about 30 feet. (the lamp draws 120W warmed up).

There are more commercial COB lamps that will be coming to market soon.

 

[Beta Test Team]

I am using a pair in my vegging space that are 5000K, 18W ea, 60% efficient (4.5umol/J), zero lumen depreciation.

Where did you get the 4.5 umol/J info from? (I strongly doubt there is no depreciation, but that's kind of off-topic, anyway.)

 

[SupraSPL]

The 5000K CXA3590 curve was analyzed/digitized by Mr Flux on RIU. He determined the LER as 324 and Photon flux 7.50 umol/s. So when running at 60% efficient (minimum figures used), we get 4.5 umol/s.

Because it is a 150W COB running at only 18W and is mounted on a CPU cooler, the Tj is not far above ambient, hence zero lumen depreciation and near zero temp droop. From what I understand, it may actually experience lumen appreciation. Also, the cost for this unit (for DIYers) is well within reason.

All that said, you are right it is off topic because there is nothing like that for sale in the commercial market. But it is an example of what COB technology may offer in the future, so relevant in that way.

Here is a CXA3070 3000K running at 150mA (5W, 203lm/W, 62.4% efficient, 3.93umol/s). In this case, the cost is probably not within reason. I run them at 800mA (28W, 161lm/W, 49.5%, 3.11umol/S) but this example is technically possible. Also, Cree has confirmed the LER for the 3000K CXA to be 325lm/W, so we are not relying on a digitized curve approximation in this case.

 

[SupraSPL]

More relevant to the topic at hand is the Vero 29 3000K. If we drive it at 2.55A, using a Mean Well driver and mounted on a CPU heatsink, we are looking at a total parts cost of about $78. Dissipation would be about 104W, 109.5lm/W, 34% efficient. If w were to use a generic driver we could get 2.45A for $22, bring parts cost down to $64 for 94W.

That works out to :
W Mean Well- $2.20/PAR W
W generic driver- $1.94/PAR W

That is something I suspect can really compete today with the best HIDs, if only a good manufacturer would build it.

600W HPS setup from HTG supply w digital ballast: $200
That works out to $0.93/PAR W. Pretty tough to beat that deal in terms of up front cost!

 

[Beta Test Team]

Supra, you're now driving this thread off-topic, please stop. If you want to write about what you're writing about, please make a new thread. This thread is about commercial luminaires someone can go out and buy.

I asked you how you got the efficiency datum, and you didn't answer, so please stop posting unless you're going to post on-topic. I have no idea who "Mr Flux" is, but the fact he uses LER means he doesn't know what he's doing (LER isn't for plants and is a useless metric for plants). And I doubt he properly converted lumen to umol, anyway, nor that he has accurate radiant energy data.

This thread is meant to go by provable facts about photosynthetic efficiency and irradiance uniformity, and "Mr Flux" isn't a valid citation. For example, the photosynthetic efficiency data I posted was from placing the luminaires inside an integrating sphere to calculate umol/s within PAR range.

I would be very surprised if that luminaire even had 2 umol/J photosynthetic efficiency, let alone 4.5 umol/J.

I also would disagree there is no energy deprecation from that luminaire. And it seems the way you're (or Mr. Flux is) calculating PAR W is wrong, too. But again, those subjects are off-topic.

Thanks for understanding.

 

[SupraSPL]

I am sorry you feel that way. I did just post information about a commercial luminaire that is using confirmed genuine Bridgelux Vero 18 COBs that is likely superior to the LED lamps used in this test when it comes to flowering cannabis. I think that is a very important issue to consider because it means the lamps they tested are already obsolete.

The rest of my posts demonstrate that this is just the tip of the iceberg when it comes to COBs. I am surprised you do not think that is relevant? We will be seeing even better commercial COB lamps in the market soon, maybe within days/weeks. I stand behind the numbers I posted. I can understand your skepticism regarding the efficiency figures, they are incredible. But I urge you to look into it yourself, check the Cree PDF for the CXA3590 5000K CD bin. Yes it really can create over 200lm/W at useful levels. 324 LER is typical for a neutral/cool white. This is not far fetched, it is all quite reasonable science.

Mr Flux is a well respected COB LED DIYer in the cannabis community. He used calculus to convert the manufacturer provided curves into LER. It is a PITA but it is not magic. His numbers agree closely with 2 others that have done the same work and Cree confirmed it to me personally (for the 2700K and 3000k CXA curves). LER is very important because it allows us to convert from lumens/W to PAR W/ppf/or umol/J for white LEDs. Also for the record, Mr Flux efforts helped us expose the weak radiant output of the high CRI LED and he deserves a lot of credit for that. Manufacturers really do read our work and it affects the development of commercial lamps. High CRI was in danger of becoming a fad and that would have resulted in less bud for us LED growers.

 

[Beta Test Team]

I am sorry you feel that way. I did just post information about a commercial luminaire that is using confirmed genuine Bridgelux Vero 18 COBs that is likely superior to the LED lamps used in this test when it comes to flowering cannabis. I think that is a very important issue to consider because it means the lamps they tested are already obsolete. I am surprised that does not seem relevant to you?

It's not relevant because it's not verifiable data, it's just something "Mr Flux" came up with. And this thread isn't about spectrum, other than PAR (which is defined as PPF).

Thank you for understanding.

You may not be aware, but this thread has had a lot of drama from people with totally wrong ideas and understandings. I don't want it go down the drain again.

Mr Flux is a well respected COB LED DIYer in the cannabis community. He used calculus to convert the manufacturer provided curves into LER. It is a PITA but it is not magic. His numbers agree closely with 2 others that have done the same work and Cree confirmed it to me personally (for the 2700K and 3000k CXA curves). LER is very important because it allows us to convert from lumens/W to PAR W/ppf/or umol/J for white LEDs.

This is the last I will write about this in this thread because it's very off-topic, so please don't keep this going or I will request the posts are removed:

I know well the math involved in converting lumens to umol, and I know you don't need to use LER to do it (and using LER is not a great idea); that's just knna's old method (and he also messed up in calculating RQE in his spreadsheet). It doesn't matter the 'color' of the LED when converting from lumen to umol, all that matters is the SPD (digitized) and photopic curve (i.e. CIE luminosity function from 2006).

So again, Mr Flux's data isn't reliable as compared to the data in this thread and therefore isn't on-topic. This thread seeks to only provide factual data from testing, not math.

To find umol/s from lumen, as well as umol/J (per second), from SPD of the radiation source you can use our soon-to-be-released HISSA ("Harvest Irradiance Spectral-System Analyzer") spreadsheet, which is a complete re-write of knna's spreadsheet with corrections (e.g. calculating RQE correctly), and only useful data (so no LER and no CRI and no kelvins, for example, because that's not useful to plants).

 

[SupraSPL]

BTT, I understand how this is a slippery slope regarding getting off topic and scientific misunderstandings. But I am presenting you with a scientific challenge that cannot possibly be more on topic. In other words, if we are going off topic then that probably needs to happen. The test that this thread is based on is incomplete/obsolete. Do you disagree?

Yes there are slight imperfections in the calculated and estimated LER. Mr Flux actually updated his numbers slightly and the white LEDs curves have changed significantly since KNNAs time. I am not suggesting that is the ideal way to arrive at ppf, but it is what we have got to work with for now. Manufacturers only give us mW output for deep blue and deep red LEDs. But the lumen data is far from useless and Cree has provided the LER. How can you ignore a 200lm/W output? This is not a green LED we are talking about. The LER is in the low 300s no matter which way you cut it. That means we have 60+% efficient COBs on our hands, for sale, right now. Awesome!

Rather than engineer toward maximum efficiency, commercial lamps are geared toward up front cost. The Vero 29 is the perfect tool for that job and it is already being used for commercial COB lamps. We will soon have access to flowering lamps with efficiency in the high 30's. That is great news for us growers! That is bad news for the HID industry.

 

[Beta Test Team]

BTT, I understand how this is a slippery slope regarding getting off topic and scientific misunderstandings. But I am presenting you with a scientific challenge that cannot possibly be more on topic. In other words, if we are going off topic then that probably needs to happen. The tests that this thread is based on is incomplete/obsolete. Do you disagree?

The tests in this thread are not incomplete/obsolete, are you kidding me? The photosynthetic efficiency data is by integrating sphere. They (two Ph.D. scientists in published research) just didn't include the luminaire you wanted them to include. So, get it tested like all the other luminaires in this thread and I'd be very happy to have you post the results. If they're really 4.5 umol/J than that's amazing and great, but until it's tested it's not a fact and it's off-topic.

Get that luminaire tested in an integrating sphere and the data is on topic, otherwise it's not. This thread is about comparing apples to apples, which means using the same quantification method (which isn't math).

Really, another post about this and I'm asking they get removed. Including mine.

If you want to write about this topic, please start a new thread.

 

[SupraSPL]

Luckily, Cree (engineers and PHDs galore) has already done that for us and they are brutally honest about the capabilities of their LEDs. The PDF says 200lm/W @ 270mA Tj 30C. That is good enough for my money. Have a nice day!

 

[positivity]

incoming bud alert...

You should try using good leds one day. Much more productive than crunching numbers from outdated lights.

Are big led buds off topic too?