Strips vs COBs

Hello,

It's been a while since I checked upon the latest development in growth LEDs, though I noticed that more an more people are experimenting with led Strips. Back in 2015, when I came up with the first citizen COB in this forum, it was all about single, powerful, under-driven chips to achieve maximal efficiency. The observed trend with Bridgelux, Cree and especially Citizen COBs was that the denser the COB array (more single LEDs packed into a large die), the better the efficiency when severely under-driven.

So my question is: is this still true ? Or can led strips now compete with large COBs in term of efficiency ? Is there a comprehensive study somewhere ?

While I do like COBs because the 20-23mm LES models fit very nicely into cheap chinese track light, powering lots of large chips rapidly creates a cooling headache and a cabling mess. Plus, for folks like me growing all kinds of stuff, they tend to concentrate the light too much, and burn plants.

It seems to me, that long arrays of small chips offer better cooling characteristics (you can go almost all passive) and are way easier to install, because there are so many good quality aluminium bars available with awesome cooling performance. I'm actually dreaming of a false ceiling made of 3m long alu bars populated with led strips. Of course, this only makes sense if the overall chip efficiency of strip led can match high quality cobs...

cheap chinese track light....20-23 mm ?

OLD MOTHER SATIVA said:

cheap chinese track light....20-23 mm ?

Click to expand...

Yes, just search some of my posts. Here you can find some photos of my current track lights: https://www.rollitup.org/t/heatsinks-for-diy-led-lamps.851907/page-62#post-12015456
The aluminium heatsink offers 23mm diameter surface to attach the cob, which fits exactly even cree cx3070s. Though the citizen clu046/48 18 is a better fit.

qwerkus said:

So my question is: is this still true ? Or can led strips now compete with large COBs in term of efficiency ? Is there a comprehensive study somewhere ?

Click to expand...

they always could, its just a matter of cost

tech is the same (blue die w/phosphor)

qwerkus said:

It seems to me, that long arrays of small chips offer better cooling characteristics (you can go almost all passive) and are way easier to install, because there are so many good quality aluminium bars available with awesome cooling performance. I'm actually dreaming of a false ceiling made of 3m long alu bars populated with led strips. Of course, this only makes sense if the overall chip efficiency of strip led can match high quality cobs...

Click to expand...

a lot of people agree with you.

for example a 288-die QB is about equal to 4 small cobs like citi 1212 in efficiency

at 150-200W either needs cooling and after cooling the cost comes out to be the same

while in theory the qb would spread light more evenly than the cobs you are still working with a fixture of finite size vs 4 cobs which can be spread out so its kind of a wash on coverage

strips kinda get the best of both worlds (easy to cool dies but spread out)

theres some really great strip builds on here @OLD MOTHER SATIVA etc

The only reason to build with cobs over strips is if you are a spectrum junkie and you dont want to mess around with supplementing a 3000k strip build.

Other than that strips are, watt for watt... cheaper, just as efficient if not moreso, easier to install, easier to manage heat, much lighter fixtures, more uniform ppfd, better "penetration". They dont require heavy heatsinks, thermal compound, nor holders.

You can order strips, a driver, kapton tape, some wire and a cookie sheet and you've got a light.

When I say strips I mean samsung strips (h-series and f-series) specifically not quantum boards.

qwerkus said:

Hello,

It's been a while since I checked upon the latest development in growth LEDs, though I noticed that more an more people are experimenting with led Strips. Back in 2015, when I came up with the first citizen COB in this forum, it was all about single, powerful, under-driven chips to achieve maximal efficiency. The observed trend with Bridgelux, Cree and especially Citizen COBs was that the denser the COB array (more single LEDs packed into a large die), the better the efficiency when severely under-driven.

So my question is: is this still true ? Or can led strips now compete with large COBs in term of efficiency ? Is there a comprehensive study somewhere ?

While I do like COBs because the 20-23mm LES models fit very nicely into cheap chinese track light, powering lots of large chips rapidly creates a cooling headache and a cabling mess. Plus, for folks like me growing all kinds of stuff, they tend to concentrate the light too much, and burn plants.

It seems to me, that long arrays of small chips offer better cooling characteristics (you can go almost all passive) and are way easier to install, because there are so many good quality aluminium bars available with awesome cooling performance. I'm actually dreaming of a false ceiling made of 3m long alu bars populated with led strips. Of course, this only makes sense if the overall chip efficiency of strip led can match high quality cobs...

Click to expand...

nogod_ said:

When I say strips I mean samsung strips (h-series and f-series) specifically not quantum boards.

Click to expand...

Interesting reply, even though I would love some spreadsheet to sustain your claim. Also, it leads to the manufacturers question: dunno about you guys, but to me, it looks like a jungle of led strips out there. It's really complicated to get your hands on some reliable documentation about efficiency, spectrum and so on, given the zillion of cheap ebay / aliexpress products flooding the market.

Did anyone compile a list of efficient and reliable strips ? I read some good stuff about samsung diodes and especially nichia's, but would hate to jump to one particuliar manufacturer too fast. Back in the cob days, it was quite hard to break the "cree/bridgelux hype", and open minds to other models, just as good and sometimes a lot cheaper. In fact, when it comes to led strips, it seems that each and every manufacturer has some to offer...

Strips are cheap enough (especially in quantity) that i wouldnt even bother rolling the dice on alibaba or ebay.

The specs for samsung's strips are on their website, likewise for the bridgelux EB series. Last time I checked Nichias strips were on par with the others' efficiency but more expensive /watt.

There are many threads on this subforum dedicated to strips, I created one to compile a list of good options. Others have build threads and grow journals.

qwerkus said:

Interesting reply, even though I would love some spreadsheet to sustain your claim. Also, it leads to the manufacturers question: dunno about you guys, but to me, it looks like a jungle of led strips out there. It's really complicated to get your hands on some reliable documentation about efficiency, spectrum and so on, given the zillion of cheap ebay / aliexpress products flooding the market.

Did anyone compile a list of efficient and reliable strips ? I read some good stuff about samsung diodes and especially nichia's, but would hate to jump to one particuliar manufacturer too fast. Back in the cob days, it was quite hard to break the "cree/bridgelux hype", and open minds to other models, just as good and sometimes a lot cheaper. In fact, when it comes to led strips, it seems that each and every manufacturer has some to offer...

Click to expand...

nogod_ said:

Strips are cheap enough (especially in quantity) that i wouldnt even bother rolling the dice on alibaba or ebay.

The specs for samsung's strips are on their website, likewise for the bridgelux EB series. Last time I checked Nichias strips were on par with the others' efficiency but more expensive /watt.

There are many threads on this subforum dedicated to strips, I created one to compile a list of good options. Others have build threads and grow journals.

Click to expand...

Do you have a link to the thread ?

WOOT: 187-206lm/w for the samsung h series ? Has that been verified empirically ?

qwerkus said:

Do you have a link to the thread ?

WOOT: 187-206lm/w for the samsung h series ? Has that been verified empirically ?

Click to expand...

Dude Cree inc jumped fucking almost 5 bucks a share today!!!!!!!!!!!!!!! They must be the best. Buy Cree and ill see at 45 a share in no time!!!!!!!!!

qwerkus said:

Did anyone compile a list of efficient and reliable strips ? I read some good stuff about samsung diodes and especially nichia's, but would hate to jump to one particuliar manufacturer too fast. Back in the cob days, it was quite hard to break the "cree/bridgelux hype", and open minds to other models, just as good and sometimes a lot cheaper. In fact, when it comes to led strips, it seems that each and every manufacturer has some to offer...

Click to expand...

I too am wondering this. I'm 1 month into my first grow. But i've been looking for info on different strips, here is a link to a thread about Samsung vs Bridgelux EB. I just built my first lights 2 days ago (4) COB's, but i'm already planning to add 2-4 led strips in between them. That way whichever is best I'll have awesome supplemental light also.

qwerkus said:

Do you have a link to the thread ?

WOOT: 187-206lm/w for the samsung h series ? Has that been verified empirically ?

Click to expand...

But it's how hard you drive them... I can prolly get well over 200lm/watt on my lights if I run them super low mA.

Budies 101 said:

But it's how hard you drive them... I can prolly get well over 200lm/watt on my lights if I run them super low mA.

Click to expand...

Yes, but than we are back to the days of expensive cxb 3590 severly underdriven = just too expensive, if you ask me. The astonishing thing about those samsung series, is that their efficiency is given for 40% of their max. capacity (240ma/600ma) - at least theoretically.

qwerkus said:

Yes, but than we are back to the days of expensive cxb 3590 severly underdriven = just too expensive, if you ask me. The astonishing thing about those samsung series, is that their efficiency is given for 40% of their max. capacity (240ma/600ma) - at least theoretically.

Click to expand...

qwerkus said:

Do you have a link to the thread ?

WOOT: 187-206lm/w for the samsung h series ? Has that been verified empirically ?

Click to expand...

Hey qwerkus, wellcome back!
The H- and F-Series use the same diode, LM561c A1/S6 bin(same like on QB's), but F-Series has more diodes and is driven harder in the datasheet.
48 vs. 72 diodes on the 2footers, 480 vs. 1120mA. They are made of 8 diodes in series, 6 or 9 times in parallel. You can run the F-strips at lower currents and get the same or even better effiency like the H-Series.
And there is no big difference in price, 25pcs. costs you ~250$US.
Also the new Q-Series is out since a few days and already available at digikey, 40 diodes on a 2ft. strip, 200+lm/w. A better H-Series IMO, with newest LM301b.

This is the link to Sammys mid-power LED calculator, but I don't know if the new LM301b is already implemented.
http://www.samsung.com/global/business/led/support/tools/calculator-pkg

Few examples ...?
You could use a HLG-185H-C1050B and 8 2ft. F-Series strips in series for 193,2w with 170lm/w(LER is 321,6, so ~54% system eff.)
Or take 16 strips connected in an 8s2p circuit and use the same driver(525mA per strip, 184lm/w, ~57% system eff., incl. driver!!!)

You could also take a CV/CC driver like the HLG-150H-48A for the 4ft. strips(46v/1120mA). 4 strips would get ~800mA each, that's 150w with ~176lm/w or 55% system eff.....

You can get LER, QER and other values for the LM561c here:
https://horticulturelightinggroup.com/collections/diy/products/qb288-quantum-board

Edit:
Screenie below is calculated with 1, 6 and 12 2ft. strips(2s6p circuit) on a HLG-150H-48A(94% eff., 43-53v/3,2A, at 45v maybe up to 3,33A or more), left side is at 3,2A(59,25mA per diode, 533mA per strip), right side is 3,33A.

Thank you for the detailed answer. Yes, I saw the new Q series, and am very intrigued. My target for any future build would be 200lm/w @4000k and 40-50°C max.

nogod_ said:

The only reason to build with cobs over strips is if you are a spectrum junkie and you dont want to mess around with supplementing a 3000k strip build.

Other than that strips are, watt for watt... cheaper, just as efficient if not moreso, easier to install, easier to manage heat, much lighter fixtures, more uniform ppfd, better "penetration". They dont require heavy heatsinks, thermal compound, nor holders.

You can order strips, a driver, kapton tape, some wire and a cookie sheet and you've got a light.

When I say strips I mean samsung strips (h-series and f-series) specifically not quantum boards.

Click to expand...

"better penetration" ?????............strips use lube?

qwerkus said:

Thank you for the detailed answer. Yes, I saw the new Q series, and am very intrigued. My target for any future build would be 200lm/w @4000k and 40-50°C max.

Click to expand...

not getting that @ system level, take bare chip #s with a grain of salt.............top tested fixtures hovering around 150-160lm/w, and Plight has yet to confirm (2.7umol/j) that I know of?

qwerkus said:

Thank you for the detailed answer. Yes, I saw the new Q series, and am very intrigued. My target for any future build would be 200lm/w @4000k and 40-50°C max.

Click to expand...

I like the idea of getting 200lm/w flux, but I like the 3000°k spectrum more.
But LM561c in 4k can do 200lm/w at ~60mA. The numbers above are incl. driver loss(electr. eff. 94%). But there is still an optical loss too because of reflection and heights.

A 2ft. F-Strip at 540mA needs ~22,5v. This means, if you run the 2ft. strips at ~12w you got 2400lm flux, no optical or electric loss incl. here.
Double that for the double row strips or the 4footers and re-double again for the double row 4ft. monsters, that would be 48w/9600lm at 45v/1080mA.

PSUAGRO. said:

"better penetration" ?????

Click to expand...

yes sir. the more diffuse the light is, the better is its penetration.

For 200lm/w at plant level you would need probably 230lm/w flux or more. impossible with the most chips currently

PSUAGRO. said:

"better penetration" ?????............strips use lube?



not getting that @ system level, take bare chip #s with a grain of salt.............top tested fixtures hovering around 150-160lm/w, and Plight has yet to confirm (2.7umol/j) that I know of?

Click to expand...

Totally agree,
you can calculate 6% driver and 8% optical loss, then Sammy's mid-power calculator spurts out almost real numbers.

tomate said:

yes sir. the more diffuse the light is, the better is its penetration.

Click to expand...

* with sunlight in a greenhouse application

Quantumboard's and strips puts the light out in a more diffuse way as a COB, which almost acts like a point source, although the lights is also getting diffused by the phosphor layer.
So I see profits against single COB's on the quantumboard/strip side.

But if you use more, but lower driven COB's and spread them above the canopy the light becomes also more diffuse because of the multiple light sources.
If you want "outdoor penetration" you need to build a light dome with surrounding strips, boards or COB's. Like a suncloak, you know...?!