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

[rives]

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).

That is the second mention that I've seen from you about KNNA's "messing up". While it is perfectly possible that he made an error in his calculations, if you are going to stand on someone's shoulders while pointing out deficiencies in their work, it would probably be a good idea to either offer proof of the charge or state it as an opinion.

 

[Beta Test Team]

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!

They didn't test the luminaire, that's the whole point of this thread:

1) photosynthetic efficiency (umol/J within 400-700 nm)
2) irradiance uniformity (which has not been tested, and is very poor for most LED luminaires)

Thanks for being cool about this, really, thanks. You're being nice and I do appreciate that. I'm honestly glad you're happy with your LEDs, I'm sure it's great. And if you do get your luminaire tested please come back and post.

Happy holidays

 

[Beta Test Team]

That is the second mention that I've seen from you about KNNA's "messing up". While it is perfectly possible that he made an error in his calculations, if you are going to stand on someone's shoulders while pointing out deficiencies in their work, it would probably be a good idea to either offer proof of the charge or state it as an opinion.

What? This is NOT on-topic, you should have PM'd me this. I am not happy that you posted this (I expected more from you).

And we're not standing on knna's shoulder, if anything, we're standing on shoulders of Ivo Busko's (ca. 1999; a Ph.D. astrophysicist), and Sager, Smith, and Edwards (1988). Knna seems to have gotten his work from to start with from Dr. Busko (knna didn't come up with it on his own).

We have been in contact with Dr. Busko, about his work, just so you know.

Our work uses very little of knna's work, and in fact, we are fixing errors he made (they're called "bugs"), and it's the same thing as messing up. All work the size and scope of what knna did will have bugs, pointing them out isn't dissing the author.

Knna did a great job and raised the bar for Cannabis growers in many ways, he was far ahead of his time, but there are a few important flaws in his work. The biggest is not correctly calculating RQE for YPF, which is the final goal of his work. You want proof? Here:

-- He didn't divide K.J. McCree's action spectrum by Cannabis absorptance spectrum to find RQE when multiplying by umol/s to find YPF, he multiplied McCree's RQE by Cannabis absorptance spectrum by umol/s to find YPF.

 

[rives]

What? This is NOT on topic, you should have PM'd me this. I am not happy that you posted this (I expected more from you).

And we're not standing on knna's shoulder, if anything, we're standing on Dr. Ivo Busko's shoulder (Ph.d. astrophysicist), which is where knna got his work from to start with (knna didn't come up with it on his own). Maybe you didn't know that...

We have been in contact with Dr. Busko, about his work, just so you know.

Our work uses very little of knna's work, and in fact, we are fixing errors he made (they're called "bugs"), and it's the same thing as messing up. All work the size and scope of what knna did will have bugs, pointing them out isn't dissing the author.

Knna did a great job, but there are a few important flaws in his work. The biggest is not correctly calculating RQE, which is the final goal of his work. You want proof?! Here:
He didn't multiply Dr. McCree's action spectrum by Cannabis absorptance spectrum, he multiplied Dr. McCree's RQE by Cannabis absorptance spectrum. (And I'm not explaining it further, if you don't understand why that's wrong.)

Your happiness with the post is irrelevant. You brought the subject up (as quoted), so I'd say that probably makes it meet the very elusive nature of being "on topic" in this thread.

 

[grouchy]

How can you state as a fact that leds are inferior to "best of breed" hids when you leave out all the current, new led technology that is now commercially available. It seems your are leaving out (declaring off topic) anything that suggests otherwise.

 

[GodsCronik]

Your happiness with the post is irrelevant.

That is sig worthy.

 

[Beta Test Team]

How can you state as a fact that leds are inferior to "best of breed" hids when you leave out all the current, new led technology that is now commercially available. It seems your are leaving out (declaring off topic) anything that suggests otherwise.

No one stated they are inferior as a rule, just those tested were in terms of efficiency and irradiance uniformity. That's the whole point of this thread: testing complete commercial luminaires in the real-world.

If anyone wants to post about the data from newer LED or LES luminaires that's great, I would love that. BUT, the data needs to be from an integrating sphere, not math, to compare apples to apples.

If you, or anyone else, wants to start a thread where you compare theoretical efficiencies (using math) go right ahead, but not in this thread. Because this thread is for real-world analyses, not theoretical output.

So, I am not declaring things off-topic as long as they're not off-topic. Get the luminaires tested and post the data, the more the merrier.

What I'm saying is if people want to add data to this thread, then add correct data as it relates to this thread. Don't just add any data you want to add, as that's not only rude, it's also off-topic.

 

[grouchy]

No one stated they are inferior as a rule, just those tested were in terms of efficiency and irradiance uniformity. That's the whole point of this thread: testing complete commercial luminaires in the real-world.

If anyone wants to post about the data from newer LED or LES luminaires that's great, I would love that. BUT, the data needs to be from an integrating sphere, not math, to compare apples to apples.

If you, or anyone else, wants to start a thread where you compare theoretical efficiencies (using math) go right ahead, but not in this thread. Because this thread is for real-world analyses, not theoretical output.

So, I am not declaring things off-topic as long as they're not off-topic. Get the luminaires tested and post the data, the more the merrier.

What I'm saying is if people want to add data to this thread, then add correct data as it relates to this thread. Don't just add any data you want to add, as that's not only rude, it's also off-topic.

You have stated that they are of no benefit to commercial applications and did not indicate it was an opinion. You are leaving out large amounts of information (newer leds on the commercial market) when you are making these statements and when confronted you brush it off because they have not been tested to your standards. It seems like your tests are out of date and it is decreasing the validity of what you are stating as fact.

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).

The point of this thread is supposed to be the truth. It seems not all facts are being ultilized to form your basis of truth.

 

[Beta Test Team]

Take this or leave this, I won't respond to you again. And I won't respond to anyone else who makes similarly flawed arguments:

I already told you I would be very happy to get new data to compare. And as I've written a few times already, if new data shows LEDs or LESs are better than HID, then great. I don't care either way. What I DO care about is accurate comparisons (so, the same tests useful to plant growth), which thus far has never happened for LED vs. HID, even though there are tons of threads like "LED vs HPS watt for watt."

You're conflating issues or maybe confused. That second quote you cited was me writing about the uniformity of the footprint, and the efficiency. And that still holds true as it relates to this thread (like I wrote in that quote), because we can only compare data apples to apples (and we don't have data for other luminaires).

Furthermore, there's no evidence to suggest LED or LES luminaires have better (or even close to the same) uniformity as compared to HID luminaires (with good reflectors), which is the main reason I (and three other scientists, one of whom as a Ph.D.) wrote LED luminaires not well suited for large canopies.

Many times I have written and agreed that in the future, LED or LES may be better than HID. And that future may be now as far as I know, but I can't assume it's now, like you are assuming it's now - because there is no data for newer (or otherwise non-tested) LED, LES, and HID luminaires.

That study was published 5 months ago, so I wouldn't call "outdated" in terms of the luminaires they studied. And all of those LED luminares are good models, just like all of the HID luminres are good models.

Its obvious many people in this forum want LEDs to win, which is fine if they do, but until someone gets them tested (like say, the manufacturer via. 3rd party) how could I include them in what I'm writing? That's my point to you and others: newer luminaires, LED, LES, and HID, are not included in the study not due to nefarious reasons, but simply due to lack data to compare.

So again: To you, and anyone else reading, if you want to add info to this thread about LED or HID luminaires not listed (and please do!), get them tested and then post the info.

 

[SupraSPL]

BTT, maybe you have some insight on some of the things I have been trying to make sense of in the study. The first thing that jumps out at me, what could explain the very large difference between the Sunlight Supply 1000W and the Gavita 1000W? Their results claiming that the Gavita system makes 150% as much light as the Sunlight Supply using the same power. Gavitas own study shows a much closer performance from bulb to bulb.

 

[SupraSPL]

The other thing that has me wondering, were the LEDs warmed up to thermal stability before the measurements were taken? What was the ambient temp, was there any airflow etc. It is not mentioned in the materials and methods.

Check out the design of the BML, it is passive cooled and using Osram reds and deep reds. Red chips are notorious for temp droop and the BML lamp is using passive cooling in an insufficient manner. So if they tested it cool, it is grossly overstating the capability of the BML. If they did warm it up is is grossly understating the capabilities of a well designed LED lamp.

I recently tested a commercial LED lamp using white COBs (which stand up very well to heat) and the output decreased by 19% from cold to warm. 5% due to the drivers warming up/Vf shift and 14% due to temp droop. For reference, the LED lamps I use lose less than 1% from cold to warm. So clearly that is an important variable.

 

[Beta Test Team]

Yes, those are good questions. Thanks for asking. I wish I had better answers. You could email the authors (probably better contacting Nelson than Bugbee) and ask them; I bet you'd get a response (if you do please post responses here).

In terms of warmed up pre-testing I would assume yes (but not sure), because they did the proper 100 hour burn-in for the HID lamps (I believe - if so it stands to reason they'd do pre-testing warm up for LED).

In terms of air flow for photosynthetic efficiency testing, they tested commercial units plug and play style, and they were placed in an integrating sphere for the efficiency measurement, so it wouldn't take long and air flow would not happen (no air ports).

When referring to uniformity testing, it was ambient room temp (I assume around 70'F to 75'F) in a large open area, and uniformity testing wouldn't be affected by these issues, anyway (not to the same degree as photosynthetic efficiency, at least).

 

[OIBI]

Where exactly would someone send a light to get tested?

 

[Beta Test Team]

Where exactly would someone send a light to get tested?

Great question, and here you go. This is for photosynthetic efficiency testing (umol/J from 400-700 nm):

TÜV SÜD America
http://www.tuv-sud-america.com/us-en

When testing ask that they use the IES LM79-08 measurement standard*, which is same as used in the study, and in fact, that's where the authors got all those luminaires tested for the study.

I don't know the cost, but it shouldn't be more than a couple hundred dollars per test (if that much).

* http://asmetec-shop.de/images/infobox/IESNA-LM-79-08_EN.pdf

 

[Beta Test Team]

The other thing that has me wondering, were the LEDs warmed up to thermal stability before the measurements were taken? What was the ambient temp, was there any airflow etc. It is not mentioned in the materials and methods.

I was thinking about your questions last night, and it occurred to me the answers may be in the testing protocol, found here: http://asmetec-shop.de/images/infobox/IESNA-LM-79-08_EN.pdf

Or, you can email or call the company that did the testing for the study (the authors didn't carry out the photosynthetic efficiency testing), and ask them directly: http://www.tuv-sud-america.com/us-en

 

[SupraSPL]

Thanks BTT, I will check into it.

BTW you were right, 4.5umol/s for my vegging lamp was (has to be) incorrect. It was not Mr Flux mistake, it was my hasty calculation and lack of experience working with umol numbers, I usually work with PAR W. I realized the discrepancy when I saw table 1 of the study you posted, they list the cool white with a maximum of 4.56umol/s. I looked more closely at Mr Flux work and realized his max for the CXA3590 5000K is 4.6 so my vegging lamp should be approx 2.85umol/s. The actual number will vary of course, but not by much. It has been tested by a mechanical engineer friend of mine but it was just a quick look, non official. I will see what I can do about getting some solid PPF and or PPFD data.

 

[Beta Test Team]

Sounds good.

I looked at some of Mr. Flux's posts and am intrigued. It would be nice to get his Python script, does he share the script? (I'm not very good with Python, but can manage.) I bet Mr. Flux would like to get our data sets for the various action spectra and RQE, so he can calculate not only YPF but all kinds of other values, as well. If you know him it may be worth PMing him these links:

"Action spectra and RQEs weighted with Cannabis absorptance"
https://www.icmag.com/ic/showthread.php?t=295933

"2006 CIE luminous efficiency functions & PC (1 nm step size)"
https://www.icmag.com/ic/showthread.php?t=298428

(Here are graphs of all those RQEs:
https://www.icmag.com/ic/showthread.php?p=6718314)

(Here are graphs of Cannabis abosrptance spectra:
https://www.icmag.com/ic/showthread.php?t=293045)


Oh yea, just to clarify: PPF and PPFD are the same thing (umol/m2/s from 400-700 nm), and in fact, PPF is preferred to PPFD. They both are the true definition of "PAR." PPFD was popular in the late 20th century, but fell out of favor, though it's still used and is meant to mean the same thing as PPF. The definition of "PAR" were researched and created by McCree ca. 1975, and accepted by the plant science community thereafter.

To write about area-less PPF it's often best to write something like "photosynthetic umol/s," "umol/s within PAR range" or "radiant PPF." Note that many companies, including big one like Philips, misuse the term "PPF" by using it to mean "umol/s within PAR range" (from an integrating sphere, like lumens), which isn't the same thing as the true definition of PPF (which is umol/m-2/s-1 from 400-700 nm).

 

[Beta Test Team]

Where exactly would someone send a light to get tested?

Hey OIBI, aside from the comapny I already listed, I have spoken with a different company about testing luminaire efficiency as defined in this thread (using NIST cert. method). Their rate is about $285 per luminaire, for 360-830 nm waveband range. Spectral units from the integrating sphere can be radiometric, as milliwatt/nm, which can be converted into quantum units (umol/s/nm) easily, then PAR range values can be integrated to find photosynthetic umol/s per joule.

The company is ITL, Inc., out of Colorado.

Personally I think every luminare manufacture should have their units tested in this way, especially LED and LES luminaries, and share those data with their customers.

 

[Beta Test Team]

An interesting metric I came up with the other day that uses data from this thread is what I called reflector 'photosynthetic radiation efficiency.' This provides the relative percentage of PAR range umol/s that exit the reflector as compared to the PAR range umol/s that is emitted by the lamp.

Just divide the reflector's umol/s per joule in PAR range by the lamp's umol/s per joule in PAR range, and multiply by 100. The lamp's total output is often reported by the manufacturer, as total PPF, or PPF per watt.

I was a bit surprised to learn, for example, Gavita PRO DE 1000W HPS reflector has a photosynthetic radiation efficiency of about 81%, that is, about 19% of radiant PPF emitted by the lamp is absorbed by the reflector/fixture.
https://www.icmag.com/ic/showpost.php?p=6738657&postcount=6837

 

[sprinkl]

got any evidence to prove a hps is down to 70% after 3 grows or 1 year ?

most industry lighting is changing to led technologhy ,,i recently had a job changing all the streetlights in Birmingham uk to leds ,the efficency is amazing ,,lots of shops and factories also switching over ,,

the horti industry is lagging i feel

They are still using really old high wattage street lights around here, anything produced in the last 10 years would be a great update.
LEDs are a great technology but not yet the end-all light technology many claim it is. Commercial leds for in the house are not much more efficient than CFL. If you calculate the extra cost of the leds towards CFL it takes years to win back the little energy saved. Are you going to replace 8W CFL with 4W led? The current gen leds save more energy but the increased cost is absurd. Too much can go wrong or they could burn out before you get the returns.

If you look at the graph it would still be more efficient to replace bulbs yearly than to go with a new led panel each 5 years.