Rider's first foray into LED strip lighting: Samsung H Series GEN3

Graytail, I'm curious about your LED system. Is it a DIY? Which light engines are you using and what are you driving them with? Those are respectable numbers against a DE 1000W HPS.

Have you checked out the Samsung F Series Gen3 which uses the LM561C in a 44" double row configuration? It's a 100W beast. With a Vf of 46V and a nominal drive current of 2260mA, up to a max of 3600mA, I'm wondering how best to drive them. I have a couple of spare 320H-C2100B drivers so I could fit three of them on each within the constant current range but they'd be under-driven at 2100mA. They'd run cool and the efficacy would be high, so not a bad combination.
 
It was the strips that first caught my attention for the LM561c. I scoured 'baba for a couple weeks looking at all the options, worked up a spreadsheet, etc, and then checked out the tapes, which led me to the other forum and the DIY stuff being done with tape ...

But then I got a tip about a Chinese company that can produce boards like HLG has, the ones with 304 diodes on a 8"x12" board? So I took the plunge and ordered enough boards to be able to run 'em nice and soft and far away. It's something I've wanted for a couple years. Penetration and canopy depth are so much better if you can cover the sky with light and mount it at the top of the grow space. With the price I paid for these I can do just that. So it's basically the same as 16 3590s in a 4x4, but cheaper and with significantly better efficacy. :slide:

When I was looking at strips, I was thinking of running them at 1050ma, and mounting them an inch or two apart. That would have satisfied my requirements, but then I saw that tape was even cheaper and easier. Like the strips, you can cut them every 2-3 inches depending on design, so I was going to buy a couple sheets of 1/8 aluminum for mounting and heatsink. But then I discovered the Lu/W stats - much lower than they should be - due to the resistors needed to keep the segments at 12 or 24 volts. I seem to recall 140 Lu/W? Bad. Not much better than Citi 1818s. So, back to LM561s in long series for me. They run individually at 2.50 to 3.00 volts, and with 38 in series, the 108 volts that the 304 boards draw is a bit annoying, but I managed to work out combinations of parallel and parallel/series wiring and got away with only four drivers. HLG has another improved version using 288 diodes that's split into two channels in parallel, so that one only runs at 54 volts, and drivers are easier/cheaper.

I'm running mine with a 120H-C700 for 4 boards at 36 watts, a 120HC1050 for 3 boards at 36 watts, a 120H-C1400 for 3 boards at 49 watts and a 240H-C1050 for 4 boards at 55 watts.


Here they are laid out on the floor.
DSCN80741.JPG


And this is how I mounted them. The filter got in the way of a perfect design, and I have that slanted back wall, so I used that to mount three across on the angle - should be interesting for penetration to the rest of the canopy anyway. :cheesygrinsmiley: They run at 36 watts and the ones under the filter at 49 watts. The middle 4 draw 36 watts each and the other 4 are the 55 watters. (Already straightened the cock-eyed one and slipped them all a little farther back.)

DSCN80772.JPG
 
HA! You ungrateful buttdart! I go outta my way to put it in terms even a Canadian can understand... and this is the thanks I get? bwahahahahahaha :rofl:

:theband::Namaste:

It was going well till someone mentioned ABBA at which point I lost my shit...
 
Graytail... that is pure genius. Thank you so much for sharing that with us. You've given me a lot to think about, and I think it's safe to say that the LM561C strips I bought for testing are going to make great shop lights. :)
 
It was the strips that first caught my attention for the LM561c. I scoured 'baba for a couple weeks looking at all the options, worked up a spreadsheet, etc, and then checked out the tapes, which led me to the other forum and the DIY stuff being done with tape ...

But then I got a tip about a Chinese company that can produce boards like HLG has, the ones with 304 diodes on a 8"x12" board? So I took the plunge and ordered enough boards to be able to run 'em nice and soft and far away. It's something I've wanted for a couple years. Penetration and canopy depth are so much better if you can cover the sky with light and mount it at the top of the grow space. With the price I paid for these I can do just that. So it's basically the same as 16 3590s in a 4x4, but cheaper and with significantly better efficacy. :slide:

When I was looking at strips, I was thinking of running them at 1050ma, and mounting them an inch or two apart. That would have satisfied my requirements, but then I saw that tape was even cheaper and easier. Like the strips, you can cut them every 2-3 inches depending on design, so I was going to buy a couple sheets of 1/8 aluminum for mounting and heatsink. But then I discovered the Lu/W stats - much lower than they should be - due to the resistors needed to keep the segments at 12 or 24 volts. I seem to recall 140 Lu/W? Bad. Not much better than Citi 1818s. So, back to LM561s in long series for me. They run individually at 2.50 to 3.00 volts, and with 38 in series, the 108 volts that the 304 boards draw is a bit annoying, but I managed to work out combinations of parallel and parallel/series wiring and got away with only four drivers. HLG has another improved version using 288 diodes that's split into two channels in parallel, so that one only runs at 54 volts, and drivers are easier/cheaper.

I'm running mine with a 120H-C700 for 4 boards at 36 watts, a 120HC1050 for 3 boards at 36 watts, a 120H-C1400 for 3 boards at 49 watts and a 240H-C1050 for 4 boards at 55 watts.


Here they are laid out on the floor.
DSCN80741.JPG


And this is how I mounted them. The filter got in the way of a perfect design, and I have that slanted back wall, so I used that to mount three across on the angle - should be interesting for penetration to the rest of the canopy anyway. :cheesygrinsmiley: They run at 36 watts and the ones under the filter at 49 watts. The middle 4 draw 36 watts each and the other 4 are the 55 watters. (Already straightened the cock-eyed one and slipped them all a little farther back.)

DSCN80772.JPG

Nice,
 
Not sure what I like more... growing, or getting new toys for growing. I've got everything in for this build except the LEDs which should show up tomorrow.
I'm going to do a head to head with four CXB3590s on a 320H-C2100B, six of the LM561C strips on the same driver, and eight LM561C strips on this new 185H-C1050B.
The proof is in the PAR!

420-magazine-mobile1088888385.jpg
 
Awwww yeaahh! I'm glad your here to do this cause I'm definitely not smart enough XD looking forward to your experiment results ! :cheesygrinsmiley:
 
Hey rider I subbed up back on page one. This stuff is a little over my head honestly, but I'm trying. Are these Samsung units comparable to the quantum boards? I also saw u mention a much cheaper price. I am looking to light a (roughly) 4.5 by 4 area for flowering how many panels would you suggest?
 
That's tough to answer because there are a number of ways to approach it. My goal, besides the fun factor of new toys, is to exceed the specs of my sixteen cob Cree CXB3590 build while providing a more uniform distribution of light. It's generally accepted that one cob provides coverage for one foot square, so sixteen cobs for a 16 foot square space similar to yours. But then you have to decide how hard you want to run them in order to choose an appropriate driver. I went balls out and decided to maximize the potential of the cobs and am able to run each group of four cobs to 300 watts. In total I'm capable of pushing the system to 1200 watts at the lights. Of course it's a bit more at the wall due to system losses. I'm running four Meanwell HLG-320H-C2100B drivers. Some consider it overkill but it got my PAR readings up to the range where additional CO2 has a beneficial effect.

So to duplicate the system I built for a 4'x4' tent would require 24 of the Samsung units; four groups of six running on the four drivers I already have. You can see that the Samsung strip cobs are respectable performers and exceed the Cree cobs in output and efficacy.

..................flux.......Vf......W ....m/W
CXB3590.....44263...143...300...147.4
LT-HB22D....49173...143...306...160.5

Admittedly, it's not a huge difference, but I'm hoping to see better penetration with the more diffuse light. The other benefit is that without the concentrated hotspots of the Cree/Vero/Citizen type round cobs I should be able to push the system a bit more without cooking the plants.

The second option is to not build a top-fuel dragster just because you can. Using the same 24 Samsung units but driving them in three groups of eight with three HLG-185H-C1050B drivers still provides 600 watts of power at the lights. That's more usable light energy, owing to an efficacy of 176.7 lm/W, than you'll get with a 600 watt MH/HPS and at a much lower heat index. And you can run the LEDs much closer to the canopy which increases the PPFD at the plants.

It's a generally accepted axiom that you need 35 watts of light energy per square foot of grow space. My Cree cob system is capable of 75 watts per square foot as is my preferred Samsung design. A 600 watt HPS bulb is 37.5 watts per square foot. The second option also comes in at 37.5 watts but at a much lower operating cost, less heat, and with no bulb to change every season.

No matter how ya slice it though, an LED system is going to put a bite on the wallet. Figure $420 for 25 3500K Samsungs, $150 for the three Meanwell drivers, and the cost of the goodies to mount and wire it all. So about $600 to $650 for an HPS killing beast.

edit: the reason I quoted the price at 25 units is because of the quantity price break at that number
 
Hey rider I subbed up back on page one. This stuff is a little over my head honestly, but I'm trying. Are these Samsung units comparable to the quantum boards? I also saw u mention a much cheaper price. I am looking to light a (roughly) 4.5 by 4 area for flowering how many panels would you suggest?

I answered the "how many" in the post above with just two designs for a roughly 4'x4' space, one at the accepted minimum of 35W ftsq and the other at 75W ftsq. If I was starting from scratch I'd probably design to the 50W ftsq mark. There are many ways to do that with different series/parallel configurations and different drivers. It's not something I've looked into because I was designing around drivers I already had. Keep in mind also that the 35W ftsq minimum axiom is from the days when HPS was the standard choice and that other factors come into play with LED. The true test of a system is how much PAR light gets put down on the canopy.

As for a comparison to the Quantum boards, they both use the same Samsung 561C LED chips. A QB288 board has the same number of LEDs as three of the 44" LT-HB22D light strips. Some of the initial design parameters have been made for you with the QB boards so in multiple unit configurations the design and DIY assembly is a little easier. The QB boards are awesome, reasonably priced, and hard as hell to come by. They tend to sell out each production run pretty quickly.
 
I spent some time last night looking at the branded Samsung strips, and I was impressed. :cheesygrinsmiley:

The cost initially put me off because I was working on a Citi 1818 design at that time, and they're a pretty reasonable cost. And then I immediately saw all the offerings from the Chinese at $10 or less using the same chips, so I spent my time researching those. But the Lu/W is much better on the LT-HB22Ds at 170+, and they run 96 diodes instead of the 60-72 that the cheap ones run. :hmmmm:

I can see why they kept your interest. 25 strips at 25 watts would draw similar current as my 14 boards at 45 watts. Your example costs $600-700, while mine was just under $800 with sinks and framing. You'll have a hundred or so into framing?

So, very similar cost. :thumb:

Ideally, I'd like to add a couple strips to my design later, loaded with reds and far reds for Emerson effect, etc.
 
I just love reading all this up, one day I'll understand it XD good work guys :cheesygrinsmiley:
 
Framing costs are still up in the air. I'm either going to go with 3/8" 6061 flat stock or 1/8" 6061 flat stock with L-channels as support. I'm leaning towards the 3/8" because then I only have to drill four holes. Also, I can round the corners so I don't dent my noggin on a sharp corner. :)

So $100 for four units is about right.
 
Well hell. I got the order in from Digikey but there's a problem. And it's not even Digikey's doing! They're being very helpful in resolving the issue.

I ordered the H Series Gen3 but what showed up is the M Series Gen2 with an H Series ID sticker. The PCBs identify as M, the form factor is M, but the manufacturer applied sticker says H. I'm bummed.

420-magazine-mobile866118383.jpg


420-magazine-mobile4283304.jpg
 
Finally had a chance to read the "fluence" articles. A great read for an LED newb like me. I was a little overwhelmed with all the terminology and units used to describe LED lighting but this basic understanding of PAR, PPF and PPFD now gives me a good starting point to work from. That article answered 3 questions but now I have 10 more. I suppose that s a good thing though. Thanks Rider.
 
Finally had a chance to read the "fluence" articles. A great read for an LED newb like me. I was a little overwhelmed with all the terminology and units used to describe LED lighting but this basic understanding of PAR, PPF and PPFD now gives me a good starting point to work from. That article answered 3 questions but now I have 10 more. I suppose that s a good thing though. Thanks Rider.

I wrote my descriptions in a way that built on terms sequentially. It's hard to understand what "photosynthetic photon flux" means if you don't understand each term. There's a guy with a great blog, the LED Gardener, that takes it to a whole 'nother level of well explained and documented LED system exploration. I've chatted with him and he's a hell of a nice guy. Def check him out.
 
Back
Top Bottom